BRIDGE MAINTENANCE, SAFETY, MANAGEMENT, HEALTH MONITORING AND INFORMATICS

PROCEEDINGS OF THE FOURTH INTERNATIONAL CONFERENCE ON BRIDGE MAINTENANCE, SAFETY AND MANAGEMENT, SEOUL, KOREA, 13–17 JULY 2008

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics

Edited by

Hyun-Moo Koh Seoul National University, Seoul, Korea Dan M. Frangopol Lehigh University, Bethlehem, PA, USA CRC Press/Balkema is an imprint of the Taylor & Francis Group, an informa business

© 2008 Taylor & Francis Group, London, UK

Typeset by Charon Tec Ltd (A Macmillan Company), Chennai, India Printed and bound in Great Britain by Antony Rowe (A CPI-group Company), Chippenham, Wiltshire

All rights reserved. No part of this publication or the information contained herein may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, by photocopying, recording or otherwise, without written prior permission from the publishers.

Although all care is taken to ensure integrity and the quality of this publication and the information herein, no responsibility is assumed by the publishers nor the author for any damage to the property or persons as a result of operation or use of this publication and/or the information contained herein.

Published by: CRC Press/Balkema P.O. Box 447, 2300 AK Leiden, The Netherlands e-mail: [email protected] www.crcpress.com – www.taylorandfrancis.co.uk – www.balkema.nl

ISBN 13: 978-0-415-46844-2 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Table of Contents

Preface XXXV Conference organization XXXVII

T.Y. Lin Lecture Fatigue of steel bridge infrastructure 3 J.W. Fisher & S. Roy

Keynote Lectures Managing seismic performance of highway bridges – Evolution in experimental research 13 M. Saiidi Cost-effective and durable cable-stayed bridges 17 H. Svensson A new concept of orthotropic steel bridge deck 25 M.-C. Tang Health monitoring of structures & related education and training needs of civil engineers 33 A.A. Mufti & J.P.Newhook Practical implementation of probability based assessment methods for bridges 41 I. Enevoldsen Bridge monitoring in Japan: The needs and strategies 49 Y. Fujino & D.M. Siringoringo Bridging capacity innovations on cable-supported bridges 53 Y.J. Ge & H.F.Xiang Overcoming technological challenges to create new values for bridges 63 S.P.Chang

Technical Contributions

Advanced and high performance materials Durability of bridges made of advanced composite materials 83 J.R. Correia, F.A.Branco, J.G. Ferreira, S. Cabral-Fonseca, M.I. Eusébio & M.P.Rodrigues V Analytical study on the performance of reinforced high-strength concrete bridge columns 85 D.J. Seong, H.M. Lee, H.M. Shin, J.H. Choi & M.S. Oh Use of steel fibre concrete to eliminate shear reinforcement in pretensioned concrete beams 87 P.De Pauw, L. Taerwe, N. Van den Buverie & W. Moerman Durability of Structural Composite Lumber in bridge applications 88 N. Yazdani, E.C. Johnson & S. Duwadi Development of Self-Consolidating Concrete for bridge construction and repairs 89 P.Paczkowski, A.S. Nowak, G. Morcous & M. Kaszynska On the use of duplex stainless steels in bridge construction 90 O. Hechler & P.Collin Self-consolidating lightweight concrete – Excellent material for bridge applications 91 M. Kaszynska

Assessment and control of bridge vibrations Detection of bridge damages by recognition of non-linear dynamic effects 95 H. Wenzel Experimental dynamic analysis of steel concrete composite railway bridges: The Sesia viaduct on the high speed line Turin-Milan 97 G. Chellini, L. Nardini, W. Salvatore, G. De Roeck, K. Liu, E. Reynders, B. Peeters, M. Tisalvi & G. Sorrentino Excitation of pedestrian structures during marathon events 98 C. Sahnaci & M. Kasperski A novel active mass damper for vibration control of bridges 99 U. Starossek & J. Scheller Analysis of fuzzy active control for traffic-induced vibration of highway girder bridge 100 M. Kawatani, Y. Nomura, C.-W. Kim & Y. Otsubo Sensitivity-based optimal design of damper connecting system for vibration control of parallel bridges under wind excitations 103 D.-S. Kim, S.-Y. Ok, K.-S. Park, H.-M. Koh & C.-Y. Choi

Bridge 200 toward durable bridge Research project “Bridge 200” 107 B.S. Kim, S.Y. Kim, K. Cho, J.R. Cho & S.T. Kim Development of FRP bridge decks in Korea 108 K.-T. Park, Y.-H. Lee, J. Jeong & Y.-K. Hwang Development of FRP-concrete composite bridge deck system 109 S.Y. Park, K. Cho, J.R. Cho, S.T. Kim & B.S. Kim Development of Ultra High Performance Cementitious Composites (UHPCC) in Korea 110 S.W. Kim, J.J. Park, S.T. Kang, G.S. Ryo & K.T. Koh VI Flexural strengthening of RC structures with externally unbonded prestressed CFRP plates 111 Y.H. Park, J.S. Park & W.T. Jung Bridge scour countermeasures to minimize bridge failures during floods 112 J. Lee, J. Park, M. Chung & K. Kwak

Bridge codes Distribution of demand in single-column-bent viaducts with irregular configuration in longitudinal direction 117 R. Akbari & S. Maalek Effective length factor of X-bracing system 118 J. Moon, H.-E. Lee & K.-Y.Yoon Monitoring system of suspension bridges and the utilization of recorded data 119 C. Kawatoh, S. Kusuhara, S. Fukunaga & K. Endo Reliability analysis of composite girder under positive and negative flexure designed by LRFD method 120 D.K. Shin, J.S. Roh & E.Y. Cho Flexural design of prestressed high-strength concrete girders 121 W. Choi, H.C. Mertol, S. Rizkalla, P.Zia & A. Mirmiran Strength prediction on the stiffened plates in compression 122 Y.B. Kwon, D.W. Kang, B.H. Choi & T.Y.Yoon Strength of fillet welded splices of SM570-TMC, extra thick plates 123 J.B. Jo & J.W. Kim Safety of ductility demand based seismic design for bridge columns 124 J.-H. Lee, J.-H. Choi, J.-K. Hwang & H.-S. Son An experimental investigation of the ultimate flexural behavior of steel tub girders with top lateral bracing 125 B.H. Choi, T.Y.Yoon & Y.S. Park

Reliability based calibration of limit state bridge design code with material and member resistance factors 126 I.Y. Paik & D.J. Bang

Bridge E-8 in the new railroad of high velocity to the northwest of Spain 127 C. Jurado

A simple iterative method for determining the effective length of structural members in steel cable-stayed bridges 129 D.-H. Choi, H. Yoo, D.-S. Lee & Y.-S. Kim The influence of friction/sliding behavior of rubber bearing to the seismic performance of highway bridges 130 K.Y. Liu, K.C. Chang, W.I. Chen & J.S. Hwang Ultimate flexural strength of hybrid composite girders at sagging bending 131 S.G. Youn,Y.T. Kim & D.B. Bae VII Bridge inspection and diagnostics Application of data fusion technology on scour and siltation monitoring in river bed 135 Z.J. Chen, W.Wang, S. Chen & S.H. Cao Special tests of two post-tensioned concrete viaducts 136 J. Ciesla, M. Lagoda & P.Olaszek Active lamb wave propagation-based damage detection and location for steel plate 137 W. Jeong, J. Seo & H. Kim Settlement prediction model for pile foundation based on field observation 138 X. Li, Z. Chen & X. Dong Ambient vibration of stay cables used for damage detection in cable-stayed bridge 139 C.-C. Chen, W.-H.Wu & J. Lin Specials inspections and maintenances of prestressed concrete in Rio-Niterói Bridge 140 C.H. Siqueira Non-destructive testing of suspender ropes with magnetostriction 141 M.S. Higgins & O. Tozser Judging suitability of arch bridges for higher axle loading by load testing 142 R.K. Gupta Health assessment of pre-stressed concrete girder bridges by non-destructive testing 143 R.K. Gupta Optimal inspection and maintenance strategies for bridge network using supply and 144 demand approach A.D. Orcesi & C.F.Cremona Experimental investigations on the strength behavior of box beam and circular column connections 146 Y.P.Kim & W.S. Hwang Nondestructive evaluation of effective prestress using the core-drilling method 147 S. Pessiki & M.J. McGinnis Bridge safety management system by using Bridge Inspection Robot 148 D.-j. Park, H.-g. Jung, B.-j. Lee, W.-t. Lee & J.-o. Kim Matrix based cable-stay bridge cable force and deck elevation adjustments and FEM updating 149 A. Turer A regularization scheme for displacement reconstruction using measured structural acceleration data 150 Y.H. Hong, H.W. Park & H.S. Lee Dynamic testing of existing bridges for high speed trains 152 A. Turer Concrete bridge deck condition assessment with automated multisensor techniques 153 D. Huston, J. Cui, D. Burns & F.Jalinoos Parameter estimation of concrete bridge using ambient acceleration measured by wireless measurement system 154 S.J. Lee, S.B. Kim, K.Y. Choi, G.Y. Song, D.O. Kang & Y.H. Lee VIII Impact-Echo scanning for grout void detection in post-tensioned bridge ducts – Findings from a research project and a case history 155 Y. Tinkey & L. Olson

Bridge management systems Design and implementation of a new bridge management system for the Ministry of Transport of Québec 159 R.M. Ellis, P.D. Thompson, R. Gagnon & G. Richard Comprehensive lightning protection technologies for mechanical and electrical systems of Sutong bridge 161 B. Yao, W. Zhang, G. Chen & C. Jiang Risk evaluation and management for road maintenance on urban expressway based on HELM (Hanshin Expressway Logic Model) 162 Y. Sakai, K. Kobayashi & H. Uetsuka Asset management system development 163 J. Radic, J. Bleiziffer & G. Puz Development of a smart-client based bridge management and maintenance system for existing highway bridges 164 D.Shan&Q.Li A new Bridge Management System for the National Department of Transportation of Argentina 165 M.E. Ruiz, E.A. Castelli & T.A. Prato Development of a bridge maintenance decision support module for Taiwan Bridge Management System 166 H.-K Liao, C.-I. Yen & N.-J. Yau Optimization of bridge management policies on the French national roads network 167 N. Odent, J. Berthellemy, C.F.Cremona, A.D. Orcesi & M. Toriel Bridge management: A challenge for local authorities 169 B.M. Kamya A condition index based on the concept of apparent age 170 D. Zonta, F.Bortot & R. Zandonini Design of the standardized measuring system for the integrated safety management of bridge structure 171 W.S. Lee & K.T. Park

Bridges for high-speed railways Steel bridges for high speed railways – Design regarding fatigue and durability 175 W. Hoorpah Dynamic response of the Cahir Viaduct – An investigation into the derailment of a freight train 176 M. Majka & M. Hartnett A comfort limit for evaluating the serviceability due to bridge vibration 177 B.G. Jeon, N.S. Kim & S.I. Kim IX Vibration control through TMDs in high-speed railway bridges 178 J.F.Henriques & J.M. Proença DETAILS: A research project for improvement of analysis, design and durability of HS railway bridges 179 G. Chellini, L. Nardini & W. Salvatore Dynamic testing and numerical modelling of a typical short span high-speed railway bridge 180 V.Zabel & M. Brehm Development of an efficient finite element model for the dynamic analysis of the train-bridge interaction 181 S. Neves, A. Azevedo & R. Calçada Fatigue assessment of composite bridges for high speed railway traffic 182 H. Figueiredo, R. Calçada & R. Delgado Experimental modal analysis of a twin composite filler beam railway bridge for high-speed trains with continuous ballast 183 T. Rauert, B. Hoffmeister, R. Cantieni, M. Brehm & V.Zabel A study of the lateral dynamic behaviour of high speed railway viaducts and its effect on vehicle ride comfort and stability 184 R. Dias, J.M. Goicolea, F.Gabaldón, M. Cuadrado, J. Nasarre & P.Gonzalez

Damage assessment Evaluating composite steel girder-concrete slab bridge beams using simplified plastic analysis 187 P.S.McCarten Improving bridge component deterioration forecasting precision 188 H.S. Kleywegt Stochastic subspace-based structural identification and damage detection – Application to a long span cable-stayed bridge 189 W.Zhou&H.Li Structural safety of historical stone arch bridges in Korea 190 N.K. Hong, H.-M. Koh, S.G. Hong & B.S. Bae Realistic estimation method of moment redistribution in reinforced concrete beams based on the analytical methods 191 J.H. Cheon, J.G. Park, S.C. Lee, M.S. Oh & H.-M. Shin Probabilistic analysis of the structural behaviour of a bridge prestressed concrete beam 192 C. Cremona, S. Mohammadkhani-Shali, B. Richard, C. Marcotte & B.Tonnoir Gi-Lu Cable Stayed Bridge – From earthquake damage to full recovery 194 Z.K. Lee, K.C. Chang, C.C. Chen & C.C. Chou Stress monitoring of steel girder bridges with different boundary conditions 195 N. Namatame, K. Tani, T. Tsuji, M. Kawatani, M. Kano, N. Tanaka & H. Hattori Evaluation of compressive stiffness of elastomeric bearings 197 H.J. Yoon,Y.J. Kim, C.B. Cho & I.J. Kwahk X Effect of soil-bridge interaction and continuity on live load distribution in integral bridges 199 M. Dicleli & S. Erhan Application of the analytic hierarchy process in performance evaluation of existing concrete cable-stayed bridge 200 Q. Li, D. Shan & W.Yan Analysis of box girder bridges using finite elements and AASHTO-LRFD 201 R.R. Doerrer & R.A. Hindi Nonlinear finite element analysis of precast segmental prestressed concrete bridge piers 203 H.-M. Lee, D.-J. Seong, J.-G. Park, K.-S. Kim, H.-M. Shin, T.-h. Kim, Y.-J. Kim & S.-W. Kim Advanced numerical study of asphaltic surfacings on orthotropic steel deck bridges 204 X. Liu, T.O. Medani & A. Scarpas Estimation of damping characteristics for cable using system identification scheme 205 S.-K. Park, K.W. Lyu & H.S. Lee Self-adapting models of bridge degradation 206 J. Bien & A. Banakiewicz Computational model generation based on 3D CAD digital data of RC bridges 208 J. Lee & M.-S. Kim Artificial Intelligence: Historical development and applications in civil engineering field 210 L. Sgambi Safety factor prediction for steel cable-stayed bridges by iterative eigenvalue analysis 211 D.-H. Choi, H. Yoo, D.-S. Kim & H.-S. Na

Damage assessment of existing bridges Evaluation of load-carrying capacity of the damaged bridge model using the updated FE model 215 D.S. Jung & C.Y. Kim Detection of sudden damages of structure by regularized autoregressive model using measured acceleration 216 J. Kang & H.S. Lee Internal damage localization in a thick plate using moving sensing windows 217 Y.H. Kim, H.W. Park, J.W.Whang & H.S. Lee Hybrid health monitoring technique for PSC girders using wireless sensing and embedded monitoring algorithm 218 J.H. Park, J.T. Kim, Y.S. Ryu, D. Mascarenas & M.D. Todd Regularization of inverse problem for damage detection 219 I. Yoshida, C.W. Kim & M. Kawatani Feasibility investigation of health monitoring from traffic-induced vibration data of bridge 221 M. Kawatani, C.-W. Kim & T. Fujimoto XI Localization of damage in a bridge using measured response signals 223 S. Shin & H. Park

Design and analysis Stonecutters bridge – Design for operation 227 M. Carter & N. Hussain Strut-and-Tie Method for FRP strengthened deep RC members 228 S. Park & R.S. Aboutaha Design of the Machang Mainbridge 229 E.-H. Bae, S.-Y. Kim, R.-G. Kim, S. Hopf, A. Patsch & P.Walser Durability of suspension bridge with multi main spans 230 M. Inoue, M. Kudo, K. Doi & Y. Takizawa Effect of main steel corrosion on the stiffness of corroded reinforced concrete beams 232 F.J. O’Flaherty, P.S.Mangat, P.Lambert & E.H. Browne Optimum life-cycle-cost design for bridge structures considering damage probability 233 Y.S. Shin, J.H. Park & T.H. Kim The study on the methods for slimming bridge 234 K.H. Kim, C.S. Lee, K.S. Hong & C.-G. Lee Design of deck continuity details for steel and prestressed concrete bridges 235 A.P.Ranasinghe & W.L. Haugeto Reliability-based optimum design of high-speed railway bridges considering structure-rail interaction 236 J.-S. Lee, H.-N. Cho & Y.-R. Ihm Cyclic response of the precast SRC bridge piers 237 Y.-S. Chung, C.S. Shim, J.-Y.Yoon & J.-H. Park General design method for the shear connection according to failure modes 238 C.S. Shim, S.-M. Jeon & P.-G. Lee Structural system durability through jointless bridge decks 239 U.B. Attanayake, A.E. Ulku & H.M. Aktan Modal flexibility-based FEM model updating for bridges 241 J. Cui, D. Kim, K.Y. Koo & H.Y. Jung Optimal design of a steel box girder bridge considering aesthetics 243 Y.S. Shin, J.H. Park & G.O. Kim Complex shapes and innovative technologies for bridges 244 I. Paoletti Reliability-based design optimization using a response surface method 245 S.C. Kang & H.-M. Koh Durability design procedure of concrete structures in Korea using partial safety factor format 246 J.S. Kim, K.J. Shin, J.H. Kim, K.M. Lee & S.H. Bae Analysis of static behavior of CFTA girder 247 H. Lee, K.-H. Park & J.-S. Kong XII Durability analysis of RC bridges using Monte-Carlo simulation 248 W. Raphael, R. Faddoul & J. Hokayem System stability design of cable-stayed bridges based on elastic/inelastic system buckling analyses 249 Y.-S. Kyung, J.-S. Lee & M.-Y. Kim Experimental research on passive cable dampers’ performance 251 S.-s. Ahn, J.-H. Park & S. Lee Conformity control of concrete based on the “concrete family” concept 252 R. Caspeele & L. Taerwe Effective slab width in steel-concrete composite girder bridges 253 D. Bae, S.G. Youn & Y.S. Park Preference-based optimal maintenance planning for deteriorating bridges 255 S.Y. Lee, W. Park, H.-M. Koh & H.J. Kim Analysis of steel-soil bridge structure made of corrugated plate 256 D. Beben & Z. Manko Introduction of design for concrete filled steel tubular arch girder with external tendons 258 E. Lee, H. Park, M.G. Park, K.H. Park, S.Y. Lee & J.H. Kim

Design expectations, monitoring response and maintenance decisions Design expectations and monitoring response of the Certosa fly-over in Milan (Italy) 263 P.G. Malerba, A. Giussani, G. Pezzetti & C. Malerba Statistical monitoring of concrete structures and cable-stayed bridges 264 A. Fassò & G. Pezzetti Short and long term monitoring for maintenance and retrofitting of existing bridges 266 C. Modena, P.Franchetti, M. Grendene & M. Frizzarin The structural analysis of the Messina Strait Bridge 267 F.Bontempi Influence of large displacements on the structural stability of cable supported bridges 268 F.Biondini, P.Limonta, P.G. Malerba & R. Stucchi Wind-induced fatigue assessment in main cables and hangers of suspension bridges 269 F.Ubertini & F.Bontempi

Development of the advanced robot systems for bridge inspection and monitoring Introduction of the Bridge Inspection Robot Development Interface (BIRDI) 273 J.S. Lee, I. Hwang, H.S. Lee & S.H. Hong Robotic diagnosis system for detection of bridge structures 274 D.-J. Moon, K.-T.Yang, S.-S. Nam & K.-H. Im Development of bridge inspection robot system: Wall climbing robot and flying robot 275 I.M. Koo, C.M. Lee, S.-H. Whang, D.-H. Kim, M.-S. Kang, K. Cho, W.-H. Son, S. Park, S.K. Park & H.R. Choi XIII Bridge inspection robot system with novel image processing 276 J.-K. Oh, A.-Y. Lee, S.M. Oh, Y. Choi, B.-J. Yi, H.W.Yang, J.H. Lee & Y.S. Moon Intelligent bridge management system based on the image data from robotic devices 277 S. Kim, J.S. Lee, Y. Choi & Y.S. Moon Optimum NDT using infrared thermography for defected concrete 278 G. Zi, J.G. Sim, H. Oh & J. Lee A development of repair mechanism and control technologies for bottom part of concrete bridge 279 K.-Y. Lim

Fatigue analysis Evaluation of fatigue strength of one riveted historical railway bridge 283 A. Pipinato, C. Pellegrino & C. Modena Global-local finite element analysis of riveted railway bridge connections for fatigue evaluation 284 B.M. Imam, T.D. Righiniotis & M.K. Chryssanthopoulos Fatigue of riveted metal structures 285 T. Larsson & O. Lagerqvist Managing fatigue susceptible details on critical railway bridges at CN 286 J.A. Cavaco Influence of the fatigue resistance of duplex steel on the bridge design – A span length investigation 287 T. Rauert, B. Hoffmeister, A. Gieseking & O. Hechler Fabrication procedure effects on fatigue resistance of steel orthotropic deck welds 288 H.B. Sim, C.M. Uang & C. Sikorsky Estimation of low-cycle fatigue strength of steel structural members under earthquake loading 289 J. Iyama & J.M. Ricles Variability analysis of fatigue crack growth rates of materials from ancient Portuguese steel bridges 290 J.A.F.O. Correia, A.M.P.Jesus, M.A.V.Figueiredo, A.S. Ribeiro & A.A. Fernandes Assessment of the coupled effect of corrosion-fatigue on the reliability of RC bridges 292 E. Bastidas-Arteaga, M. Sánchez-Silva, Ph. Bressolette, A. Chateauneuf & W. Raphael An optimal design of TMD for the improvement of fatigue reliability of steel-composite high-speed railway bridges using target performance approach 293 S.-J. Kim, S.-C. Kang, H.-M. Koh & W. Park Fatigue damage of orthotropic steel bridge decks and its retrofit 294 T. Shimozato, T.Yabuki, Y. Arizumi, Y. Hirabayashi, N. Inaba & S. Ono Fatigue design for highway bridge ancillary structures 295 Y.C. Park, S. Roy & R. Sause XIV Fiber reinforced composites in bridges On mechanical performance of different type of FRP beams as reinforcement of pedestrian bridge 299 G. Boscato & S. Russo Shaping composite bridges for traffic and the environment 300 R.A. Daniel A study on the dynamic behavior of a CFRP cable 301 J. Park & K.J. Hong Composite ‘Delta Deck’: The promising bridge deck for new and rehabilitated bridges 302 S.W. Lee & K.J. Hong Advanced removable connection for glass fiber reinforced polymer bridges 303 D.-U. Park, K.-J. Hwang & J. Knippers Full pultruded FRP profile structures 304 M.D.G. Pulido

Health monitoring Detachable sensor for bridge cable maintenance and safety 307 S. Sumitro, H. Hoashi, T. Okamoto & M.L. Wang Stock Condition Index analyses response to Bridge Condition Index determinations 308 P.S.McCarten Traffic-induced vibration of bridges: Input force identification 309 C.-H. Loh, A.-L. Wu & J.-H. Weng Turning the Humber Bridge into a smart structure 310 N.A. Hoult, P.R.A.Fidler, C.R. Middleton & P.G. Hill Application of data fusion in the safety monitoring of Sutong Bridge foundation 311 Z.J. Chen, L. Bian, T. Xue & X.W. Zhang Remote measurement of crack length in sacrificial test piece by self-reference lock-in thermography 312 Y. Sakino, T. Sakagami & Y.-C. Kim Wind-induced vibrations and countermeasures for cable systems on long-span bridges 313 I. Yamada, S. Kusuhara, K. Fumoto & N. Toyama Structural monitoring of a bridge prestressed concrete beam under loading 314 C. Cremona, C. Tessier,V.Le Cam, B. Tonnoir, R. Leconte & V.Barbier Application of InSAR on ground deformation in the location of Sutong Bridge 316 Z.J. Chen, N.N. Zhang, X.Y. Li & L.Y. Feng Long-term structural health monitoring of the Torino’s pedestrian cabled-stayed bridge 317 L.M. Giacosa & A. De Stefano Identifying bridge damage using Brillouin Optical Fiber Sensing 318 W. Zhang, B. Shi, Y.Q. Zhu & Y.F.Zhang Innovative treatment of monitoring data for reliability-based structural assessment 319 T.B. Messervey & D.M. Frangopol XV Life-cycle monitoring of the structural configuration of a suspension bridge 320 H.-K. Kim, H. Lee, J.-H. Jang, Y.-H. Kim & S.-K. Ro Bridge monitoring while partial demolition under traffic 321 K. Zilch, E. Penka, M. Hennecke, U. Willberg, Th. Wunderlich & Th. Schäfer Development and implementation of a low-cost, continuous bridge health monitoring system 322 Y.-S. Lee, B.M. Phares & T.J. Wipf Structural health monitoring of civil infrastructures using MEMS-based technologies 324 T. Miyashita & M. Nagai Error-resilient routing for wireless SHM powered by solar cells 326 J. Ryu, J. Kim, I. Yeo, Y. Cho & H. Shin Correlation analysis on long term monitoring data of Donghai Bridge 327 L. Sun, Z. Min & D. Dan Low-latency routing for wireless SHM powered by solar cells 328 J. Kim, Y. Cho, J. Ryu, I. Yeo & H. Shin Health monitoring for corrosion detection in reinforced concrete bridges 329 A. Del Grosso, F.Lanata, L. Pardi & A. Mercalli Investigation of the relationship between displacement and acceleration in nonlinear dynamics using chaos theory analysis 330 R.A. Livingston & S. Jin The application of the frequency-shifted feedback laser optical coordinates measurement system for field measurement of bridges in service 331 S. Umemoto, N. Miyamoto, K. Kubota, T. Okamoto, T. Hara, H. Ito & Y. Fujino Bridge-Weigh-In-Motion based on strain measurement of vertical stiffeners 332 E. Yamaguchi, Y. Naitou, K. Matsuo, Y. Matsuki & S. Kawamura An FIS and AHP based on line evaluation system on Donghai Bridge 333 D. Dan, L. Sun, Z. Yang & D. Xie Structural health monitoring of complex structural systems using adaptive models 336 S. Arangio

Information technology for lifetime management of bridge A Strategy for IT-based lifetime management of bridge 339 S.-H. Lee, B.-G. Kim, H.-J. Kim & S.-J. Kim Use of information technology in a regional bridge management contract 340 R. Kiviluoma & M. Tervo Development of a structural health monitoring system with wireless sensor networks 342 H. Emoto, A. Miyamoto & K. Kawamura Development of a multi-purpose remote health monitoring system for existing bridges 344 A. Miyamoto, J. Sonoda & K. Kawamura Automated identification of modal properties in a steel bridge instrumented with a dense wireless sensor network 345 A.T. Zimmerman, R.A. Swartz & J.P.Lynch XVI Downsizing seismic sensing system and its implementation 346 Y. Mizuno & Y. Fujino SHM sensor networking with remote powering and interrogation 348 M.D. Todd, D. Mascarenas, E. Flynn, B. Lee, K. Lin, D. Musiani, T. Rosing, R. Gupta, S. Kpotufe, D. Hsu, S. Dasgupta, G. Park, K. Farinholt, M. Nothnagel & C. Farrar Development of an advanced inspection system for weathering steel bridges based on digital image recognition 349 S. Goto, T. Aso & A. Miyamoto A study on the self-anchored suspension bridge behavior using GPS 350 H.J. Ham, S.H. Oh, I.H. Bae & G.H. Ha SHM role in Bridge Life Cycle Analysis (BLCA) 351 S. Alampalli & M. Ettouney Interoperable information model based on IFC for the cable-stayed bridge monitoring system 352 J.-H. Yi, H.J. An, H.-J. Kim & S.-H. Lee An automatic crack recognition system for concrete bridge inspection by image processing approach 353 A. Miyamoto

Innovative construction technology Erection of asymmetric pylon table and geometry control of Machang cable-stayed bridge 357 H. Lim, M. Kim & J. Seo Experimental study on the stability of temporary support for girder construction 358 K. Ohdo, S. Takanashi & H. Takahashi Design of Cheong-Poong (steel-concrete hybrid cable-stayed) bridge 359 D.-H. Yoo, J.-S. Ko & J.-G. Paik Influence of initial imperfections on stability of temporary support for bridge girder 360 H. Takahashi, K. Ohdo & S. Takanashi Analysis for initial equilibrium condition and erection stages of Sorok (Self-Anchored Suspension) Bridge 361 Y. Son, D. Yoo, S. Jeong & T.Yoon Seosang bridge movable scaffolding system 362 A.A. Póvoas The construction of Machang cable-stayed bridge 363 M. Kim, J. Seo, J. Song & H. Lim Construction planning and analysis of six continuous extradosed PSC bridge 364 S. Kim, J.W. Seo, Y. Lee & I. Seo Underpass in the street O’Donnell in Madrid 365 C. Jurado Construction of P.S.C. Box girder bridge (F.C.M) which pre-compensation method is applied to 367 H. Jee, J.H. Shim, M.K. Min & H.J. Lee XVII Large scale cyclic tests of precast segmental concrete bridge columns with unbonded post-tensioning tendons 368 Y.-C. Ou, G.C. Lee, P.-H.Wang, M.-S. Tsai & K.-C. Chang Key-segment closing method using artificial heat for partially earth-anchored cable stayed bridges with classical span length 369 J.H. Won, K.I. Cho, J.H. Yoon & S.H. Kim Highway bridges made of circular hollow sections 370 U. Kuhlmann & M. Euler Monitoring of early stage prestress change of long span stream-cured concrete box girder with pretension method 371 K.-Y. Choi, G.-Y. Song, D.-O. Kang & S.-W. Cha Effect of deformation of spans on serviceability of composite highway bridge 373 Z. Manko

Integrated assessment – Practical application of probabilistic methods Guideline for probabilistic assessment of deteriorated bridges 377 J. Lauridsen, F.M.Jensen & S. Engelund Optimization of special inspections of concrete bridges 378 S. Engelund & M. Sloth Probability based assessment of railway bridges in Denmark – Previous applications, current state and future possibilities 379 J.S. Jensen, D.F.Wisniewski & O.B. Ulstrup Probability based assessment of motorway bridges in Denmark 380 J. Bjerrum, A. O’Connor, C. Pedersen & I. Enevoldsen Probability based assessment of a large riveted truss railway bridge 382 A. O’Connor, C. Pedersen, I. Enevoldsen, L. Gustavsson & J. Hammarbäck

Integrating health monitoring and lifecycle management of bridge and highways Integrating human, natural and engineered systems and associated paradigms for infrastructure asset management 387 F.Moon, P.Gurian, F.Montalto & A.E. Aktan Structure and infrastructure health monitoring as a key enabling paradigm for integrated asset management 389 F.L. Moon, A.E. Aktan, F.Jalinoos & H. Ghasemi Integration of health monitoring in asset management in a life-cycle perspective 391 T.B. Messervey & D.M. Frangopol Challenges for asset management in W-Europe 392 L.(H.E.) Klatter Effective bridge management using ABMS 393 H. Kawamura, K. Kudo, M. Soma, H. Kawaragi & M. Kaneuji Development of BMS and possibility of performance based contracting using BMS 394 M. Kaneuji, H. Kawamura, M. Soma & E. Watanabe XVIII A study on LCC prediction for bridge management taking future uncertainty into account 396 K. Mitsunari, Y. Takahashi & Y. Otani Load testing and analysis of bridges missing critical documentation 397 J. Prader, J. Weidner, H. Hassanain, F.Moon, E. Aktan, F.Jalinoos, B.Buchanan & H. Ghasemi Integrating health monitoring in asset management 398 H. Furuta, H. Hattori, T. Ohama, K. Yoshida & D.M. Frangopol Integrative research supporting decision making for bridges 399 F.N. Catbas, D.M. Frangopol & A.E. Aktan Development of affordable GPS displacement monitoring system 400 M. Saeki, K. Oguni & M. Hori Methods for measuring structural deflection and applications to bridge deck performance monitoring 401 J.M.W. Brownjohn & X. Meng Experimental identification of multiple oscillation frequencies using GPS 402 P.A.Psimoulis, S. Pytharouli & S. Stiros Real-time dynamic monitoring with GPS and georobot during Sutong Bridge construction 403 S.X. Huang & B.C. Yang The safety assessment method of existing large span steel structural members 404 X. Liu & Y. Luo A mechanical model of steel frames with joint damages 405 Y. Luo & H. Song Statistic analysis of a prototype structural health monitoring system for the Nanpu Bridge in Shanghai, P. R. China 406 R. Wang, X. Meng, Y. Luo, L. Yao & W. Huang The analysis of GPS single epoch positioning algorithm based on the deformation monitoring 407 L. Yao, P.Yao & X. Meng

Deformation analysis of the supporting towers of the Nanpu Bridge from GPS measurements 408 L. Yao, Y. Xie, Y. She & X. Meng

Deformation monitoring and analysis of high pylon of Su-Tong Bridge in construction 409 D.Yue,C.Wang&H.Li The statistical investigation on one year GPS monitoring data from Donghai Bridge Health Monitoring System (DHBHMS) 410 D. Dan, L. Sun, X. Meng & D. Xie Nonlinear dynamic responses of large span hybrid structures under multi-dimensional seismic excitation 411 Y. Huang & Y. Luo

XIX Research into the use of GNSS to monitor the deflections of suspension bridges, and the role of the FIG in deformation monitoring of bridges 412 G.W. Roberts, X. Meng & C.J. Brown Deflection monitoring of bridges: A case study of the Forth Road Bridge 413 X. Meng, G.W. Roberts & C.J. Brown Recent progress in GNSS-based long bridge deformation monitoring 414 X. Meng, G.W. Roberts, A.H. Dodson, L. Xu & Z.Wan Investigation on the severe corroded steel girder bridge, Hakkeibashi-Bridge 415 H. Furuta, M. Kawatani, T.Yamaguchi, I.H. Kim & M. Soma System of partial safety factors in reliability-based bridge assessment 417 I. Paik, D. Kim & S. Shin Bridge management system for national highway network in Korea 418 H.Y. Kim Autonomous bridge inspection and monitoring based on the robotic systems 419 J.S. Lee, S. Kim, I. Hwang & J.F.Choo Development of inspection robot to PSC box bridge using digital image processing 420 J. Kim, B. Lee, D. Park, J. Shin & C. Park Quantification models of bridge condition and performance 421 K.-J. Lee, S.-H. Park, J.-S. Kong, K.-H. Park & C.-H. Park Structural behaviors of Seohae cable-stayed bridge affected by temperature 422 S.G. Kang, J.B. Kwon, I.K. Lee & G.H. Lee Influences of diffusion coefficient and verification of validity on prediction of chloride induced deterioration of concrete bridges 423 H. Tsuruta, H. Furuta, I. Iwaki, A. Kamiharako, M. Soma & M. Suzuki

Life cycle costing Life time assessment of steel bridges via monitoring and testing 427 U. Peil, M. Frenz & I. Schendel Life cycle cost evaluation of neutralized reinforced concrete bridges subjected to earthquake 428 Y.C. Sung, C.K. Su, C.C. Hsu, M.C. Lai, K.Y. Liu & K.C. Chang Degradation, repair methods and real service life of soil steel composite bridges in Sweden 429 H.-Å. Mattsson & H. Sundquist Study on function extension of an existing PC rigid frame bridge during its life cycle 430 X.-X. Li, X.-F.Shi, X. Ruan & T.-Y.Ying Residual life assessment of steel girder bridges 431 R.K. Gupta Application of a new metal spraying system for steel bridges Part4. Reference product service life prediction for the system 432 T. Kondo, S. Okuno, A. Yamazaki & H. Matsuno Resource allocation for seismic retrofit of highway network 433 U.J. Na, M. Shinozuka, P.Franchetti, E. Da Lozzo & C. Modena XX Probabilistic cost model for bridge integrated project delivery and management 434 M.G. Huang, B.G. Kim, S.H. Lee & Y.H. Park Reliability analysis for bridge piles 435 A.S. Nowak, M. Kozikowski, T. Lutomirski & J. Larsen Influence of chloride ion diffusion coefficient on the service life of concrete structures subjected to coastal environment 436 J.I. Park, S.H. Bae, K.G. Yu, K.M. Lee, H.Y. Shin & D.O. Kang Damage to structures due to increasing traffic numbers related to service life predictions 437 A. de Boer & B.(M.H.) Djorai Optimal design of cable-stayed bridges based on minimum life-cycle cost 439 S.-H. Han & A. H-S. Ang Design planning decision for deteriorating wearing surfaces based on whole-life design considering life-cycle cost 440 J.X. Peng, X.D. Shao & M.G. Stewart Optimal seismic design of cable-stayed bridges based on LCC concept 442 D. Hahm, H.-M. Koh, W. Park, K.-S. Park & S.-Y. Ok

Life-cycle structural engineering Structural geometry effects on the life-cycle performance of concrete bridge structures in aggressive environments 445 F.Biondini, D.M. Frangopol & P.G. Malerba FRP reinforced concrete: Reliability assessment for life-cycle analysis 446 S.M.C. Diniz Seismic performance upgrading of existing bridge structures 447 G. Furlanetto, L. Ferretti Torricelli & A. Marchiondelli Estimation of fatigue life for suspension bridge hangers under wind action and train transit 448 F.Petrini, F.Giuliano & F.Bontempi Life-cycle bridge management considering member interference 449 K.-H. Park, S.-Y. Lee & J.-S. Kong Bridge maintenance strategy based on life-cycle cost and rebuild cost stabilization 450 A. Miyamoto & J. Ishida Damage modeling and life-cycle reliability analysis of aging bridges 452 F.Biondini, D.M. Frangopol & E. Garavaglia

Lifetime-perspective design of Kwangyang suspension bridge with main span 1545m The planning and design of the long-span suspension bridge connecting Myodo and Gwangyang in Korea 455 J.-H. Kim, M.-J. Lee, S.-H. Shin & S.-B. Chun Wind resistance design of Kwangyang Bridge 456 S.-D. Kwon, S.-H. Lee, H. Uejima & M.-J. Lee XXI Planning, design and construction of the largest concrete pylon in the world 457 S.-H. Lee, W.-S. Jang, S.-B. Oh & K.-T. Kim The design for anchorage of sea-crossing long-span suspension bridge 459 H.-S. Jang, Y.-I. Jang, Y.-S. Choi, K. Park & K.-T. Kim IDC for economical and safe design for the suspension bridge connecting Myodo and Gwangyang 460 C.-S. Kim, W.-J. Kim, K.-S. Cho, J.-H. Kim & Y.Yamasaki The innovative construction method for the long-span suspension bridge connecting Myodo and Gwangyang in Korea 461 S.-H. Shin, P.-J. Yu, S.-W. Jeong & Y. Takizawa The planning of ship collision protections based on risk analysis 462 H.-C. Kwon, M.-J. Lee, J.-H. Park & H. Andersen Measurement and evaluation of data from wind observation station in Gwangyang 463 S.-L. Lee, G.-M. Han, Y.-S. Gwon & Y.-G. Bae Vehicle-structure dynamic interaction by displacement constraint equations and stabilized penalty method 464 K.-Y. Chung, J.-M. Kim, M.-K. Song & J.-G. Paik

Loads and capacity assessment Evaluation of ultimate capacity of deteriorated reinforced concrete bridge columns 467 M. Tapan & R.S. Aboutaha Load test of bonded post-tensioned concrete beams with corroded tendon 468 S.G. Youn, S.H. Park, C. Lee & E.K. Kim The bridges and the floods 470 V.Popa 3 DOFs collision model for the analysis of bridge super-structures and deck house collision 471 G.H. Lee & S.L. Lee Modelling granular soil to predict pressures on integral bridge abutments 472 J. Banks, T. Knight, J. Young & A. Bloodworth Assessment of bridge capacity through proof load testing 473 J.D. Gómez & J.R. Casas Micro-simulation modelling of traffic loading on medium- and long-span road bridges 474 E.J. OBrien, A. Hayrapetova & C. Walsh Nonlinear analysis of PSC structures with internal tendon by strengthened using external tendon 475 J.G. Park, J.-H. Cheon, M.-Y. Kim, H.M. Shin, B.-J. Lee & J.-H. Choi Enhancement of bridge serviceability due to a strong wind 476 A. Krecak, P.Sesar & M. Masala-Buhin Comparison of theoretical and measured temperature distributions for concrete slab bridges 477 E.-S. Hwang & J.J. Lee XXII Numerical analysis of old masonry bridges supported by field tests 478 J. Bie´n,T. Kami´nski& Ch. Trela Input ground motion for seismic design considering near fault effects in stable continental regions 479 J.H. Kim & J.K. Kim Calculation of dynamic interaction of train and an arch bridge 480 J. Györgyi & G. Szabó Structural behavior of corroded reinforced concrete structures 481 K.Z. Hanjari, K. Lundgren, P.Kettil & M. Plos Field evaluation of dead and live load hanger rod stresses in a continuous steel girder bridge 482 S. Pessiki & I. Hodgson Dynamic behaviour of soil-steel road bridge made from corrugated plates 483 D. Beben & Z. Manko Development of live load model using Weigh-In-Motion data 485 E.-S. Hwang, I.R. Paik & J.-J. Lee Bridge safety analysis considering heavy truck loading 487 J. Du & D.-J. Han Behaviors of bracing members in U-type trapezoidal steel box girders 488 K. Kim & J.H. Park

Monitoring and assessment of bridges using novel techniques Wireless sensor networks for model based bridge monitoring 491 S. Deix, M. Ralbovsky & R. Stütz Structural health monitoring and passive vibration control of an Austrian road bridge 492 M. Reiterer & L. Praxmarer Recent Austrian activities in bridge monitoring 493 R. Geier Computational model updating for bridge maintenance planning 494 S. Deix, M. Ralbovsky & H. Friedl AIFIT – user orientated identification for infrastructure, theory 495 R. Wendner, S. Hoffmann, A. Strauss & K. Bergmeister AIFIT – user orientated identification for infrastructure, application 496 S. Hoffmann, R. Wendner, K. Bergmeister, M. Mautner & W. Steinhauser Stochastic nonlinear finite element analysis of bridges 498 R. Pukl, M. Voˇrechovský & D. Novák Overview of 40 bridge monitoring projects using fiber optic sensors 499 D. Inaudi & B. Gliši´c Reliability assessment of an existing bridge using long-term monitoring 500 A. Strauss, D.M. Frangopol & S. Kim XXIII Evaluation of the redundancy of bridge superstructures and substructures 501 M. Ghosn & D.M. Frangopol Damage detection by pattern recognition at bridge components 502 H. Wenzel & R. Veit-Egerer Dynamic damage identification of Colle Isarco viaduct 504 D. Lehký, D. Novák, P.Frantík, A. Strauss & K. Bergmeister Virtual testing of bridges for life cycle reliability assessment 505 R. Pukl, V. Cervenka,ˇ B. Teplý, D. Novák & K. Bergmeister Degradation modelling of bridge components based on cellular automata 507 D. Novák, B. Teplý, J. Podroužek, M. Chromá & A. Strauss Stochastic aging model for infrastructure buildings 508 M. Petschacher

New developments in large-scale model studies of bridge components and systems subjected to earthquakes Hybrid distributed simulation of a bridge-foundation-soil interacting system 511 A.S. Elnashai, B.F.Spencer, S.J. Kim, C.J. Holub & O.S. Kwon Research and application of precast segmental bridge columns for seismic regions 512 K.-C. Chang, M.-S. Tsai, Y.-C. Ou, G.C. Lee, J.-C. Wang & P.-H.Wang Seismic performance of a two-span bridge subjected to fault-rupture 513 H. Choi, M.S. Saiidi, P.Somerville & S. El-Azazy Nonlinear modeling of a two-span reinforced concrete bridge model from pre-yield through failure utilizing contemporary analytical methods 514 N. Johnson, M. Saiidi & D. Sanders Development of an innovative seismic damper for large-scale bridges and sub-structured hybrid earthquake loading tests 515 H. Iemura, A. Igarashi & A. Toyooka

Practical application of BMS and BMS-DB Decision making processes and deterioration models of bridge management systems in Korea 519 B.-G. Kim, J.-N. Park, S.-H. Lee & M.-S. Park J-BMS database system 2007 for management of existing bridges in Yamaguchi prefecture 520 K. Kawamura, A. Miyamoto & J. Ishida Practical application of J-BMS to existing bridges in Yamaguchi Prefecture 521 A. Miyamoto, K. Kawamura & J. Ishida Rational approach for the management of a medium size bridge stock 522 E. Brühwiler Proposal for BMS deterioration curves based on the analysis of Hanshin Expressway inspection data 523 H. Nakajima, T.Yamagami, T. Kagayama & M. Hayashida XXIV Methodology for determination of financial needs of gradually deteriorating bridges 524 B.T. Adey & R. Hajdin The measure towards advanced of Bridge Management System for Expressway Bridges in Japan 525 Y. Wada, S. Sakai, T. Ohshiro, A. Homma & N. Ogata Application of China bridge management system in Qinyuan city 526 B.F.Yan & X.D. Shao

Practical applications of SHM techniques for railway systems Structural monitoring of a maglev guideway with wavelength division multiplexed FBG sensors 529 W. Chung, D. Kang, I. Yeo & J.S. Lee Evaluation of modal parameters of a full scaled prestressed concrete beams for railway bridges 530 S.I. Kim & N.S. Kim Practical acceleration reducing method in high-speed railway bridges 531 W.J. Chin, J.W. Kwark, J.R. Cho, E.S. Choi & B.S. Kim Real-time damage detection of railroad bridges using acceleration-based ANN algorithms 532 J.T. Kim, J.H. Park, D.S. Hong & J.H. Yi Active piezoelectric sensor nodes and sensor self-diagnosis for structural health monitoring 533 S. Park, C.B. Yun, G. Park & D.J. Inman Modal parameter extraction of high-speed railway bridge using TDD technique 534 B.H. Kim, J.-W. Lee & T.-Y.Yoon

Reliability and risk management Challenges for structural maintenance in coastal and offshore zones – Floating structures 537 E. Watanabe Reliability analysis of Steel – Concrete Hybrid Cable-Stayed Bridge during construction 539 J.H. Yun, C. Moon, J.W. Sun & H.N. Cho Reliability analysis of a high-speed railway bridge system based on an improved response surface method 540 A.S. Nowak, T. Cho, D.H. Lee & M.-K. Song ANN-based reliability analysis of a fiber reinforced polymer deck 542 J. Cui, D. Kim & D.H. Kim Strength of the chain for suspended scaffolds 543 Y. Hino Multimode analysis of extraneously induced excitation due to turbulence on cable-stayed bridges, including temporary stabilizing measures 544 J.-Y. Cho, Y.-R. Cho & H.-E. Lee XXV Reliability assessment of seismic expansion joints in bridges 545 J.E. Padgett & R. DesRoches System-level reliability evaluation of bridge structures and networks by matrix-based system reliability method 546 J. Song & W.-H. Kang Rehabilitation and monitoring of a marine bridge in Ireland 547 A. Farrell, L. Duffy, A. O’Connor & J. Kelly Repair of damaged footbridge after strike of excavator 549 A.G. Mordak & Z. Manko Effect of corrosion on the reliability of a bridge based on Response Surface Method 550 S.I. Jo, T. Onoufriou & A.D. Crocombe Vibration control of cable-stayed bridge and derrick crane system during construction 551 H.-J. Pae, D.-S. Kim, W. Park, K.-S. Park & H.-M. Koh

Repair and strengthening Ductility of CFRP strengthened concrete bridge girders 555 S. Kim & R.S. Aboutaha Modal analysis and step-by-step repair operation of a two span concrete skew bridge to replacement of its elastomeric bearings 556 R. Akbari, S. Maalek & H. Ashayeri Bridge widening – technical, economical and aesthetical aspects 557 G. Boro´nczyk-Plaska& W. Radomski Behavior under compressive loads of steel structural members repaired by heating and pressing 558 M. Hirohata & Y.-C. Kim Effectiveness of prestressed Carbon Fibre Reinforced Polymer (CFRP) sheets for rehabilitation of prestressed concrete girders 559 Y.J. Kim, M.F.Green, C. Shi, J. Ford, L. Bizindavyi & R.G. Wight The use of polymer concrete materials for construction, maintenance, rehabilitation and preservation of concrete and steel orthotropic bridge decks 560 A.M. Dinitz & S. Park The correlation between crack and residual stress generated by repair welding in service 561 Y.C. Kim, S.H. Lee & Y. Agano Repair of a concrete bridge by composites CRFP 562 M. Abdessemed, S. Kenai, A. Kibboua, J.-L. Chatelain, B. Guillier & A. Bali Innovative rehabilitation of a damaged prestressed concrete girder bridge using prestressed CFRP sheets: Design and specification 563 Y.J. Kim, M.F.Green, G.J. Fallis, R. Eden & R.G. Wight Rehabilitation of bridges with concrete overlays 564 C.A.M. de Smet & J. Kunz Bond and flexural behaviour of RC members strengthened with CFRP composites 566 D.S. Yang, J.M. Park, S.N. Hong & S.K. Park XXVI Seismic performance improvement of bridges by earthquake protection systems in Korea 567 D.-H. Ha, H.-M. Koh, S.Y. Lee, H.J. Kim & I.J. Kwahk Application of new prestressing method using carbon fiber plates 568 T. Ohshiro, Y. Wada, A. Takeuchi, K. Morikita, H. Yasumori & T. Takahashi Experimental study of bolted joint with ultra thick plate and M30 bolt 569 J. Kim, J. Byuoun, J.B. Jo & K. Jung Monitoring for fatigue crack propagation of steel plate repaired by CFRP strips 570 H. Nakamura, K. Maeda, H. Suzuki & T. Irube Reevaluation of stresses and displacement of horizontally curved girders of a continuous span bridge 573 D.J. Kim, C.P.Fan & B.T.Yen Railroad bridge replacement in the US today: Current technology and future possibilities 575 F.Moreu, T. Nagayama, J. Zeman, G. Rus, S.Y. Lee & T. Park Assessment of repair cost and service life of repaired concrete structures after chloride attack 577 H.-W. Song, A. Petcherdchoo & H.-B. Shim Tests on cast iron carried out to repair bearings in Tumski Bridge in Wroclaw (Poland) 579 Z. Manko

Research & applications for bridge health monitoring Residual structural performance of corroded steel tubes submerged in seawater 583 K. Sugiura, E. Watanabe, K. Nagata & I. Tamura Modal flexibility and curvature for damage assessment: Laboratory demonstrations 585 M. Gul & F.N. Catbas Benchmark studies for Structural Health Monitoring using computer vision 587 R. Zaurin & F.N. Catbas Field monitoring of continuous steel-concrete composite girder during internal force adjustment of the Siyuan Bridge 588 W. Lu & D.M. Frangopol Application of GPS monitoring technology to the construction of the pylon 589 J.S. Lee & J.G. Yoon Bridge fatigue reliability assessment and prediction 590 K. Kwon & D.M. Frangopol Eigenfrequency estimation for bridges using the response of a passing vehicle with excitation system 591 Y. Oshima, Y. Kobayashi, T.Yamaguchi & K. Sugiura Monitoring and inspection of a 30 years old prestressed concrete bridge 592 M. Pimentel, J. Santos, J.R. Casas & J. Figueiras Boundary condition parameter estimation for structural identification 593 Y. Dere & F.N. Catbas XXVII Long-term monitoring of stochastic characteristics of a full-scale suspension bridge 594 H.-B. Yun, S.F.Masri, R.D. Nayeri, F.Tasbihgoo, E. Kallinikidou, M. Wahbeh, R.W.Wolfe & L.-H. Sheng The need, challenges, and opportunities for research and application of Bridge Health Monitoring, a Turkish Experience 595 A. Turer Crack detection in steel bridges 596 F.Jalinoos & A. Rezai Systems-based monitoring approaches for improved infrastructure management under uncertainty: Novel approach 597 F.N. Catbas & D.M. Frangopol

Seismic and dynamic analysis Prioritization and seismic risk assessment of bridges 601 D. Cardone, G. Perrone, M. Dolce & L. Pardi Ambient vibration test and seismic evaluation of steel-deck truss bridge 603 S. Jung, S.-T. Oh, S. Kim, Y.-W. Shim & S.S. Chen Finite element model updating of a concrete arch bridge through static and dynamic measurements 605 H. Schlune, M. Plos, K. Gylltoft, F.Jonsson & D. Johnson Seismic response of highway viaducts under design live load considering vehicle as a dynamic system 606 M. Kawatani, C.-W. Kim, S. Konaka & R. Kitaura A substructures approach in the dynamic analysis of continuous beams under moving oscillators 608 V.De Salvo, G. Muscolino & A. Palmeri Improvement of seismic analysis concerning the characteristic difference of HDR-S between the stages of design and inspection 610 H. Yosuhisa, M.-S. Yoo, D.-H. Ha & K.-Y. Kim Vulnerability assessment of an existing highway bridge by 3-D nonlinear time history analyses and proposing its retrofit design 611 M. Hosseini & S.R. Khavari Shake table studies of scaled reinforced concrete bridge piers subjected near-fault ground motions 612 Y.-S. Chung, C.-Y. Park, H.-K. Hong, D.-H. Lee & C.-S. Shim Computer wind investigations for long bridge crossings 613 D. Janjic & A. Domaingo Identification of the dynamic characteristics of long span bridges using ambient vibration measurements 614 A.L. Hong & R. Betti Application of indexing and detailed seismic risk assessment approaches to existing bridges 615 D. Cardone, G. Perrone & L. Pardi XXVIII Effectiveness of rupture controllable steel side blocks for elevated girder bridges with isolation bearings 617 N. Asada, M. Matsumura, T. Kitada, M. Sakaida & M. Yoshida Influence of bullet train as dynamic system on seismic performance of Shinkansen viaducts 620 M. Kawatani, X. He, K. Shinagawa & S. Nishiyama Vibration-based tension identification of ultra long stay cables 623 J. Liu, N. Fang & Q. Zhang Calculation of the influence line of a bridge using a moving vehicle 624 A. González & E.J. OBrien Seismic assessment and evaluation of 520 highway bridges in Western Kentucky 625 C.C. Choo, I.E. Harik, W. Zatar & H.S. Ding Cross-sectional stress distribution of short suspenders in arch bridges 626 Y.B. Li & Q.W. Zhang Experimental investigation on the Bi-lateral seismic behavior of a two-span bridge model isolated by rolling-type bearings 627 K.-C. Chang, M.-H. Tsai & Z.-Y. Lin Modal analysis of corrugated steel flexible shell bridge structure before backfilling 628 D. Beben & Z. Manko Experimental study on the shear characteristics of seismic isolation bearings 629 I.J. Kwahk, C.B. Cho & Y.J. Kim

Seismic design and performance issues for highway bridges Post-earthquake evaluation of reinforced concrete bridge columns 633 A. Vosooghi, M. Saiidi & S. El-Azazy Identification of effective seismic retrofits for common bridge classes on the basis of failure probability 634 J.E. Padgett & R. DesRoches Enhancement of axial ductility of circular concrete bridge columns 635 L.A. Marvel, J.C. West & R.A. Hindi Retrofitting structures with a combination of seismic isolation and attenuation 636 A. Caner, M.J. Abrahams, E. Dogan & C. Ozkaya Supplemental device to improve the performance of seismic-isolated bridges in near-fault zones 638 M. Dicleli

Smart sensing and monitoring technologies for bridge maintenance, safety and management Baseline knowledge discovery from one-year structural monitoring measurements of Donghai Bridge 641 Z. Sun, Z.H. Min & Z.F.Zhou XXIX Test-bed implementation of piezopaint-based acoustic emission sensor for crack initiation monitoring 643 Y. Zhang & X. Li Development of experimental benchmark problems for international collaboration in structural response control 644 C.-H. Loh, A.K. Agrawal, J.P.Lynch & J.N. Yang Advance sensor technologies on Korean Bridges: Field benchmark opportunities 645 J.P.Lynch, J.H. Kim, Y. Zhang, M. Wang, H. Sohn & C.B. Yun Remotely controllable structural health monitoring systems for bridges using 3.5 generation mobile telecommunication technology 646 K.Y. Koo, J.Y. Hong, H.J. Park & C.B. Yun Structural damage assessment using optical Fiber Bragg Grating vibration sensing system 647 R.J. Sun, Z. Sun & L.M. Sun Self-sensing and power harvesting carbon nanotube-composites based on piezoelectric polymers 648 K.J. Loh, J. Kim & J.P.Lynch A nonlinear impedance method and its potential application in baseline free crack detection in metallic structures 649 D. Dutta & H. Sohn Piezoelectric sensor system for structural health monitoring 650 B. Kim & Y. Roh Fatigue reliability updating through inspections and monitoring data of steel bridges 651 C. Wang, X. Yu, Y. Feng & X. Liu Advanced signal processing for ultrasonic structural monitoring of waveguides 652 M. Cammarata, P.Rizzo, D. Dutta, H. Sohn & K.A. Harries Development of an optical fiber corrosion sensors based on light reflection 653 H. Huang & N. Gupta Stochastic Subspace Identification (SSI) model analysis using wireless data logger system in grand bridge under wind load effect 655 Y.-S. Kim, S.-Y. Park, C.-B. Yun & J.-S. Choi Sequential health monitoring in steel plate-girder bridges by using combined vibration-impedance signatures 656 D.S. Hong, J.H. Park, J.T. Kim & W.B. Na Smart bearings for structural behavior monitoring 657 F.M.Wegian, G. Fu, J. Feng, Y. Zhuang & P.-J. Chun Smart structural elements for the condition monitoring of bridge structures 658 D. Zonta, M. Pozzi, H.Y. Wu & D. Inaudi Sensing capability of electromagnetic induction system for vibration control of structures 659 H.J. Jung, D.D. Jang, H.J. Lee, S.W. Cho & J.H. Koo

Special session on Incheon bridge Incheon Bridge project outline 663 J.H. Yang Son, M.G. Yun, H.S. Kim, H.Y. Shin & I.S. Shim XXX Design and construction of approach bridge in Incheon Bridge project 665 J.-Y. Song, K.-Y. Choi, H.-Y. Shin, W.-S. Lee, B.-C. Cho & D.-W. Hwang Development of geometry control system for cable-stayed bridges and application to the Incheon Bridge 667 K. Jung & H.S. Lee Seismic design and performance assessment of pile-bents in Incheon Bridge viaduct 668 H.-S. Son, M.-S. Oh, K.-L. Park & J.-H. Yang Geometry control for the concrete pylon of Incheon Cable Stayed Bridge 669 D.K. Im, J.G. Yoo, C.H. Kim & H.S. Kim Case study of Osterberg-Cell pile load test on large diameter drilled shaft in Incheon Bridge project 670 S.-H. Shin, Y.-K. Lee, Z.-C. Kim, J.-H. Kim & H.-G. Lee Design of ship impact protection in Incheon Bridge 671 J.H. Kim, H.Y. Shin, H.T. Kim & S.H. Lee

Structural health monitoring on cable supported bridges Development of bridge WIM systems without axle detector using artificial neural network 675 M.-S. Park, J. Lee, B.-W. Jo & S. Kim Development of local live load truck model for long span bridges based on BWIM data of Seohae cable-stayed bridge 676 M.-S. Park, C.-H. Park & J. Lee Long-term structural behaviors of Seohae cable-stayed bridge based on results from SHM and surveys 678 J.C. Park, C.M. Park, M.S. Park, I.K. Lee & B.W. Jo Modal parameter extraction of Seohae cable-stayed bridge using TDD technique 679 B.H. Kim, J.C. Park, M.S. Park & I.K. Lee Analysis model updating of the Seohae cable-stayed bridge 680 H.K. Kim, S.D. Park, K.T. Kim, W. Park, S.H. Lee, J.F.Choo, J.C. Park & M.S. Park Development of prediction method of non-linear observed data from cable-stayed bridge using support vector regression 681 M.-Y. Park, H.-N. Cho, K.-W. Park, J.-C. Park, M.-S. Park & I.-K. Lee SHMS and wind engineering on the Busan-Geoje Fixed Link bridges 684 Y.M. Kim, D.Y. Kim, C.H. Kim, A. Galmarini, P.D. Frederiksen & J.E. Andersen

Structural robustness Robustness investigation of long suspension bridges 687 F.Bontempi & L. Giuliani Dynamic analysis for structural robustness evaluation 688 L. Giuliani & F.Bontempi Measure of structural robustness under damage propagation 689 F.Biondini & S. Restelli XXXI Robustness assessment of a cable-stayed bridge 690 M. Wolff & U. Starossek Approaches to measures of structural robustness 691 U. Starossek & M. Haberland Collapse resistance and robustness of bridges 692 U. Starossek Evaluation of the dynamic amplification factor for cable breakage in cable-stayed bridges 693 Y. Park, H.-M. Koh, J.F.Choo, H. Kim & J. Lee

Sustainable bridges Test of a concrete bridge in Sweden. – I. Assessment methods 697 A. Puurula, O. Enochsson, H. Thun, B. Täljsten, L. Elfgren, J. Olofsson & B. Paulsson Test of a concrete bridge in Sweden. – II. CFRP strengthening and structural health monitoring 698 B. Täljsten, M. Bergström, H. Nordin, O. Enochsson & L. Elfgren Test of a concrete bridge in Sweden. – III. Ultimate Capacity 699 O. Enochsson, A. Puurula, H. Thun, L. Elfgren, B. Täljsten, J. Olofsson & B. Paulsson Test to failure of a railway reinforced concrete through bridge in Örnsköldsvik, Sweden. – IV.Evaluation of damage detection methods 700 P.J.S. Cruz & R. Salgado Assessment and monitoring of an old railway steel truss bridge in northern Sweden 701 O. Enochsson, L. Elfgren, A. Kronborg & B. Paulsson Railway bridge loaded to failure test in Örnsköldsvik, Sweden – Strain measurement using Fiber Bragg Grating system incorporated in Carbon Fibre Reinforced Polymer 702 A. Kerrouche, J. Leighton & W.J.O. Boyle Single and multiple crack monitoring in concrete bridges 704 P.J.S. Cruz, A. Diaz de León & C.K.Y. Leung Structural assessment of concrete railway bridges 705 M. Plos, K. Gylltoft, K. Lundgren, L. Elfgren, J. Cervenka,ˇ A. Herwig, E. Brühwiler, S. Thelandersson & E. Rosell Safety requirements in the capacity assessment of existing bridges 707 J.R. Casas & D.F.Wisniewski Guideline for load and resistance assessment of existing European railway bridges 708 J.S. Jensen, M. Plos, J.R. Casas, C. Cremona, R. Karoumi, & C. Melbourne Improved assessment methods for static and fatigue resistance of old metal railway bridges 710 C. Cremona, A. Patron, S. Hoehler, B. Eichler, B. Johansson & T. Larsson Consideration of dynamic traffic action effects on existing bridges at ultimate limit state 712 E. Brühwiler & A. Herwig XXXII High cycle fatigue strength of brick masonry. A probabilistic approach 714 J.R. Casas Probabilistic models for resistance of European concrete railway bridges 715 J.R. Casas & D.F.Wisniewski Probabilistic models of material properties for design and assessment of concrete bridges 716 D.F.Wisniewski, P.J.S. Cruz, A.A.R. Henriques & R.A.D. Simões Safety assessment of railway bridges by non-linear and probabilistic methods 718 J. Cervenka,ˇ V. Cervenka,ˇ Z. Janda & R. Pukl Complex multi-tool inspection of a masonry arch bridge using non-destructive testing 719 R. Helmerich, E. Niederleithinger, C. Trela, J. Bien & G. Bernardini Evaluation of corrosion situation on reinforced concrete by portable electrochemical technique 721 R. Bäßler, A. Burkert & T. Frølund Assessment of a railway concrete arch bridge by numerical modelling and measurements 722 G. He, Z. Zou, O. Enochsson, A. Bennitz, L. Elfgren, A. Kronborg, B. Töyrä & B. Paulsson Field test – strengthening and monitoring of the Frövi Bridge 723 A. Kerrouche, W.J.O. Boyle, Y. Gebremichael, L. Alwis, K.T.V.Grattan, B. Täljsten & A. Bernnitz

Author index 725

XXXIII

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Preface

The number of deteriorating bridges is increasing worldwide. Costs of maintenance, repair and rehabilitation of these bridges far exceed available budgets. Maintaining the safety of existing bridges by making better use of available resources is a major concern for bridge management. Internationally, the bridge engineering profession has taken positive steps to develop more com- prehensive bridge management. It was therefore considered appropriate to bring together all of the very best work that has been done in the field of bridge maintenance, safety, management, life- cycle performance, health monitoring, informatics and cost at the Fourth International Conference on Bridge Maintenance, Safety and Management (IABMAS’08), held in Seoul, Korea from July 13 to 17, 2008. The First (IABMAS’02), Second (IABMAS’04) and Third (IABMAS’06) Inter- national Conferences on Bridge Maintenance, Safety and Management were held in Barcelona, Spain, July 14–17, 2002, Kyoto, Japan, October 18–22, 2004, and Porto, Portugal, July 16–19, 2006, respectively. The International Association for Bridge Maintenance and Safety (IABMAS), which serves as the organizing association of IABMAS’08, encompasses all aspects of bridge maintenance, safety and management. Specifically, it deals with : health monitoring and inspection of bridges; bridge repair and rehabilitation issues; bridge management systems; needs of bridge owners, financial planning, whole life costing and investment for the future; bridge related safety and risk issues and economic and other implications. The objective of IABMAS is to promote international cooperation in the fields of bridge maintenance, safety, management, life-cycle performance and cost for the purpose of enhancing the welfare of society. The interest of the international bridge community in all these fields has been confirmed by the high response to the call for papers. In fact, more than 660 abstracts were received at the Conference Secretariat. About 70% of them were selected for final publication as full-papers and presentation at the Conference within four plenary sessions and 72 technical sessions. Compared to IABMAS’06 the total of number of papers scheduled for presentation has increased from 421 to 465. IABMAS’08 covered all major aspects of bridge maintenance, safety, management, health moni- toring and informatics including advanced materials, ageing of bridges, assessment and evaluation, bridge codes, bridge diagnostics, bridge management systems, composites, design for durability, deterioration modeling, emerging technologies, fatigue, field testing, financial planning, health monitoring, high performance materials, innovations, inspection, life-cycle performance, load capacity assessment, loads, maintenance strategies, new technical and material concepts, non- destructive testing, optimization strategies, prediction of future traffic demands, rehabilitation, reliability and risk management, repair, replacement, residual service life, safety and serviceabil- ity, service life prediction, strengthening, sustainable materials for bridges, sustainable bridges, informatics, and whole-life costing, among others. Bridge Maintenance, Safety, Management, Health Monitoring and Informatics contains the lectures and papers presented at IABMAS’08. It consists of a book of abstracts and a CD-ROM containing the full texts of the lectures and papers presented at IABMAS’08, including the T.Y. Lin Lecture, eight Keynote Lectures and 456 technical papers from 32 countries. This set provides both and up-to-date overview of the field of bridge engineering and significant contributions to the process of making more rational decisions in bridge maintenance, safety, management, life-cycle performance, and cost for the purpose of enhancing the welfare of society. On behalf of IABMAS, the chairs of the Conference would like to take this opportunity to express their sincere thanks to the authors, organizers of special sessions and mini-symposia, and participants for their contributions, to the members of the Conference Scientific Committee for

XXXV their dedicated work, and to the members of the Local Advisory and Organizing Committees for the time and effort they have devoted to making IABMAS’08 a successful event. Finally, we would like to register our sincere thanks to all the sponsors of IABMAS’08.

Hyun-Moo Koh and Dan M. Frangopol Chairs, IABMAS’08 Seoul and Bethlehem, April 2008

XXXVI Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Conference organization

ORGANIZING ASSOCIATION

The International Association for Bridge Maintenance and Safety (IABMAS) (http://www.iabmas.org/)

CONFERENCE CHAIRS

Hyun-Moo Koh, Seoul National University, Seoul, Korea Dan M. Frangopol, Lehigh University, Bethlehem, PA, USA

HONORARY CHAIR

Sung-Pil Chang, Seoul National University, Seoul, Korea

HOST ASSOCIATION

Korea Bridge Design & Engineering Research Center, Seoul National University, Korea Under the auspices of Ministry of Land, Transportation and Maritime Affairs, Korea Seoul Metropolitan Government, Korea Korea Institute of Construction & Transportation Technology Evaluation and Planning, Korea Korea Expressway Corporation, Korea Korea Infrastructure Safety & Technology Corporation, Korea Korea Institute of Construction Technology, Korea

CONFERENCE SCIENTIFIC COMMITTEE

Joan Casas (Co-Chair) Technical University of Catalonia, Barcelona, Spain Young Soo Chung (Co-Chair) Chung-Ang University, Seoul, Korea Andrzej Nowak (Co-Chair) University of Nebraska, Lincoln, NE, USA Emin Aktan Drexel University, Philadelphia, PA, USA Alfredo H-S. Ang University of California, Irvine, CA, USA Giuliano Augusti University of Rome La Sapienza, Rome, Italy György L. Balázs Budapest University of Technology and Economics, Budapest, Hungary Konrad Bergmeister University of Natural Resources and Applied Life Sciences, Vienna, Austria Jan Bien Wroclaw University of Technology, Wroclaw, Poland Fabio Biondini Politecnico di Milano, Milan, Italy John Bjerrum Danish Road Directorate, Copenhagen, Denmark Franco Bontempi University of Rome La Sapienza, Rome, Italy

XXXVII Fernando Branco Technical University of Lisbon, Lisbon, Portugal Eugen Bruehwiler EPFL, Lausanne, Switzerland Christian Bucher Vienna University of Technology, Vienna, Austria Necati Catbas University of Central Florida, Orlando, FL, USA Michael Chajes University of Delaware, Newark, DE, USA Kuo-Chun Chang National Taiwan University, Taipei, Taiwan, R.O.C. Sung-Pil Chang (Ex-Officio) Seoul National University, Seoul, Korea Steven Chase Federal Highway Administration, McLean, VA, USA Moe Cheung The Hong Kong University of Science & Technology, Hong Kong, P.R.C. Hyo-Nam Cho , Ansan, Korea Marios Chryssanthopoulos University of Surrey, Guilford, Surrey, UK Marcello Ciampoli University of Rome La Sapienza, Rome, Italy Joel Conte University of California, San Diego, CA, USA Ross Corotis University of Colorado, Boulder, CO, USA Christian Cremona LCPC, Paris, France Paulo Cruz University of Minho, Guimarães, Portugal Lennart Elfgren Luleå University of Technology, Luleå, Sweden Bruce Ellingwood Georgia Institute of Technology, Atlanta, GA, USA Ib Enevoldsen RAMBOLL, Copenhagen, Denmark Allen Estes California Polytechnic State University, San Luis Obispo, CA, USA Glauco Feltrin Swiss Federal Laboratories for Materials Testing and Research(EMPA), Dübendorf, Switzerland João Almeida Fernandes National Civil Engineering Laboratory, Lisbon, Portugal Joaquim Figueiras University of Porto, Porto, Portugal John Fisher Lehigh University, Bethlehem, PA, USA Michael Forde University of Edinburgh, Edinburgh, UK Dan M. Frangopol (Ex-Officio) Lehigh University, Bethlehem, PA, USA Ian M. Friedland Federal Highway Administration, Washington, DC, USA Yozo Fujino University of Tokyo, Tokyo, Japan Hitoshi Furuta Kansai University, Takatsuki, Japan Yaojun Ge Tongji University, Shanghai, China Hamid Ghasemi Federal Highway Administration, USA Michel Ghosn City University of New York, NY, USA Paul Grundy Monash University, Victoria, Australia Rade Hajdin Infrastructure Management Consultants GmbH, Zürich, Switzerland Geir Horrigmoe NORUT Technology, Narvik, Norway Michael C. H. Hui Government of Hong Kong Special Administrative, Hong Kong, P.R.C. Naeem Hussain ARUP,Hong Kong, P.R.C. Daniele Inaudi SMARTEC SA, Manno, Switzerland Jens Sandager Jensen COWI A/S, Lyngby, Denmark Ahsan Kareem University of Notre Dame, Notre Dame, IN, USA Maria Kaszynska Szczecin Technical University, Szczecin, Poland Malcolm Kerley Virginia Department of Transportation, Richmond, VA, USA Sungkon Kim Seoul National University of Technology, Seoul, Korea Sang-Hyo Kim , Seoul, Korea Risto Kiviluoma WSP Finland Ltd., Helsinki, Finland Wayne Klaiber Iowa State University, Ames, IA, USA Leo Klatter Public Works and Water Management, Utrecht, The Netherlands C.G. Koh National University of Singapore, Singapore

XXXVIII Hyun-Moo Koh (Ex-Officio) Seoul National University, Seoul, Korea Ulrike Kuhlmann University of Stuttgart, Stuttgart, Germany John Lane Railway Safety and Standards Board, London, UK Jørn Lauridsen Danish Road Directorate, Copenhagen, Denmark Hakeun Lee , Seoul, Korea Aiquin Li Southeast University, Nanjing, China Chin-Hsiung Loh National Taiwan University, Taipei, Taiwan, R.O.C. Giorgio Malerba Politecnico di Milano, Milan, Italy Ayaz Malik Rensselaer Polytechnic Institute, Troy, NY, USA René Maquoi University of Liège, Liège, Belgium Barney Martin Modjeski & Masters, Poughkeepsie, NY, USA Sami Masri University of Southern California, Los Angeles, CA, USA Ayaho Miyamoto Yamaguchi University, Ube, Japan Aftab Mufti ISIS Canada, Research Network, Winnipeg, Manitoba, Canada Hani Nassif The State University of New Jersey, New Jersey, USA Jan M. van Noortwijk HKV Consultants, Lelystad, The Netherlands Jinping Ou Dalian University of Technology, Dalian, China Tso-Chien Pan Nanyang Technological University, Nanyang, Singapore Livia Pardi Autostrade per l’Italia, Rome, Italy Udo Peil Technical University of Braunschweig, Braunschweig, Germany Stephen Pessiki Lehigh University, Bethlehem, PA, USA Victor Popa Search Corporation, Bucharest, Romania Mark Reno Quincy Engineering Inc., Sacramento, CA, USA James Ricles Lehigh University, Bethlehem, PA, USA Salvatore Russo University Iuav of Venice, Venice, Italy Mehdi Saiidi University of Nevada, Reno, NV,USA Richard Sause Lehigh University, Bethlehem, PA, USA Robert Sexsmith University of British Columbia, Vancouver, Canada Richard Shepard County of Eldorado, Department of Transportation, Placerville, CA, USA Dong-Ku Shin Myongji University, Gyeonggido, Korea Hyun-Mock Shin , Gyeonggido, Korea Soobong Shin Inha University, Incheon, Korea Arunprakash M. Shirole Arora and Associates, P.C.,Minneapolis, MN, USA Jongsung Sim Hanyang University, Ansan, Korea Marja-Kaarina Söderqvist Finnish Road Administration, Helsinki, Finland Mark Stewart University of Newcastle, Newcastle, Australia Luc Taerwe University of Ghent, Ghent, Belgium Man-Chung Tang T.Y. Lin International, San Francisco, CA, USA Palle Thoft-Christensen Aalborg University, Aalborg, Denmark Paul Thompson Castle Rock, CO, USA Tomoaki Utsunomiya Kyoto University, Kyoto, Japan Pedro Vellasco State University of Rio de Janeiro, Rio de Janeiro, Brazil Thomas Vogel ETH Zürich, Zürich, Switzerland Eiichi Watanabe Kyoto University, Kyoto, Japan Zhishen Wu Ibaraki University, Hitachi, Japan Kentaro Yamada Nagoya University, Nagoya, Japan Chung-Bang Yun Korea Advanced Institute of Science and Technology, Daejon, Korea Riccardo Zandonini University of Trento, Trento, Italy Yunfeng Zhang Lehigh University, Bethlehem, PA, USA

XXXIX CONFERENCE LOCAL ORGANIZING COMMITTEE

Young Suk Park (Chair) Myongji University, Gyeonggido, Korea Hae Sung Lee (Vice-Chair) Seoul National University, Seoul, Korea Jeeho Lee (Secretary) Dongguk University, Seoul, Korea Doobyong Bae Kookmin University, Seoul, Korea Jae-Yeol Cho Seoul National University, Seoul, Korea Dong Ho Choi Hanyang University, Seoul, Korea Dong-Ho Ha Konkuk University, Seoul, Korea Chul-Young Kim Myongji University, Gyeonggido, Korea Ho-Kyung Kim Mokpo University, Jeonnam, Korea Jee Sang Kim Seokyong University, Seoul, Korea Jeonghwan Kim Samsung C&T Corporation, Seoul, Korea Jeong-Tae Kim Pukyong National University, Busan, Korea Nam-Sik Kim Pusan University, Busan, Korea Youngjin Kim Korea Institute of Construction Technology, Gyeonggido, Korea Jung Sik Kong Korea University, Seoul, Korea Sang-Cheol Lee Korea Infrastructure Safety & Technology Corporation, Gyeonggido, Korea Jae Hoon Lee Yeungnam University, Gyeongsangbukdo, Korea Yun Mook Lim Yonsei University, Seoul, Korea Chan Min Park Korea Expressway Corporation, Gyeonggido, Korea Jaegyun Park , Gyeonggido, Korea Jang Ho Park , Suwon, Korea Kwan-Soon Park Dongguk University, Seoul, Korea Seon Kyu Park Sungkyunkwan University, Gyeonggido, Korea Young Ha Park Korea Expressway Corporation, Gyeonggido, Korea Chang Su Shim Chung-Ang University, Gyeonggido Korea Won-Sup Hwang Inha University, Incheon, Korea Seok-Goo Youn Seoul National University of Technology, Seoul, Korea

CONFERENCE LOCAL ADVISORY COMMITTEE

Chang Se Kim (Chair) Korea Institute of Construction & Transportation Technology Evaluation and Planning, Gyeonggido, Korea Won Joong Kim Korea Institute of Construction & Transportation Technology (Secretariat) Evaluation and Planning, Gyeonggido, Korea Woo Jong Kim (Secretariat) DM Engineering Co., Ltd., Seoul, Korea Yoon Chul Chun Samsung C&T Corporation, Seoul, Korea Chun Yang Jung Shinsung Engineering Co., Ltd., Seoul, Korea Kyoung Sup Jung Chungbuk National University, Chungbuk, Korea Byung Suk Kim Korea Institute of Construction Technology, Gyeonggido, Korea Dae-Young Kim Korea Railroad Technical Corporation, Seoul, Korea Il-Gon Kim Korea Infrastructure Safety & Technology Corporation, Gyeonggido, Korea Sun Won Kim BnS Engineering Co., Ltd., Seoul, Korea Sung-Hwan Kim Korea Expressway Corporation, Gyeonggido, Korea Woo Kim Chonnam National University, , Korea Young-Bong Kwon Yeungnam University, Gyeongsangbukdo, Korea

XL Chul Soo Lee Chungsuk Engineering Co., Ltd., Seoul, Korea Doo Hwa Lee Sambo Engineering Co., Ltd., Seoul, Korea Guem-Sook Lee Yongma Engineering Co., Ltd., Seoul, Korea Hae Kyung Lee Dasan Consultants Co., Ltd., Seoul Korea Kil Yong Lee Korea Consultant International Co., Ltd., Seoul, Korea Myeong-Jae Lee Yooshin Engineering Corporation, Seoul, Korea Sung Chul Lee Dongguk University, Seoul, Korea Sung Min Lee Saman Engineering Co., Ltd., Gyeonggido, Korea Sung Woo Lee Kookmin University, Seoul, Korea Kyung Kook Lim Ministry of Land, Transportation and Maritime Affairs, Gyeonggido, Korea Hyung-Ghee Park University of Inchoen, Incheon, Korea Dong Ho Yoo ENVICO Consultants Co., Ltd., Seoul, Korea DukHee Yoo Seoyeong Engineering Co., Ltd., Seoul, Korea Jae So Yoo Dohwa Consulting Engineering Co., Ltd., Seoul, Korea Tae YangYoon Research Institute of Industrial Science & Technology, Gyeonggido, Korea

IABMAS’08 SPECIAL SESSIONS

– Assessment and control of bridge vibrations, organized by Álvaro Cunha – Bridge 200 toward durable bridge, organized by Byung Suk Kim – Bridges for high-speed railways, organized by Rui Calçada – Damage assessment of existing bridges, organized by Soobong Shin and Jeong Tae Kim – Design expectations, monitoring response and maintenance decisions, organized by Giorgio Malerba – Development of the advanced robot systems for bridge inspection and monitoring, organized by Jong Seh Lee and Sungkon Kim – Fiber reinforced composites in bridges, organized by Sung Woo Lee – Integrated assessment – Practical application of probabilistic methods, organized by Alan O’Connor and Jens Sandager Jensen – Life-cycle structural engineering, organized by Fabio Biondini – Lifetime-perspective design of Kwangyang suspension bridge with main span 1545m, organized by Soon-Duck Kwon and Jaehong Kim – New developments in large-scale model studies of bridge components and systems subjected to earthquakes, organized by Mehdi Saiidi – Practical application of BMS and BMS-DB, organized by Ayaho Miyamoto and X. Shao – Practical applications of SHM techniques for railway systems, organized by Soobong Shin and Jeong Tae Kim – Seismic design and performance issues for highway bridges, organized by Murat Dicleli – Special session on Incheon bridge, organized by Jae-Yeol Cho and Hyun-Yang Shin – Structural health monitoring on cable supported bridges, organized by Chan Min Park – Structural robustness, organized by Franco Bontempi and Uwe Starossek

IABMAS’08 MINI-SYMPOSIA

– Information technology for lifetime management of bridge, organized by Sang-Ho Lee, Ayaho Miyamoto, Risto Kiviluoma and Jerome P.Lynch – Integrating health monitoring and lifecycle management of bridge and highways, organized by Emin Aktan, Xiaolin Meng, Leo Klatter, Hitoshi Furuta, and Hae Sung Lee

XLI – Monitoring and assessment of bridges using novel techniques, organized by Alfred Strauss and Dan M. Frangopol – Research & applications for bridge health monitoring, organized by Necati Catbas, Joan Casas, and Hitoshi Furuta – Smart sensing and monitoring technologies for bridge maintenance, safety and management, organized by Chung-Bang Yun and Hoon Sohn – Sustainable bridges, organized by Paulo Cruz, Lennart Elfgren and Jens Sandager Jensen

XLII CONFERENCE SPONSORS (as of April 1, 2008)

MAIN SPONSOR

Samsung C&T Corporation, Korea

SPONSORS

AP – Bridge Construction Systems, Portugal ASCE-SEI, American Society of Civil Engineers – Structural Engineering Institute, USA ASCP,The Portuguese Group of IABMAS, Portugal ASSISi, Anti-Seismic System International Society, Italy ATLSS, Center for Advanced Technology for Large Structural Systems, Lehigh University, USA Betar, Betar Consultores, Lda, Portugal CAU, Chung-Ang University, Korea CERIC, Civil Engineering Research Information Center, Korea COSEIK, Computational Structural Engineering Institute of Korea, Korea COWI A/S, Denmark CU, University of Colorado at Boulder, USA DGU, Dongguk University, Korea ECCS, European Convention for Constructional Steelwork, Belgium IABSE, The International Association for Bridge and Structural Engineering, Switzerland FHWA, Federal Highway Administration, USA JSCE, Japan Society of Civil Engineers, Japan JSMS, The Society of Materials Science, Japan JSSC, Japanese Society of Steel Construction, Japan KBRC, Korea Bridge Design & Engineering Research Center, Korea KCI, Korean Concrete Institute, Korea KEC, Korea Expressway Corporation, Korea KICT, Korea Institute of Construction Technology, Korea KICTEP, Korea Institute of Construction & Transportation Technology Evaluation and Planning, Korea KISTEC, Korea Infrastructure Safety & Technology Corporation, Korea KG-IABSE, Korean Group of IABSE, Korea KSCE, Korean Society of Civil Engineers, Korea KSSC, Korean Society of Steel Constructions, Korea KU, Kansai University, Japan KU-CER, Department of Civil and Earth Resources Engineering, Kyoto University, Japan MJU, Myongji University, Korea MLTM, Ministry of Land, Transportation and Maritime Affairs, Korea NCREE, National Center for Research on Earthquake Engineering, Taiwan, R.O.C OZ – DiagnOstico, levantamento e Controlo de Qualidade, Portugal RCEAS, P.C. Rossin College of Engineering and Applied Science, Lehigh University, USA SEOUL, Seoul Metropolitan Government, Korea SISTeC, Smart Infra-Structure Technology Center, KAIST, Korea SNU, Seoul National University, Korea TRB, Transportation Research Board, USA

XLIII T.Y. Lin, T.Y. Lin International, San Francisco, USA UM, University of Minho, Guimaraes, Portugal UniS, University of Surrey, Guilford, UK UPC, Technical University of Catalonia, Barcelona, Spain

XLIV TECHNICAL EXHIBITION PARTICIPANTS

Bridge Inspection Robot Development Interface (BIRDI), Korea Cervenka Consulting, The Czech Republic/CNG Softtek Co., Ltd., Korea Daelim Industrial Co., Ltd., Korea Daewoo Engineering & Construction/Busan Geoje Fixed Link Corporation, Korea Fixon Inc., Korea Geomonitoring, Korea High Performance Construction Material Research Center, Korea Hyundai Engineering and Construction Co., Ltd., Korea Korea Bridge Design & Engineering Research Center (KBRC), Korea Korea Construction Engineering Development Collaboratory Program (KOCED), Korea Korea Expressway Corporation (KEC), Korea Korea Institute of Construction Technology (KICT), Korea MIDAS Information Technology Co., Ltd., Korea Pure Technologies, USA Pyungsan SI Ltd., Korea Samsung C&T Corporation, Korea Smart Infra-Structure Technology Center (SISTeC)/Infra-Structures Assessment Research Center (ISARC), Korea SK Engineering & Construction Co., Ltd., Korea Strand7 Pty Ltd., Australia/CNG Softtek Co., Ltd., Korea Transpo Industries, Inc., USA

XLV

T.Y. Lin Lecture

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Fatigue of steel bridge infrastructure

J.W. Fisher & S. Roy Center for Advanced Technology for Large Structural Systems, Lehigh University, Bethlehem, PA, USA

ABSTRACT: This paper presents an overview of the development of fatigue design provisions and detailing for steel bridge structures. The historical performance of welded steel bridges in the USA is reviewed with focus on fatigue cracking at cover plate and similar attachment details, as well as distortion induced cracking at web gaps. The role of materials including modern High Perfor- mance Steels (HPS) is reviewed. Variable amplitude loading is also examined and compared with current design of steel bridges for fatigue resistance. Also examined are methods to improve and retrofit fatigue sensitive details, including modern post-weld enhancement by Ultrasonic Impact Treatment (UIT). Orthotropic steel decks are reviewed based on the results of two full-scale proto- type laboratory fatigue tests which identified the complex behavior that occurs at fatigue sensitive details and were verified by field measurements on field installations. The merits of thin epoxy concrete wearing surfaces is examined in terms of the critical role the deck plate thickness has on the epoxy concrete fatigue resistance and durability.

1 INTRODUCTION

Welded and bolted details for bridges are designed based on the nominal stress range that is calculated using mechanics of material equations and does not include the local effect of stress concentrations of welds and attachments. Since fatigue is typically only a serviceability problem, members are designed for fatigue using service loads. It is standard practice in fatigue design of welded structures to separate the weld details into categories having similar fatigue resistance in terms of the nominal stress. Each category of weld details has an associated S-N curve. The S-N curves for steel details in the highway and railway bridge specifications (AASHTO 2004, AREMA 2005) are shown in Figure 1, which are based on a lower bound to a large number of full-scale fatigue

Figure 1. AASHTO design S-N curves. 3 test data with a 97.5% survival limit. Figure 1 also shows the fatigue thresholds or constant amplitude fatigue limits (CAFL) for each category as horizontal dashed lines. When constant amplitude tests are performed at stress ranges below the CAFL, noticeable cracking does not occur. Typically, small-scale specimen tests will result in longer apparent fatigue lives. Therefore, the S-N curve must be based on tests of full size structural components such as girders (Keating & Fisher 1986). Testing on full-scale welded members has indicated that the primary effect of constant amplitude loading can be accounted for in the live-load stress range. The mean stress or the dead load stress is not significant since locally very high residual stresses from welding exist at the weld toe. Cover plate details and equivalent details such as gusset plates are common on older bridges. Hence it is necessary to examine the service loading that bridges are subjected to.

2 VARIABLE-AMPLITUDE FATIGUE

All service load histories for bridges consist of stress range cycles of varying amplitudes, hereafter called variable-amplitude (VA) loading. However, the design curves of stress range versus number of cycles, commonly called design S-N curves in the AASHTO LRFD bridge design specifications (2004), are based on fatigue tests that were performed under constant-amplitude (CA) loading. A histogram of VA stress ranges can be converted into an equal number of CA stress ranges that produce the same amount of crack growth as is the case for the VA stress ranges. This so-called equivalent CA stress range is based on Miners rule and given by:

where Sri = the ith stress range in the histogram; φi = frequency of occurrence of Sri; and 3 is the slope of the log-log linear S-N line for CA fatigue. This equivalent stress range of Sre is referred to as the root-mean-cube (RMC) stress range. When stress ranges in the variable distribution are higher than the CAFL, and have a frequency of 0.05% or more of the cumulative frequency, the equivalent stress range Sre is used with the log-log linear S-N curves the same way as a CA stress range would be used. This design rule is based on the results of long life variable amplitude fatigue tests conducted by Tilly & Nunn (1980) and Fisher et al. (1983, 1993).

3 FATIGUE DESIGN TRUCK

The fatigue design truck in the LRFD specifications is based on the design truck, which consists of 35-kN front axle and two 145-kN rear axles. The axle spacings are 4.3 m on the tractor and 9 m on the trailer. Two assumptions were made: (1) the stress range induced at a detail by a truck crossing the bridge is proportional to the gross truck weight; and (2) all trucks crossing the bridge produce the same fatigue damage as is done by an equal number of trucks with axle loads of 75% of the design truck, and axle spacings of 4.3 and 9 m. The equivalent fatigue truck weight was calculated (Moses et al. 1987) analogous to calculating the equivalent stress range of a VA spectrum with Equation 1. A constant axle spacing of9mwasfound to best approximate the axle spacing of typical four and five axle trucks responsible for most fatigue damage to bridges.

4 EXPERIENCE WITH FATIGUE CRACKING

Fatigue cracking in steel bridges in the United States has become more frequent in its occurrence since the 1970’s. A large crack was discovered in 1970 at the end of a cover plate in one of the Yellow Mill Pond (YMP) multi-beam structures located at Bridgeport, Connecticut (Fisher 1984). 4 Figure 2. Yellow Mill Pond Bridge: fatigue crack at the end of cover-plate.

Between 1970 and 1981, numerous fatigue cracks were discovered at the ends of cover plates in this bridge as seen in Figure 2. In most cases, fatigue cracking in bridges resulted from an inadequate experimental base and overly optimistic specification provision developed from the experimental data in the 1960’s.The fatigue problems with the older bridges can be avoided in new construction if good detailing practice is followed and each detail is designed such that the stress range due to applied live load is below the design allowable stress range. It is also possible to retrofit or upgrade the fatigue strength of existing steel bridges with poor details. Peening works primarily by inducing a state of compressive residual stress near the weld toe (Fisher et al. 1979). Because the benefit of peening is derived from lowering the effective tensile stress range, it has been found to be most effective when conducted under dead load. In this case, the peening only needs to be effective against live load. Air-hammer peening (AHP) can be a successful repair as long as the crack depth does not exceed the zone of compressive stress. The YMP Bridge structure was retrofitted by peening and gas tungsten arc (GTA) re-melting in the 1970’s. This prevented subsequent crack growth in this heavily used structure until it was replaced in 1997. Subsequently, several beams removed from the original structure were tested in the laboratory (Takamori & Fisher 2000). The tests verified that no fatigue crack growth had occurred in these bridge details after more than 60 million cycles of truck loading and the repairs had successfully prevented further growth at the treated details. At the higher test stress range, the laboratory tests were found to develop fatigue cracks in the weld throat and not at the treated weld toe.

5 DISTORTION-INDUCTED FATIGUE

Most cracks found in bridges were caused by distortion of member cross sections and out-of-plane deformations of webs that induced localized bending stresses (Fisher et al. 1990). Out-of-plane dis- tortion occurs mainly at attachment plates for diaphragms, transverse stiffeners and floor beams.The solution to this problem lies in proper detailing that eliminates the secondary stresses. In most cases, problems with web-gap-cracking can be solved by rigidly connecting the attachment plate to the tension flange. In the case where the distortion is limited, holes may be drilled or cored at the crack tips to temporarily arrest propagation (Fisher et al. 1980). Holes essentially blunt the tip of the crack. In cases where the distortion is displacement limited, the stresses can be reduced by increasing the flexibility of the connection. The flexibility may be increased by allowing a small fatigue crack to remain after drilling or coring holes at the crack tips. Another way to increase the flexibility of the joint is to remove part of the stiffener or connection plate and increase the length of the gap. 5 Figure 3. Retrofitting by connection stiffening.

In case where the distortion is not displacement limited (i.e. load controlled), hole drilling or increasing the flexibility of a connection will not work. In these cases, and even in many displacement limited cases, the best solution to distortion cracking is to increase the rigidity of the connection. In new construction, the bridge specifications now recommend that stiffeners and connection plates should be rigidly connected to both flanges and the web. Figure 3 shows a tee connection bolted to the flange and transverse connection plate in order to “bridge” the web gap.

6 HIGH PERFORMANCE STEELS AND ENHANCEMENT

The development of high performance steels such as HPS Grade 485 W and HPS Grade 690 W has shown that without enhancement of welded details their fatigue resistance is no different than other high strength steels in use during the last four decades (Fisher et al. 1970). Post-weld enhancement of fatigue resistance of welded details such as cover-plates, gussets and stiffeners that are known to experience crack growth from a weld toe is essential for an efficient use of these modern high perfor- mance steels. As noted in the previous section, peening and gas tungsten arc remelting has been used in the past to improve the fatigue resistance of welded details that have experienced fatigue damage. Over the past decade, ultrasonic impact treatment (UIT) has proved to be a consistent and effective means of improving fatigue strength of welded connections. Research at Lehigh University on large scale specimens having stiffener and cover-plate welded details has demonstrated that substantial increases in fatigue strength of these high strength steel welded details can be achieved by UIT in particular the elevation of their fatigue limit (Roy et al. 2003, Roy & Fisher 2005, 2006). UIT involves post-weld deformation treatment of the weld toe by impacts from single or mul- tiple indenting needles excited at ultrasonic frequency, generating force impulses at the weld toe (Statnikov 1997). The treatment introduces beneficial compressive residual stresses at the weld toe and also reduces the stress concentration by enhancing the profile of the weld toe. The UIT equipment consists of a handheld tool consisting of an ultrasonic transducer, a wave guide, and a holder with impact needles; an electronic control box; and a water pump to cool the system. Compared with traditional impact treatment methods such as air hammer peening, shot peening and needle peening, UIT appears to be more efficient and environmentally acceptable. The large scale beam tests showed that although the treated details suppressed crack growth from the weld toe, when the usual end weld size was used the failure mode changed to fatigue crack growth from the weld root. This usually resulted in a longer life but still led to cracking and failure (Roy et al. 2003, Roy & Fisher 2005). For enhanced fatigue resistance it was desirable to prevent 6 Figure 4. Design curve for end welded cover-plate details treated by UIT; 0 ≤ R ≤ 0.1.

Figure 5. Design curve for end welded cover-plate details treated by UIT; 0.1 < R ≤ 0.5. root cracking and this was achieved by increasing the size of the end weld at the cover-plate to the plate thickness, which reduced the stress concentration at the weld root (Takamori & Fisher 2000). The test results also showed that the enhancement in fatigue resistance was dependent on both the stress range Sr and the minimum stress Smin. Substantial enhancement results when the treatment is applied under a high level of minimum stress. This was verified experimentally for weld toes treated by air hammer peening (Fisher et al. 1979). Design curves for Category E’ end welded cover-plates treated by UIT are provided in Figures4&5.Asisapparent in Figure 4, treated details under low minimum stress (i.e. the R-ratio of Smin/Smax is less than 0.1 or the detail treated under dead load) provide a design fatigue limit corresponding to Category B of the AASHTO specifications. None of the end-welded cover- plate details developed fatigue cracks below 110 MPa, the CAFL for Category B (Roy & Fisher 2006). When the treatment is applied before the dead load stress, and the R-ratio is bracketed by 0.1 < R < 0.5, the design fatigue limit is decreased as shown in Figure 5 to 70 MPa which is the fatigue limit for Category C. Although there is enhancement in the finite life, it is not as significant as the increase in the fatigue limit for this class of detail. 7 Cover-plate end welds on existing bridges are not likely to have end weld size same as the plate thickness. More likely the weld size will be about half the plate thickness. The test results indicated that when the R-ratio was less than 0.1, the enhanced fatigue resistance was applicable to the weld toe. There is a high probability that fatigue crack growth will initiate at the weld root as was demonstrated in the girders that were removed from the Yellow Mill Pond Bridge that were treated by air hammer peening and gas tungsten arc re-melting (Takamori & Fisher 2000). This would indicate that inspections should focus on the weld throat to ascertain if root cracking would subsequently develop. Fortunately, there is a significant increase in life for root cracking to occur and the cycles (time) necessary for the crack to propagate across the cover-plate end to the longitudinal welds which is the only way the crack can enter into the girder flange. Normal periods of inspection should identify such throat cracking if it ever occurs.

7 ORTHOTROPIC STEEL DECKS

During the past decade, full-scale laboratory and field testing of portions of several orthotropic bridge decks have been conducted (Tsakopoulos & Fisher 2003, 2005). These tests were carried out to minimize the possibility of fatigue cracking as orthotropic deck systems in service in Europe, UK, Australia and Japan have exhibited fatigue cracking in various components of the steel deck system including rib to floorbeam (diaphragm) connections and the rib to deck or the diaphragm to deck connection, particularly when the deck plate thickness was less than 14 mm and when fillet welds were used to connect the ribs to the deck plate. Particularly sensitive in the orthotropic deck system is the rib-to-diaphragm connection – the welded connection between the transverse (floorbeam) diaphragm plate and the continuous longitudinal ribs that are being supported. Concern for the fatigue resistance of rib-to-diaphragm connections on the replacement deck panels resulted in the development of alternatives proposed by the consulting firm Steinman, Boynton, Gronquist, and Birdsall for the Williamsburg Bridge (Gajer et al. 1996), and by Weidlinger Associates for the Bronx Whitestone Bridge (Fanjiang et al. 2004).

7.1 Prototype laboratory tests The two independent laboratory fatigue test programs on a prototype and an as-built deck panel of the Williamsburg Bridge (Kaczinski et al. 1997, Tsakopoulos & Fisher 2003), and on the proto- type deck panel for the Bronx-Whitestone Bridge were conducted in the multidirectional reaction wall test facility at the Advanced Technology for Large Structural Systems (ATLSS) Engineering Research Center, Lehigh University. The loading configuration represented a single-axle that was consistent with the characteristics of a HS20 design truck axle and the HS15 fatigue truck. This provided a more severe loading condition on the floorbeam diaphragm(s) than if a tandem-axle was used. The laboratory fatigue test programs on full-scale steel orthotropic deck panels for the Williamsburg and Bronx-Whitestone Bridge provided valuable information on the complex behav- ior and fatigue performance of orthotropic deck systems that was in good agreement with finite element models used for design. The results reflected a capacity inherent to the orthotropic deck system to redistribute stresses induced under the elevated truck loadings that were simulated. The effectiveness of design improvements that ultimately enhanced the fatigue resistance of the replace- ment deck panels currently in service are expected to exceed the required 75-year design life on both the bridges.

7.2 Results of long term remote monitoring Long-term monitoring of the Williamsburg Bridge began in August of 1998 and continued for seven months (Connor & Fisher 2001). The measurements indicated that the variable amplitude stress range spectrum has a wider band width than assumed in the AASHTO-LRFD bridge design specifications (2004) for the rib-diaphragm cut-out. Similar results were obtained on the Bronx- Whitestone Bridge (Connor et al. 2003). The field measurement results underscored a need for 8 modifications in the fatigue limit state load in the AASHTO specification for the diaphragm details at the cutout, which should be designed for a fatigue limit state stress range of 3 Sreff . Other elements such as the deck plate-to-rib connection and the floorbeam were observed to satisfy the AASHTO LRFD design fatigue limit state stress range of 2 Sreff .

7.3 Wearing surfaces Various materials have been used and/or tested for wearing surfaces on steel orthotropic-plate bridge decks including Portland cement concretes, asphaltic concretes with or without polymer modification, epoxy concretes, methyl-methacrylate concretes, hybrid epoxy-urethane aggregate surfaces, polyurethane-aggregate surfaces and polyester–sand surfaces. There is considerably more experience with use of these surfacing materials on concrete decks than on orthotropic steel decks. Flexibility of the thin steel deck plate introduces a fatigue loading and other more severe conditions than are present in a concrete deck application. Originally 50 to 70 mm asphaltic concrete surfaces were installed on the orthotropic decks of the Poplar Street Bridge across the Mississippi river in St Louis, MO, the McNaughton Bridge in Peoria, IL and the Champlain Bridge in Quebec between 1960’s and early 1990’s.The second and third asphaltic concrete wearing surfaces installed in the 1980’s on the Poplar Street Bridge were observed to provide poor performance by de-bonding, shoving and rutting, and failed after three and four years of service. All of these structures are also exposed to a wide range of temperature including below freezing. As a part of the 1989–1991 process for selecting a replacement surfacing material for the Poplar Street Bridge, MO, a laboratory test program was carried out at the University of Missouri-Columbia (Gopalaratnam et al. 1989, Gopalaratnam et al. 1999). Concurrently, test panels on the bridge were monitored under normal service conditions. Bridge deck strains were also measured under normal service conditions. In 1992 an epoxy concrete surface was installed on the bridge using the slurry method of application During the first five years of service there were no pop-outs or any other de-laminations of the epoxy concrete overlay from the steel deck. However, longitudinal fatigue cracks were found in the wearing surface at the west and east ends of the bridge over the ribs having crack widths of 1.5 mm to 3 mm. These cracks did not result in de-bonding of the wearing surface until 2003. By 2005 (year 13), some large areas of de-lamination were found in the eastbound lanes. In September 2005, repairs were made to areas where cracks had developed in the T-48 slurry along the trapezoidal rib lines. In 2006, the entire original wearing surface was removed and the deck was resurfaced using the same T-48 epoxy concrete. With its 14 mm thick steel deck plate, fatigue cracking will likely reoccur with a similar service life.

8 CONCLUSION

Fatigue of steel bridges under traffic loading is the most significant issue affecting the service performance of aging transportation infrastructure in the USA and in many other countries around the world. Research and case studies of in-service fatigue cracking of steel bridges over the past 40 years have helped in formulating design guidelines and improved detailing practices, imple- mentation of which have limited fatigue cracking in new construction. However, the risk of fatigue fracture of many steel bridges that were built prior to the implementation of the current practices and the economic impact of replacing the deficient bridges remain high. While fatigue fracture limit state in new steel bridges can be suppressed by avoiding the fatigue critical Category D, E or E’ attachments, the performance of these details in existing bridges may be enhanced by weld toe treatments such as air hammer peening, GTA re-melting, or UIT. Post-weld toe treatments should also be considered in new structures for efficient use of modern HPS, where the attachment details cannot be avoided. The orthotropic deck is the only bridge deck system likely to provide a 100 year life when the deck plate thickness equals or exceeds 16 mm. 9 REFERENCES

AASHTO 2004. LRFD bridge design specifications. Washington, D.C.: American Association of State Highway and Transportation Officials (AASHTO). AREMA 2005. Manual of railway engineering. Washington, D.C.: American Railway Engineering and Maintenance of Way Association (AREMA). Connor, R.J. & Fisher, J.W.2001. Results of field measurements on the Williamsburg Bridge orthotropic deck: Final report on Phase III. ATLSS Report No. 01-01. Bethlehem, PA: ATLSS Engineering Research Center, Lehigh University. Connor, R.J., Richards, S.O. & Fisher, J.W. 2003. Long-term monitoring of prototype orthotropic deck panels on the Bronx-Whitestone Bridge for fatigue evaluation. In K.M. Mahmoud (ed.), 2003 NewYork City Bridge Conference; Proc., New York City, October 20–21. Lisse: Swets & Zeitlinger. Fanjiang, G.-N., Ye, Q., Fernandez, O.N. & Taylor, L.R. 2004. Fatigue analysis and design of steel orthotropic deck for Bronx-Whitestone bridge, New York City. Transportation Research Record 1892: 69–77. Fisher, J.W., Frank, K.H., Hirt, M.A. & McNamee, B.M. 1970. Effect of weldments on the fatigue strength of steel beams. NCHRP Report 102. Washington, D.C.: Highway Research Board. Fisher, J.W., Hausammann, H., Sullivan, M.D. & Pense, A.W. 1979. Detection and repair of fatigue damage in welded highway bridges. NCHRP Report 206. Washington, D.C.: Transportation Research Board. Fisher, J.W.,Barthelemy, B.M., Mertz, D.R. & Edinger, J.A. 1980. Fatigue behavior of full-scale welded bridge attachments. NCHRP Report 227. Washington, D.C.: Transportation Research Board. Fisher, J.W., Mertz, D.R. & Zhong, A. 1983. Steel bridge members under variable amplitude long life fatigue loading. NCHRP Report 267. Washington, D.C.: Transportation Research Board. Fisher, J.W. 1984. Fatigue and fracture in steel bridges: case studies: John Wiley. Fisher, J.W., Jin, J., Wagner, D.C. & Yen, B.T. 1990. Distortion-induced fatigue cracking in steel bridges. NCHRP Report 336. Washington, D.C.: Transportation Research Board. Fisher, J.W., Nussbaumer, A. & Keating, P.B. 1993. Resistance of Welded Details Under Variable Amplitude Long-Life Fatigue Loading. NCHRP Report 354. Washington, D.C.: Transportation Research Board. Gajer, R.B., Patel, J. & Khazem, D. 1996. Orthotropic steel deck for the Williamsburg bridge reconstruction, 14th Structures Congress; Proc., Vol. 1, Chicago, IL. New York:ASCE. Gopalaratnam, V.S.,Baldwin, J.W.,Hartnagel, B. & Rigdon, R.A. 1989. Evaluation of wearing surface systems for orthotropic steel-plate bridge decks. MCHRP Report No. MO-FHWA-89-2. Columbia, MO: University of Missouri-Columbia. Gopalaratnam, V.S., Baldwin, J.W. & Cao, W.-M. 1999. Temperature-dependent performance of polymer concrete wearing surface system on the Poplar Street Bridge. Report No. RDT 99-001/RI 90-16. Columbia, MO: University of Missouri-Columbia. Kaczinski, M.R., Stokes, F.E., Lugger, P. & Fisher, J.W. 1997. Williamsburg Bridge orthotropic deck fatigue test. ATLSS Report No. 97-04. Bethlehem, PA: ATLSS Engineering Research Center, Lehigh University. Keating, P. & Fisher, J.W. 1986. Evaluation of fatigue tests and design criteria on welded details. NCHRP Report 286. Washington, D.C.: Transportation Research Board. Moses, F., Schilling, C.G. & Raju, K.S. 1987. Fatigue evaluation procedures for steel bridges. NCHRP Report 299. Washington, D.C.: Transportation Research Board. Roy, S., Fisher, J.W. & Yen, B.T. 2003. Fatigue resistance of welded details enhanced by ultrasonic impact treatment (UIT). International Journal of Fatigue 25(9–11): 1239–1247. Roy, S. & Fisher, J.W. 2005. Enhancing fatigue strength by ultrasonic impact treatment. International Journal of Steel Structures 5(3): 241–252. Roy, S. & Fisher, J.W. 2006. Modified AASHTO design S-N curves for post-weld treated welded details. Journal of Bridge Structures – Assessment, Design and Construction 2(4): 207–222. Statnikov, E.S. 1997. Applications of operational ultrasonic impact treatment (UIT) technologies in production of welded joint.IIW Doc. No. XII-1667–97. Paris: International Institute of Welding. Takamori, H. & Fisher, J.W. 2000. Tests of large girders treated to enhance fatigue strength. Transportation Research Record 1696: 93–99. Tilly, G.P. & Nunn, D.E. 1980. Variable amplitude fatigue in relation to highway bridges. Proceedings of the Institution of Mechanical Engineers (London) 194: 259–267. Tsakopoulos, P.A. & Fisher, J.W. 2003. Full-scale fatigue tests of steel orthotropic decks for the Williamsburg Bridge. Journal of Bridge Engineering 8(5): 323–333. Tsakopoulos, P.A. & Fisher, J.W. 2005. Full-scale fatigue tests of orthotropic deck panel for the Bronx- Whitestone Bridge rehabilitation. Journal of Bridge Structures – Assessment, Design and Construction 1(1): 55–66.

10 Keynote Lectures

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Managing seismic performance of highway bridges – Evolution in experimental research

M. Saiidi University of Nevada at Reno (UNR), Reno, Nevada, USA

ABSTRACT: Earthquakes of the past four decades have revealed the vulnerability of modern bridges to strong seismic events and have pointed out the need for improvements in design and construction. Earthquake and bridge engineering researchers and practitioners have been responsive by investigating the seismic behavior of bridges and developing new seismic design and detailing methods. Because strong earthquakes are relatively rare the waiting time to evaluate new designs in the field is too long and relying exclusively on field evaluation of new designs is impractical. The clear alternative is experimental testing of bridges under simulated seismic loadings. The utilization of experimental testing data to understand and improve the seismic performance of bridge structures is relatively new. Fifty-two years ago at a meeting of world experts and researchers of earthquake engineering, field observation was the only assessment technique that was discussed with respect to verification of design in physical structures, and laboratory tests were not included in any of the article. Many challenges have existed in experimental studies of bridges: (1) for the research results to be credible, the size of bridge models need to be sufficiently large to represent full-scale structure behavior, (2) large-scale bridge models require large and expensive testing facilities and equipment, and (3) the cost of construction, instrumentation, and data interpretation of models is beyond the available research funds. Because of these limitations, experimental studies of bridges were initially very limited. Researchers resorted to relatively small scale models. Furthermore, they focused primarily on bridge components because the facilities and funding would not allow for subsystem or system experimental studies. Fortunately, the challenges have been overcome gradually for a number of reasons that include: (a) increased awareness of the public about the importance of bridges within the highway network and the consequences of bridge failure on the local, regional, national, and, often, on international economy, (b) development of advanced and robust testing and sensor technologies, (c) availability of funds to construct large-scale testing facilities including large shake tables to simulate the dynamic effect of earthquakes. As a result of these changes, seismic experimental studies of bridges have undergone an evolution. The evolutionary path, however, has not been a single one. Rather it has been multiple path encompassing conventional bridges, retrofit studies, innovative details, materials, and systems, and repair studies. The state-of-the-art in different paths has different level of maturity, thus leaving room for considerable future research in bridge earthquake engineering.

Figure 1 shows examples of bridge failures that triggered laboratory research of bridge components. Because of the high cost and long duration of model construction in addition to high cost of laboratory testing, only a limited number of variables can be studied experimentally. As a result laboratory testing has been generally complemented by analytical studies to better interpret the performance of the models, and to develop reliable computer models that may be used for extensive parametric studies. As the public interest in improved seismic performance of bridges expanded over the decades and the available funds for research and facilities increased, researchers were able to switch from

13 Figure 1. Examples of bridge failure in recent earthquakes. individual component tests to bridge subsystems incorporating the joints and cap beams. More recently with the establishment of the Network for Earthquake Engineering Simulation (NEES) by the US National Science Foundation and the completion of the Miki E-Defense large shake table system in Japan it has become possible to study larger models subjected to dynamic testing that utilize multiple shake tables. Figures 2 shows the evolution in the complexity and size of the bridge component, subsystem, and complete system models, and Figure 3 shows a large-scale, two-span bridge model to be tested at Miki facilities. Innovation in bridge earthquake engineering has encompassed the use of new devices and new materials. Large isolators used mostly in signature bridges have shown a tremendous benefit in reducing the seismic force demand in bridges. Fiber-reinforced polymer (FRP) fabrics have also found their way in bridge seismic retrofit and repair application. Because these devices and mate- rials are new to civil infrastructure, they have required extensive studies and demonstration of performance in the laboratory environment. Figure 4 provides examples of large isolators and various forms of FRP applications. More innovative materials such as shape memory alloys or columns with built in isolators are being investigated to utilize their property in improving the seismic performance of bridges. With the drive toward accelerated bridge construction more challenging details will be required that would potentially require more innovations in detailing and materials. Figure 5 shows examples of column with built-in isolator and SMA reinforced concrete columns. The full article and the presentation will provide more details of the samples discussed in previous paragraphs and present selected experimental results and their design implications.

14 Figure 2. Evolution from single-column test to full bridge system test.

Figure 3. Two-span bridge test model to be tested in Miki.

15 Figure 4. Examples of isolator and FRP application for retrofit and repair.

Figure 5. Innovative details in column plastic hinge zone.

16 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Cost-effective and durable cable-stayed bridges

H. Svensson Leonhardt, Andrä und Partner, Consulting Engineers, Germany

1 INTRODUCTION

All engineered structures are expected to be cost-effective and durable. This is especially true for cable-stayed bridges, which represent a major public investment in infrastructure. Due to their lightness special care has to be taken in this regard. The terms cost-effectiveness and durability are closely connected. In a narrow sense cost-effectiveness can be defined as the lowest possible con- struction costs. These are determined by competition. The types of competition vary from traditional comprehensive tender where contractors can only bid on one design to design-build competitions where each contractor has to offer his own design. The decisive factor for the outcome of the com- petition is the price. In build-operate-transfer competition where the competitors have to finance the project themselves, the requested State subsidy or the toll rate may determine the outcome. In a broader sense cost-effectiveness means the life-cycle costs of a bridge. In addition to the construction costs, those for maintaining a bridge over its entire life are equally important i.e. the durability is of overriding concern. For a cable-stayed bridge the most sensitive structural members are the stay cables. Their resistance against corrosion and fatigue determines their useful life. Since cable-stayed bridges are light structures, their durability also depends on their robustness against outside attacks, e.g. against high winds and ship collision. For each characteristic discussed in the following, examples of cable-stayed bridges are given and the involvement by Leonhardt, Andrä und Partner (LAP) is mentioned.

2 COST-EFFECTIVENESS

Cost-effectiveness of cable-stayed bridges is considered initially as minimum construction cost which is best proved by bidding results. The bids by the contractors can be made either on a given tender design or on contractor’s own designs. Worldwidevarious methods exist to determine the most cost-effective design for a given location, which are outlined below. 2.1 Tender designs The classical method in Germany and other countries to determine the lowest cost for a bridge is to tender an official design. This design is usually worked out by a consultant for the client. Contractors are then asked to offer their bid for this design, sometimes including certain small improvements. One typical example is the Elbe River Crossing at Niederwartha in Germany (Fig. 1), a steel- composite superstructure with a main span of 192 m and one tower only. For the State of Saxony LAP prepared preliminary designs from which a tender design was selected and advertised. Eight groups of contractors offered bids in the range of 12.9 to 15.9 Mio. € . The lowest bidder was awarded the contract. In this particular case LAP also prepared the detailed design and construction engineering for the contractor and supervised the construction on site.

2.2 Contractor’s alternates in germany In Germany the system of so-called contractor’s alternates was introduced already during the 1930s. It means that a contractor may offer his own design in addition to offering the official 17 Figure 1. Niederwartha Bridge, Germany. Figure 2. Rhine River Bridge at Wesel, Germany. design mentioned under Section 2.1 above. The conditions for contractor’s alternates are carefully outlined in the tender documents. The tender design is basically a preliminary design with estimated quantities, and the contractor is paid for the actually built-in quantities by the unit prices of his original bid. For the contractor’s alternate the contractor has to accept full responsibility for the design and is paid by lump sum. The appropriate quality of the contractor’s alternate is ascertained by an Independent Checking Engineer and close supervision on site. The risk for the contractor is bigger when he makes a lump sum bid for his own alternate. An example for a contractor’s alternate is the Rhine River Bridge at Wesel in Germany (Fig. 2) with a main span of 335 m and one tower only. The outer shape of the bridge could not be changed, however, the contractor with the assistance of LAP proposed important structural improvements e.g. changing the stay cables from locked coil ropes to parallel strands (see Section 3.1), to change the construction of the approach bridge to incremental launching and to relocate the connection between the concrete approach and the steel main span. The contract was awarded on this contractor’s alternate for which LAP executed the detailed design and construction engineering.

2.3 Alternative designs in the US In the US the contractor was traditionally not at all involved in the design of a bridge. A consultant prepared the detailed design on behalf of the client which was bid for contractors and the lowest bidder was awarded the contract. He built the bridge according to plans, but was not responsible whether the completed structure would serve its intended purpose. In the late 1970s the cable-stayed East Huntington Bridge across the Ohio River with a main span of 274 m and one tower only had been designed in steel (Fig. 3) and the main foundations were already completed. In that stage the Federal Highway Administration decided that from now on major bridges should be designed independently both in steel and in concrete, and that the alternate with the lowest bid should be built. In this way design competitions were introduced to the US to overcome the reluctance of US contractors to assume responsibility for their own alternate designs. LAP in association with A. Grant & Associates designed the concrete alternate as light as possible (high strength concrete, steel cross girders) in order to fit onto the existing piers. In competitive bidding the concrete alternate won with a low bid of 23.5 Mio. US$ against the steel alternate with 33.3 Mio. US$ and thus proved to be the more cost-effective alternate. For the Houston Ship Channel Crossing at Baytown (Fred Hartmann Bridge), Texas, (Fig. 4) the alternates were given to the same designers. The composite alternate was designed most cost- effective e.g. by using two separate beams with four cable planes to reduce transverse bending and by minimizing the tower wall thicknesses to 30.5 cm by designing them as a truss in transverse direction. This time LAP’s composite alternate won, there were no bids for the concrete alter- nate. The US procedure of creating competitive bidding without burdening the contractor with the 18 Figure 3. East Huntington Bridge, WV,USA. Figure 4. Houston Ship Channel Crossing, TX, USA.

Figure 5. Kap Shui Mun Bridge, Hong Kong. Figure 6. Geo Geum Bridge, Korea. responsibility for a design proved to be successful. Another way of reducing costs by so-called “value engineering”, in which savings introduced by the contractor after a job has been awarded to him are split between the client and the contractor did not prove successful.

2.4 Design-Build In most countries of the world bridges are tendered by the so-called “design-build” method. The client does not prepare a tender design but only states his requirements which each design must fulfill. The contractors are then asked to offer their own designs for a lump sum price. In this way the widest variety of bridge systems, contractors’ experience and consultants’ ability are combined to find out the most cost-effective design. Since contractors today do not have comprehensive design offices within their company they commission the design of their alternate to a consultant. The resulting design evolves in discussions between designer and contractor and takes into account the special abilities of the contractor. The contractor has to accept responsibility for the design. The Kap Shui Mun Bridge in Hong Kong forms part of the connection to the new Lantau Airport (Fig. 5). It was the first design-build project in Hong Kong. A Japanese joint venture headed by Kumagai engaged Greiner Engineering with LAP as special bridge consultant. The requirements of road and light rail traffic together with typhoon conditions resulted in a unique composite double deck bridge (Fig. 5, bottom right) with a main span of 430 m. In 2002 the Geo Geum Bridge in South of Korea was tendered by the government as design-build project (Fig. 6). Hyundai Construction engaged Hyundai Engineering with LAP as foreign bridge consultant to develop the design. The cable-stayed bridge with a main span of 480 m carries a truss double deck steel-composite girder with road traffic running on top and bicycle and emergency 19 Figure 7. Myo Island Bridge, Korea. Figure 8. Orinoco II Crossing, Venezuela. lanes on the bottom flange inside the truss. Out of several bids from different joint ventures this design proved to be most cost-effective. For Lot 1 of the Myo Island Bridge, also in the South of Korea, (Fig. 7) GS Construction engaged Yooshin Engineering with LAP as foreign bridge consultant. In this case a composite steel plate girder was designed for the main span of 430 m which was then selected for construction.The contractor Odebrecht was awarded the Second Orinoco Crossing in Venezuela with a length of 3.6 km and two central cable-stayed bridges back-to-back (Fig. 8). Four lanes of highway traffic and a central railway line are supported by a composite box girder for both main spans of 300 m. Odebrecht hired LAP together with the local consultant Brave for the design and construction engineering. This bridge was completed in 2006.

2.5 Build-operate-transfer (BOT) For very major projects the clients sometimes lack the money to build the bridge and the organization to collect tolls and to manage and maintain the bridge. In this case so-called “build-operate-transfer” projects are tendered in competition between different joint ventures. These usually have to finance the bridge construction, refinance themselves by tolls over about 30 years and then return the bridge in good condition to the client. Therefore, not only the first costs but also the costs for maintenance over 30 years are of importance. This is where the durability of the structure becomes especially important. The outcome of the competition is, therefore, not decided by the price for construction but by the amount of toll or by the required subsidy. In 1998 a joint venture under the contractors Impregilo and HochTief hired LAP as main con- sultant for the design for a 4.1 km bridge across the Parana River between the cities of Rosario and Victoria in Argentina (Fig. 9). The competition was won by the amount of subsidy requested. The design comprises a central cable-stayed bridge with a main span of 350 m and an open con- crete girder built cast-in-place by free-cantilevering. The sea-going ship traffic on the Parana River required independent protections against ship collision, taking into account a possible scour around the pile foundations. The bridge was completed in 2002. When the Ma Chang Bridge in (Fig. 10) was tendered as a build-operate-transfer project, the joint venture of Hyundai Construction and Bouygues hired Hyundai Engineering with LAP as foreign bridge consultant. Their design was successful, comprising a cable-stayed composite bridge with a main span of 400 m. For the sea-going ships the Ma Chang Bridge is provided with strong foundations to withstand potentially high loadings from vessel collisions. In this case, however, seismic loads are governing. The realization of a major crossing of the Baltic Sea between the German Island of Fehmarn and Denmark has been decided by the Governments of the two countries. The bridge will be 20 km long and is designed as a two level structure with a four lane motorway on top and two railway tracks 20 Figure 9. Rosario-Victoria Crossing, Argentina. Figure 10. Ma Chang Bridge, Korea.

Figure 11. Fehmarnbelt Crossing, Germany / Denmark.

on the lower level inside the truss girder (Fig. 11). The main bridge has been designed by LAP as a cable-stayed bridge with three main spans of 724 m each. This design is currently considered to be the most economic solution. The tender process will start in 2008 and will probably take the form of BOT. Once completed this will be the longest bridge in Europe. In the times where public funds appear to be lacking, BOT projects will gain more and more importance for major projects.

3 DURABILITY

Whereas cost-effectiveness is normally meant to comprise only first or construction costs, durability takes into account the life-cycle costs. Only a bridge with a minimum total of construction cost plus maintenance cost over the life of the bridge can, therefore, be considered cost-effective. The consideration of durability thus compliments the consideration of construction costs. Specifically important for cable-stayed bridges is the durability of their stay cables, in particular their resistance against corrosion and fatigue. Furthermore, cable-stayed bridges are sensitive to wind attack because of their lightness and they are also sensitive to ship collision if located in navigable waters. 21 Table 1. Characteristics of different types of stay cables.

Modern locked Parallel wire Parallel strand coil rope cable cable

Characteristics

E·10−6 [N/mm2] 0,170 0,205 0,195 fu [N/mm2] 1470 1670 1870 σ [N/mm2] 150 200 200 biggest Ø [mm] 180 499 Ø 7 109 Ø 0,6 cables pu [MN] 31,0 32,1 24,5 fabricated so L [m] >1000 250 ≈ 200 far max G [t] >80 23 ≈ 20

Figure 12. Aerodynamic Stability of cable-stayed Figure 13. Houston Ship Channel Crossing, TX, Bridges versus Suspension Bridges. USA.

3.1 Stay cables The historical development of stay cables went from locked coil ropes to parallel wire cables to parallel strand cables. This development was intended to improve their durability with respect to corrosion and fatigue and to decrease their costs. In addition, locked coil ropes require large machinery which is only available in Europe whereas wire and strand cables can be assembled locally or even on site.

3.2 Wind resistance Cable-stayed bridges are inherently stiffer than suspension bridges. A comparison of the two bridge types shows that for a main span of 500 m the critical wind speed for a cable-stayed bridge is about double that of a suspension bridge (Fig 12). The Houston Ship Channel Crossing at Baytown (Fred Hartmann Bridge) is located in a hurricane prone area near the Gulf of Mexico (Fig. 4). The open cross sections for the twin beams were shown analytically to be stable during construction for a wind speed of up to 170 km/h, not enough for the expected considerably higher hurricane wind speeds. The beams were thus tied down to the tower foundations (Fig. 13) and indeed remained stable during a hurricane, which hit the site shortly before beam closure. Another striking example for the aerodynamic stability of cable-stayed bridges is the Helgeland Bridge in Norway (Fig. 14) which was designed by LAP in collaboration with Aas-Jakobsen. With a main span of 425 m the concrete beam has a depth of only 1.2 m and a width of 12 m thus rendering 22 Figure 14. Helgeland Bridge, Norway. Figure 15. Helgeland Bridge, Storm. a vertical slenderness of 1:354 and a horizontal slenderness of 1:35. (The horizontal slenderness is critical, because it prevents reattachment of the vortices from the leading edge.) The wind climate near the Polar Circle is very serious. Wind speeds up to 180 km/h with a turbulence intensity of up to 25% had to be expected and did indeed occur even during construction (Fig. 15). During severe storms just before center joint closure, the displacement of both cantilevers came to about 0.6 m, exactly as analytically predicted. In the final stage the bridge proved to be stable against flutter analytically and in tests up to a wind speed of 270 km/h. Both examples show that properly designed and analyzed cable-stayed bridges are durable against wind action, not only in the final stage but also during construction.

3.3 Resistance against ship impact Major waterways with sea-going ship traffic require large navigational clearances. Cable-stayed bridges are often the appropriate type of structure for such crossings. Special consideration must be given how to protect a bridge against collision with major ships which may exert high impact forces. The results of collision tests (Fig. 16) were used by the author to estimate collision forces as a function of ship size and speed (Fig. 17). The probability whether a pier under consideration will be reached by a ship must also to be investigated. Several types of protective measures can be used as outlined below. The safest method to protect piers from ship collision is to place them out of reach on land. The additional costs for an increased span length may be offset by savings for pier protections. Examples are the Yang Pu Bridge with a record span due to navigational requirements and the Second Panama Canal Crossing where the main span length was determined by taking into account not only the current Canal width but also future widening of the Canal to avoid ship collisions. If the water is too wide to be crossed without piers in navigable waters artificial islands may be possible. Their advantage is that they stop a ship slowly thus limiting the extent of damage to the ship’s hull. Their use is often limited if they reduce the flow cross-section too much. For the Houston Ship Channel Crossing (Fig. 4) an artificial island was built for the foundation of the one tower which had to be placed in shallow water. A similar situation exists for the Kap Shui Mun Bridge in Hong Kong (Fig. 5). If these two options are not possible it may be economic to strengthen the piers and their foundations sufficiently to withstand a collision. In this case the vertical loads from the bridge are favorably combined with an increased horizontal resistance of the foundation. Examples are the Ma Chang Bridge (Fig. 10) and the Helgeland Bridge (Fig. 14). All piers of the Rosario-Victoria Crossing inArgentina (Figs. 18 and 19) rest on pile foundations. In 30 m deep water with local scour of up to 12 m a free pile length of up to 42 m has to be taken into account. Independent defense structures, designed as sacrificial structures by exploiting their plastic capacities were the appropriate solution. 23 Figure 17. Equivalent Collision Forces by Figure 16. Collision Test by Voisin. Svensson.

Figure 18. Rosario-Victoria Bridge, Ship Collision. Figure 19. Rosario-Victoria Bridge, Pier Protection.

The durability of cable-stayed bridges can be safeguarded by applying one of the above counter measures against ship collision

4 CONCLUSION

Since their rediscovery in the 1950s cable-stayed bridges have developed into cost-effective and durable structures, always provided the experience in design, construction and maintenance gained since then is carefully applied. The cost-effectiveness is realistically determined by the contractors in competitive bidding. Durable cable-stayed bridges require stay cables robust for corrosion and fatigue, and require as well resistance against the attack of wind and earthquake. By optimizing the cost-effectiveness and the durability of cable-stayed bridges the minimal life-cycle costs are achieved.

24 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

A new concept of orthotropic steel bridge deck

Man-Chung Tang Chairman of the Board, T.Y. LIN International, San Francisco, USA

ABSTRACT: A new concept of orthotropic steel bridge deck is proposed to improve its durability and economy. It is suggested to hot-roll the longitudinal ribs. Thus, the ribs can be made larger and with a more effective cross section that is not possible for cold bent ribs. Together with a thicker top plate these ribs can be spaced at 800 mm apart and span 8 meters or longer. This will significantly reduce the cost of fabrication of orthotropic decks while the unit weight per square meter remains practically the same. An orthotropic deck with such a configuration will be more economical and more durable.

1 INTRODUCTION

In a contemporary design of the orthotropic steel deck, the deck, besides supporting all traffic loads, it is a part of the upper flange of the girder and is the upper flange of the transverse floor beams. This saves material and results in a light weight system which is very important for long span bridges. In mid 1950s, W. Cornelius presented a simplified method for the analysis of orthotropic deck with trapezoidal ribs. Later, G. Fischer’s contribution, with consideration of the torsional stiffness of the ribs, simplified the design of close rib orthotropic deck which is more economical. However, Cornelius, with the firm MAN, patented the trapezoidal rib configuration in late 1950s. This forced other construction firms to use many other shapes for the ribs, such as rounded shape, wine glass shape, etc. as shown in Fig. 1, which were usually less efficient. Lately, almost all orthotropic decks have the closed trapezoidal, torsional stiff ribs. The trapezoidal ribs are more efficient and more economical. However, the fabrication of orthotropic steel deck is more labor intensive and requires much more stringent quality control than a concrete deck. Therefore, for economic reasons, it was not widely used after the reconstruction period except for very long span bridges where weight saving is of utmost importance. The bridge building of Japan and later in China revitalized the idea of orthotropic deck. This is especially true in China where labor is relative inexpensive, many medium to long span bridges are finding the use of orthotropic steel bridge deck attractive. Almost all recent bridges have exclusively trapezoidal ribs. For this reason, all further discussion in this paper will concentrate on orthotropic decks with closed, trapezoidal ribs.

Figure 1. Various shapes of longitudinal ribs.

25 2 COMMON PROBLEMS IN ORTHOTROPIC STEEL BRIDGE DECKS

A new concept of orthotropic steel bridge is proposed in this paper is to avoid problems already found in existing orthotropic steel decks. It is therefore important to recognize what problems are there in the existing bridge decks. Since its application in the 1950s, various problems have been discovered in the orthotropic decks. In general the problems can be divided into two groups: the failure of the steel deck and the failure of the pavement. Very often, it has been a common practice that the steel deck is designed by the structural engineer while the pavement is designed by a material engineer. They usually do not communicate with each other. However, in reality, the steel deck acts compositely with the pavement as a monolithic structural system and they must be analyzed together as one unit. Improperly designed steel deck can cause the pavement to fail while an improperly designed pavement can create high stresses in the steel deck. This fact has only been properly addressed in recent years.

2.1 Failure in the orthotropic steel deck Cracking in orthotropic steel deck has been reported in many literatures. These cracks are mainly caused by stress concentrations in steel members due to local constraints. They can be generally grouped into the 4 types of cracks as shown in Fig. 2. A. Under the wheel load, the stress at and near point A can be very high. This is because the deck plate is supported on the one side by the web of the floor beam or diaphragm while there is no support on the other side, above the trough of the rib. Kolstein calculated 200 MPa (29 ksi) for this stress in a 12 mm deck plate. C. Seim measured 33 ksi (225 MPa) in the 11 mm deck plate of the Luling Bridge. The bending moment at this location under the wheel of the vehicle is rather constant while the stress increases when the deck plate gets thinner. Cracking at this point was found mainly in deck plates with a thickness 12 mm or less. For deck plates 14 mm or thicker, such cracks are rather uncommon.

Figure 2. Common problem spots in an orthotropic steel deck.

Figure 3. Stress concentration at discontinuity spot A.

26 One attempt to address this problem is to install a diaphragm in the trough of the rib to provide support to the deck plate. This diaphragm, or filler plate, however, is rather costly and difficult to install. Except for area where there is a compression force in the transverse direction of the deck that requires the diaphragm to transfer this force, the tendency is not to use this diaphragm in the trough. A thicker deck plate is a better solution for this problem. B. Point B is the intersection of three welds: the fillet weld (1) between the web of the rib and the deck plate, the horizontal weld (2) between the deck plate and the floorbeam web, and the weld (3) between the rib wall and the floorbeam web. For a long period of time, it has been a rather standard practice to cope the corner of the floor beam web to separate these three welds. However, to obtain a really good detail the end of both weld 2 and weld 3 must be properly rounded and ground smooth, which is usually not done due to cost of labor. Such a detail will lead to fatigue and corrosion problems. Recent practice has been not to cope the floorbeam web and let these welds intersect each other. C. When a vehicle travels on the bridge deck, the wheels advance from floorbeam to floorbeam. The longitudinal ribs deflect up and down under the wheel load depending where the wheel is located. This up and down bending of the rib creates a longitudinal movement of the rib wall at point C, which translates into an out of plane bending of the floorbeam web at this location, Fig. 4. This point C is especially vulnerable because it is also the end of a difficult weld. To reduce the stress concentration at this point, a detail as shown in Fig. 5 is usually recommended. But such a detail is very labor intensive and therefore expensive. Another option which has been used successfully is to delete the opening and use a round bottom rib and a continuous weld between the rib wall and the floorbeam web.

Figure 4. Rotation of rib and Floorbeam Web under Wheel Load.

Figure 5. Detail at Point C.

27 D. This point is away from the floorbeams. Because the ribs are closed, this weld can only be produced from the outside face. The prevailing requirement is to have a weld size of about 80 to 85% of the rib web thickness. Blow-out is not allowed. It makes this weld difficult to achieve. In addition, even if the weld is done properly, it is still an eccentric weld with respect to the thickness of the rib wall because the weld is on one side only. There are other problems involved with the orthotropic decks. However, the above mentioned problems are the major ones.

2.2 Problems in the pavement There are mainly two types of paving material used on orthotropic steel decks: bituminous type and polymer type. Mastic asphalt, Gussasphalt, etc. are bituminous type. Epoxy asphalt, polyurethane, and polyesters are polymer type. There have been some exceptions, such as concrete bonded to the deck with shear connectors. But these are really rare. The pavement on many orthotropic decks failed in a short period of time after the bridge is open to traffic. There are four main problems connected with the deck pavement: cracking, separation, or de-lamination from deck plate, rutting, and loss of skid resistance. Rutting and loss of skid resistance are mainly due to poor performance of the pavement material. Cracking and separation are mainly due to overstress or fatigue. Selection and composition of the pavement material will not be discussed here. The following discussion will concentrate on the problem of stresses. Cracking in the pavement is a result of high tensile stress in the pavement. Separation is due to the failure of the bonding layer between the pavement and the deck plate. Because the pavement is adhered to the deck plate, the entire system, including both the orthotropic steel deck and the pavement material must be treated as a monolithic structure in the calculation of the stresses in the pavement. Thus, to analyze the stresses in an orthotropic steel bridge deck with a pavement, the properties of the steel deck, which include the plate thickness, the rib spacing, the floorbeam spacing, the rib stiffness, etc. and the pavement characteristics, such as the pavement thickness and stiffness must be properly considered.. Fig. 6a shows the stress in the steel deck plate and the stress in the top of the pavement for various modulus ratio (n) between steel and the pavement material as presented by R. Wolchuk. When asphalt based pavement is used, the modulus of elasticity of the pavement material varies with temperature. The warmer the temperature, the softer the material is. Consequently, the modulus ratio can change from about 10 to about 500 in the usual range of temperature of the deck. It can be seen from Fig. 6a that the stress in both materials can change significantly. Therefore, for the design of either the pavement or the steel deck, temperature effect must be properly considered. C. Seim, R. Manzanarez and T. Ingham have explored this relationship extensively. Some of their results are reproduced in Fig. 6b to 6d. Fig. 6b shows the shear stress in the bonding layer between the steel deck plate and the pavement for an orthotropic steel bridge deck under a typ- ical AASHTO wheel load. This is based on a deck with 8 mm thick, 280 mm high trapezoidal ribs spaced at 600 mm centers. It shows two groups of curves. The upper group is for n = 10, the lower group is for n = 500, which, together, represent the two rather extreme temperature conditions a regular bridge deck may experience. The four curves in each group are for a deck plate thickness of 10 mm, 12 mm, 14 mm and 16 mm. The important message from this plot is to show, similar to the one by Wolchuk in Fig. 6a, that the maximum stress depends on many variables. It is interesting to note that the maximum shear stress is lowest for a pavement thickness around 38 mm. Fig. 6c and 6d are results extracted from the analysis for the Second Yangtze River Bridge in Nanjing, China. It is a 628 m span cable-stayed bridge with a box shaped main girder. The deck plate thickness is 14 mm uniform over the entire bridge deck. It has 8 mm thick, 280 mm high trapezoidal ribs spaced at 600 mm centers. The data shows that the maximum tensile stress in the pavement at low temperature decreases rapidly with increase of pavement thickness until it reaches about 60 mm. 28 Figure 6. Stresses in the steel deck and the pavement under AASHTO wheel load.

It is important to note that if the pavement thickness is very thin, say, less than 15 mm, both the maximum tensile stress in the pavement and the shear stress in the bonding layer can be very high. This probably explains why many of the thin pavements on orthotropic decks have failed. The problem compounds when the steel deck plate is also very thin. Kolstein and Wardenier have presented a detailed report on a bridge in Holland which has 12 mm deck plate and 8 mm thick pavement that had required extensive repair due to failures in both the steel deck and the pavement.

3 THEORETICAL ANALYSISAND PRACTICAL EXPERIENCE

Data in Fig. 6a-d show that the stress condition in an orthotropic bridge deck is very complex. Adding to this complexity is the hard to predict residue stresses in the welds and the cold bent steel ribs. Any numerical result appears to be rather approximate at best. Under such circumstances, performance record of existing bridge decks should be a more important gauge for future designs. Many of the recommendations in the Codes, AASHTO or Eurocode, have been based on past experience, with some help from laboratory tests of special details. Engineering is not science. Engineering is based on experience and aided by science as one of the tools. It is important to supplement complex numerical analysis with careful learning from successful experience. 29 4 DECK PLATE THICKNESS AND RIB WALL SPACING

The current AASHTO-LRFD Code recommends that the deck plate thickness should be at least 4% of the distance between the webs of the ribs, which represents a deck plate span to thickness ratio of 25. It is the writer’s opinion that this may be still slightly too thin. The orthotropic steel bridge deck on three bridges have performed very well: The deck plate of the San Mateo Bridge over the San Francisco Bay has a span to thickness ratio of 24. It has performed well for over 35 years. The re-decking of the Golden Gate Bridge has also been performing well for over 20 years. Its deck plate has a span to thickness ratio of 23.3. The western suspension spans of the San Francisco Oakland Bay Bridge had preformed well for over 30 years. Its deck plate has a span to thickness ratio of 24. Only its skid resistance had declined to a level that might require repair or replacement after 30 years of continuous service. All these three bridges have 50 mm epoxy asphalt pavement. Therefore, this experience should only be applied to decks with a 50 mm epoxy asphalt pavement. There have also been unsuccessful examples of bridges with similar combinations. But each can be traced to defects in the design or construction. Both the fabrication of the orthotropic deck and the placement of the epoxy asphalt pavement require very stringent quality control. If this can not be guaranteed, the deck system can not be expected to perform well. Instead of the simple span to thickness ratio, C. Seim prefers to use the span to deflection ratio as a design parameter. Obviously, these two ratios are inter-related. The span to deflection ratio has the advantage that it can be recalibrated for a modified wheel load other than the AASHTO wheel load. But, the span to deflection ratio varies with the cube of span to thickness ratio; it is also directly proportional to the intensity of the wheel load. Any transformation will be simple. For example, if a span√ to thickness ratio of 24 under the current AASHTO load is considered acceptable, a 14.3% (= 3 1.5) decrease in the span to thickness ratio to 21 should be used if the wheel load is increased by 50%. However, this transformation considers only the deformation of the deck plate. When the wheel load is increased by 50%, other factors may become more prominent so that such simple extrapolation should only be used with extreme caution.

5 RECENT APPLICATIONS IN CHINA

China is building many new long span bridges in the last two decades. The spans of these bridges are getting longer and many of them have orthotropic steel bridge deck. The deck system of several early steel bridges had problems, especially in the pavement. Most of these early bridges use a deck plate of 12 mm with bituminous type pavement. A 14 mm deck plate with 50 mm epoxy asphalt pavement was used for the first time in the design and construction of the Second Yangtze River Bridge in Nanjing. The span to thickness ratio of the deck plate of this bridge is 21.4 to allow for possible overloaded trucks that are probably unavoidable in China at the present time. This bridge was open to traffic in early 2001. The deck has been performing very well. Since then, most Chinese steel bridges, including the world’s longest cable-stayed bridge, the Sutong Bridge, have been designed with a 14 mm or thicker deck plate and most have used 50 mm epoxy asphalt pavement. The performance of these bridges has been very good until today.

6 COST OF ORTHOTROPIC STEEL BRIDGE DECK

The cost of an orthotropic steel bridge deck consists of material cost, fabrication cost, transportation and installation cost. Material cost is usually less than about 35% of the total cost. A major cost is in fabrication, with the major expenses spent on the welds at the intersections of the floorbeam and the longitudinal ribs. And, this is the one item cost saving is possible in an orthotropic steel bridge deck. 30 Figure 7. The Proposed Hot-Rolled Rib.

7 A NEW CONCEPT

A new concept is proposed to eliminate the known problems described above. It is based on the use of a hot-rolled rib with variable thickness as shown in Fig. 7. R. Wolchuk had suggested the use of hot rolled ribs to reduce the residue stresses in the ribs created by the cold bending process. However, hot rolling also offers the opportunity of optimization of the shape of the rib. The rib proposed in Fig. 7 is 380 mm deep, which is deeper than the ribs commonly used today. To optimize its effectiveness, the bottom portion of the rib is thickened to 20 mm. This thickening can be increased if needed for larger wheel loads or longer rib spans. This rib configuration is more efficient than the current trapezoidal ribs with uniform thickness. The bottom portion of the rib is rounded so a continuous weld can be produced easily from one end of the rib to other end of the rib. C. Seim has also suggested that the bottom 25% of this weld be a full penetration weld to further improve its resilience. Use of a rounded shape without the cut-out has performed well in a number of earlier bridges. This will eliminate the difficult and expensive weld at the termination points C as shown in Fig. 3. In connection with this round bottom rib, the web of the floorbeam or the diaphragm must be flexible. R. Wolchuk suggested that stiffeners or bottom flange that can restrain the out of plane bending of the web should be placed at a distance of at least twice the depth of the rib from the bottom of the ribs. To improve on the weld between the web of the rib and the deck plate, a local thickening is introduced to the top portion of the rib walls. This is to address the welding problem at point D in Fig. 3. The rib web thickness is increased to 11 mm at the top so it will be easy to produce an 8 mm to 9 mm weld from one side only without risking blow out. The taper of the upper face of the ribs can be made in the hot roll process to save an additional step in the fabrication. A 50 degree taper is suggested. However, this can be modified to suit local conditions. With a stronger rib, a thicker deck plate should be used to allow larger rib spacing. An 18 mm deck plate with 800 mm rib spacing will give a span to thickness ratio of 22.22 which is a satisfactory ratio based on experience described above. Such a longitudinal rib can easily span 8.00 m. With such a span, there will not be a need for intermediate floorbeam in most cases, which means, there will be only diaphragms that spaced at 8.00 m intervals. The advantage of a diaphragm over a floorbeam is that it can be made more flexible in the out of plane direction. This flexibility will reduce the stresses in the weld between the rib wall and the diaphragm from the out-of-pane bending of the diaphragm web caused by the rotation of the ribs at the diaphragm as the wheel proceeds from diaphragm to diaphragm on the deck. Such a deck is also more efficient in load transfer because it is closer to a plate action in carrying the wheel. When the rib span is very small, most of the wheel load is carried by the one single rib under the wheel. As the rib span is increased, the orthotropic deck acts more like a plate and more ribs are participating in carrying the wheel load. Without better performance record of other pavement materials, it is recommended a 50 mm epoxy asphalt pavement be used until testing can prove other pavement types satisfactorily. 31 8 COMPARISON

Based on available data from recently completed long span bridges in China, the typical config- uration of the orthotropic deck has 8 mm thick, 280 mm high trapezoidal ribs spaced at 600 mm center to center; Deck plate is 14 mm thick and the diaphragm, or floorbema is 12 mm thick spaced at 4.00 m. The box girder is usually 3.5 m deep. Take a 12 m wide and 16 m long deck panel. The conventional deck will have 20 longitudinal ribs and 4 floorbeams or diaphragms. Thus, there are 20 × 4 × 80 intersections of ribs and diaphragms. Such a panel, including the diaphragms will weigh about 2.81 kN per square meter. If the proposed new configuration is used, the same 12 m by 16 m panel will have 15 longitudinal ribs and 2 diaphragms with 15 × 2 = 30 intersections. The weight of such a panel is about 2.89 kN per square meter. The saving in welding is very significant. The number of intersections is reduced from 80 to only 30. And the number of welds between the ribs and the deck plate is reduced by 25%, while the weld between the diaphragms and the deck plate is reduced by 50%. In addition, the welds at the intersections are much easier to produce, without any grinding.

9 SUMMARY

Modern orthotropic steel bridge deck has been in service for about 50 years. Its performance is generally quite satisfactory. However, there are still problems that have been causing failure in some bridges. It is therefore prudent to modify the current design to deal with these problems. With advanced computer software available today the analysis of any such bridge deck is not difficult. However, due the complexity of such a structure, computer analysis can only reproduce a part of the actual behavior of the deck. Various assumptions in the establishment of the computer model are idealized assumptions. The most important item missing is the residue stress due to welding and manufacturing. These stresses could easily be as high as the stresses due to live loads. These stresses do not have much effect on global stresses. They are usually very local in nature, mainly due to unavoidable local constraints in the structure. Adding to the unknown is the behavior of the pavement. The pavement is bonded to the steel deck and form a monolithic structure with the deck. Therefore, any analysis of the steel deck must include all proper characteristics of the pavement. Consequently, it is recommended that past performance record of existing decks be incorporated in the design, with the theoretical analysis giving a comparative value to the design. Past performance record of existing bridges with orthotropic steel deck has provided clues on what may or may not work. The problems in the steel deck and in the pavement have been summed up above. A new concept, or configuration is proposed to overcome or avoid these problems, both in the steel deck and in the pavement. For example, high stress concentration at Point A, Fig. 2, is reduced by using an 18 mm thick deck plate. The fatigue problem at Point C is eliminated. The welding problem at Point D is significantly improved. By using an 18 mm thick deck plate with a span to depth ratio of 22.22 the stress problems in the pavement should be eliminated based on experience gained from successful existing structures. The proposed cross section of the longitudinal rib is more efficient because it has a thickened bottom flange. With extended rib spacing and rib spans, or diaphragm spacing, the proposed deck configuration also saves extensively in fabrication costs. Obviously, to produce this new rib, the steel mill must establish a new production line to roll them. However, such a production line is not expensive compared to the possible savings in the fabrication and other advantages. A single medium size bridge will recoup all costs involved in such a production line.

32 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Health monitoring of structures & related education and training needs of civil engineers

A.A. Mufti ISIS Canada, University of Manitoba, Winnipeg, Manitoba, Canada J.P. Newhook ISIS Canada, Dalhousie University, Halifax, Nova Scotia, Canada

1 INTRODUCTION

The Intelligent Sensing for Innovative Structure (ISIS) Canada Networks of Centres of Excellence (NCE) program has been pursuing research and field implementation of structural health monitor- ing technology for innovative infrastructure using Fibre Reinforced Polymers (FRPs) to reinforce concrete since 1995. The authors of this paper have a collective record of over 30 field projects, which have employed some aspect of health monitoring technology. Concurrent with this research work, the authors have been part of an ISIS team developing resources and strategies for educating the engineering community about SHM. This includes creating a series of design manuals, such as the Guidelines for Structural Health Monitoring and the Civionics Specifications. The members of ISIS Canada have produced a variety of documents and guidelines to assist practicing engineers with the implementation of SHM systems. More recently, ISIS Canada has produced an educational module for usage in the undergraduate engineering curriculum. Concurrently, a graduate course SHM for civil engineering graduate students has been developed. Finally, the authors have been active in developing workshops, which introduce SHM to researchers and practicing engineers. Based on these activities, the authors have formulated an opinion on the education needs and changes required such that SHM can be successfully implemented in civil infrastructure manage- ment. The changes suggested will not only impact the adoption of SHM technology but will also lead to a better integration of civil engineering practices in general. This paper discusses three key elements of for implementation of performance engineering: Structural Design, Structural Health Monitoring and Civil Engineering Education. It is hoped that in doing so, further discussion and activity related to this objective will be stimulated.

2 CIVIL STRUCTURAL DESIGN PROCESS

Traditionally, structural components were designed by the working stress method, which required that the maximum stress due to nominal dead loads and live ‘service’ loads was a fraction of the maximum stress that the component could withstand. The ratio of the failure stress to the ‘actual’ stress was and is still known as the factor of safety. As long as the factor of safety was a sufficiently large number, say 2 or 3, the design was deemed to be safe. While the working stress design has served the engineering community well for a long time, it has led to structures with non-uniform margins of safety. Works of researchers in the late 1960s and early 1970s laid the foundations for modern structural design codes, which are based on the concept of structural reliability. In Canada and some parts of Europe, the probabilistic-based design method is referred to as the limit states design method. In the USA, the same method is referred to the load and resistance factor design (LRFD) method. It is important to note that the working stress method, which is still practiced in many parts of the USA as well as in many other countries, was used to design a very large part of the current 33 Figure 1. Reduction in value of safety index with time. stock of infrastructure in practically every country. The earlier working stress design methods requiring manual calculations and based on simplifying assumptions usually led to structures with large reserves of strength. By careful use of SHM, these untapped reserves of strengths in civil structures can be utilized, thus affecting economies in the design, building and evaluation of civil structures. The use of SHM is also expected to encourage innovative structures. The level of safety of a structural component is determined by the probability of its strength exceeding the maximum load effect it would receive during its lifetime. Safety margins in a struc- tural component are measured by the probabilistic-based safety index, β. A detailed explanation of the determination of safety index is given in the paper through reference to live load distribution in a girder bridge. The resistance of a structure ®, such as a bridge, can decrease with time due to environmental and other time-dependent effects including material deterioration and fatigue. Similarly, the live loads (S) that a bridge is called upon to carry can also increase with time, which has been experienced in Canada with the changes in vehicle weight regulations demanded by the economics of freight hauling by highway trucks. As illustrated in Figure 1, the net result of the decrease of resistance and increase of loads over time is the increase of the overlapping areas of two curves. Hence, it can be seen that the notional safety indices of the components of a bridge can decrease with time. Should such a reduction in the safety index always give us a cause for concern? Fortunately, a bridge can continue to be used even if its safety index is smaller than 3.5, but only if there is reliable knowledge about the structure. Bakht and Jaeger have shown that each time a bridge was tested surprising results were observed. The actual behaviour of bridges was usually found to be different than that of the mathematical model used in the original design. In most cases, the actual strength of bridges was considerably higher than the theoretical strength.

3 STRUCTURAL HEALTH MONITORING

An interesting question that should be asked is, “Can structural health monitoring in the field be useful in managing risk?” The answer would be a resounding, “yes.” The term ‘civionics’was coined recently to be parallel to ‘avionics,’ which is defined in American English language dictionaries to denote the application of electronics to aviation and astronautics. The term ‘civionics’ refers to the application of electronics to civil structures to determine the state of their health. The use of sensors to monitor the response of a structure or its model to applied loads is not new, nor is 34 bridge evaluation by field testing, which includes both diagnostic and proof testing. What is new, however, is the use of SHM through civionics. In the past, civil engineers have gained knowledge about the integrity of civil structures largely by means of manual inspections, and rarely by nondestructive evaluation (NDE) and interpretation of data using conventional technologies. The structural engineering profession has relied heavily on the evaluation parameters given in codes of practice that lead to conservative and often costly conclusions about the strength of existing structures. In order to remain competitive in today’s global economic environment, the owners of civil structures need to minimize the user costs involved with the unnecessary closing of the structures and service disruption caused by outdated and time-consuming inspections following extreme events, such as strong-motion earthquakes, hurricanes, flash floods, or other extreme events. In the evaluation of any structural system, it is important to be able to assess specific performance issues related to serviceability, reliability and durability. To effectively quantify the system’s performance requires a means to monitor and evaluate the integrity of these large civil structures with minimal interruption of service. The SHM data, besides allowing owners to better allocate their resources towards repair, replacement or rehabilitation of the structures, will also be useful in future projects in estimating the life cycle costs of the structural system compared to the initial cost. An efficient SHM system should be autonomous and capable of continuous monitoring, inspection and damage detection. The direct benefits from SHM systems are very great and include: • monitoring and evaluating structures in real-time under service conditions; • reducing downtime; • improving safety and reliability; and • reducing maintenance costs. With reduced downtime and improved reliability, in-service structures can be used more productively with less cost. The CHBDC (2000) contains clauses for the strength evaluation of existing bridges; the clauses are based on the concept of a target reliability index that can change with (a) system behaviour, (b) component behaviour, and (c) the level of inspection. The system behaviour relates to the effect of the failure of a component to the failure of the whole structure; the component behaviour corre- sponds to the ductility of its failure; and the inspection level refers to the degree of confidence in the inspection process in determining the actual condition of the bridge and its components. In the paper, the effect of the three factors on the target reliability index is explained with the help of examples. The examples illustrate that depending upon the system and element behaviour and confidence on inspection, there can be a significant difference between two useable live load capacities for the same structure. If the condition of a component of a structure were determined with the help of sensors in an SHM system, the degree of confidence in the determination would be greater than in any visual inspection, with the consequence that the evaluator of the component would be able to utilize a larger portion of its live load capacity. The concept of the target reliability index changing with the inspection level does not exist in design of new structures.Yet,it can be appreciated that if the designer of a structure were confident that the condition of the structure and the load that it receives would be determined continually and accurately by an SHM system, he/she can afford to be less conservative, thus reducing the capital cost of the structure. Drawing upon the comparison between a fully instrumented aerospace structure and a civil structure without any sensor to report on its health, it can be stated confidently that the designers of civil structures are risk-averse to an extent because of the absence of information about the field performance of these structures.

4 CIVIL ENGINEERING EDUCATION

Both industry and university participants in the ISIS Canada network recognized that the production of leading edge research results would not find widespread usage unless the innovations and ideas 35 were integrated into the engineering curriculum. ISIS Canada therefore has taken a unique step for a research organization; it has adopted education as a primary element of its mandate. To facilitate its education mandate, ISIS Canada has developed a series of Educational Modules on both FRPs and SHM technologies for use in undergraduate engineering curricula. Teaching resources, consisting of lecture notes, presentations, worked examples, case studies, and sample assignments and laboratories, have been developed. Recently, targeted educational materials have also been developed for use within engineering technology programs at the technical college level for the training of engineering technicians and technologists. Concurrent with the development of student-oriented education modules, ISIS has been developing practice-oriented manuals for use by professional engineers. The manuals were prepared for two reasons. The first is to provide practicing engineers with detailed guidelines regarding the use of SHM technology and new materials in the design and construction of civil engineering structures. The second is to provide a practical set documents completed with guide- lines and background reference to supplement developing national Codes and hence facilitate and accelerate adoption of these new ideas in practice. While this approach has met with success in its objective of increasing awareness, the authors are proposing that a more fundamental change in curriculum is required. This change is not simply the production of course material or the introduction of a new course. A fundamental change in the integration of the various aspects of civil engineering is required throughout the curriculum in general. The paper discusses the authors’ thoughts on developing an integrated approach to infrastructure engineering. The traditional education approach and the proposed performance- oriented engineering approach are compared. With the aid of a case study the performance-oriented engineering approach is illustrated in detail. Finally, specific examples of education curriculum changes are provided.

4.1 Traditional education approach The two cornerstones of the traditional engineering education approach are engineering analy- sis and engineering design. Engineering analysis includes topics such as mechanics, determinate and indeterminate structures, linear and non-linear analysis and numerical modeling. Engineer- ing design includes such topics as material properties and behaviour, safety and codes, sectional resistance, and selection of component geometry. A startling realty of many curriculums is that these two cornerstones are generally taught in isolation of each other. Design education still includes many simplifying assumptions in behaviour ignoring many secondary effects or including them by approximation or simply coefficients. While the resistance of a structural component in ultimate limit state may be highly non-linear, the determination of load effects is often based on approximate linear analysis and often ignores a detailed consideration of load paths. While some effects, such as cracking of concrete, are stressed in design detailing considerations they are ignored or simplified in analysis. A third component of structural engineering education is construction, including project man- agement, which often receives very little attention in undergraduate engineering education. Even less frequently addressed are issues of service life maintenance and management of structures. Figure 2 illustrates these three subject areas with very little linkage (emphasized by the use of dashed connecting lines). The process is generally one-directional with very little feedback.

4.2 Proposed education model Currently structural health monitoring is normally discussed in the literature as a technology in its own right. Its separation from main stream infrastructure engineering is reinforced by the majority

Figure 2. Representation of traditional education methodology and linkages.

36 of forums for discussion and for publishing literature being specialty conferences and journals. There is a real risk therefore that, if SHM was introduced into the current curriculum, it would simply be another box in Figure 2 with less or even no linkage to the other elements. If this were to unfold then SHM would remain an interesting but little-used research idea. The authors propose that for SHM to achieve wide-spread usage a fundamental change in our education approach must occur. Indeed, the change is not simply about redesigning education curriculum or philosophy but about changing the way future engineers will practice engineering as it relates to civil infrastructure. Even without SHM, it can be readily acknowledged that the model shown in Figure 2 is inappro- priate. Firstly, a management activity box should be added to reflect the importance of the life-cycle engineering of infrastructure (i.e. the majority of money and effort often does not end with con- struction but comes throughout the remaining life of the structure). Secondly, these activities need to be integrated with strong linkages and feedback between all components. The education system needs to teach and reinforce this integrated infrastructure engineering approach at all levels of its curriculum. This is not a new concept and many progressive educators have also extolled the need for such linkages. However, it is not enough to simply state this need, particularly if the practice of engineering remains unchanged. What is required is a focal point for this integration. While SHM may seem to be a likely candidate for the focal point, it is in reality another tool or component for the infrastructure engineer. The authors propose that the proper focal point is performance. Performance in this sense needs to be considered in its broadest possible context. It includes qualitative or quantitative measures to assess the following issues:

• Determine actual behaviour and compare to assumptions used in analysis. • Determine actual loading and compare to design loading. • Evaluate durability of structure to environmental, operational and load conditions. • Assess secondary design issues such as temperature effects, joints, movements, vibrations. • Identify deterioration or damage of the structure and its components. • Evaluate safety and reliability of structure at any stage of its service life. • Evaluate remaining service life. • Assess the impact of maintenance or rehabilitation on future performance.

Whereas design and analysis are very abstract and rely on idealizations, performance is very structure specific and relies on the actual field data and assessments. SHM then becomes a vital component to obtain the field information necessary to make proper performance assessment. The model is illustrated in Figure 3. It may be argued that current infrastructure practices do have a focus on performance. This may be true for infrastructure managers but it is not fully integrated in the other stages of design. Design is often a task that ends with the production tender documents. Very few design engineers are involved in field inspections and assessment. The experience from previous poor performance is generally communicated to the designer via generic code criteria and standards. For specific structures, once the design is complete, the designer normally has no responsibility for its future performance. In many cases, construction activities are also not linked with performance. Individuals involved in construction supervision and quality control often do not see first-hand the impact of these choices during the service life of the structure. This is not merely an issue of practice but a philosophy that is reinforced by the education curriculum. Analysis is viewed as the most sophisticated of tasks yet is taught in isolation from measuring actual structural response. Design is the next sophisticated task but often relies on rigid code rules rather than reinforcing by observation of failures. The education in construction is normally more focused on scheduling and production rather performance. Field inspection and evaluation is often not even taught in engineering curricula but left to technical schools to train inspectors. 37 Figure 3. Proposed education model and linkages.

The student engineer must be taught that the prime objective of infrastructure engineering is performance and that the individual tasks shown in Figure 3 are only successful if the performance is successful. The message cannot be fully taught in the classroom and therefore must be reinforced by laboratory and field work. The paper contains a detailed discussion of a long-term field project that the authors have been involved in for over 12 years which illustrates the integrated approach necessary to assess performance. The project used both SHM data and laboratory work to establish that fatigue cracking is the key performance characteristic of an innovative concrete bridge deck system.The performance was assessed on component and superstructure level and compared against theoretical analysis and design assumptions. Experimental work was necessary to supplement an understanding of the SHM data being collected. A rigorous theoretical approach is now being developed to establish a better performance diagram for fatigue monitoring. Based on the performance evaluation of this and other similar structures, changes in design and construction details were implemented in future projects and eventually into code documents. The SHM system was constantly evolving as the knowledge and the understanding of the performance issues and the key response parameters evolve. Throughout all this activity, understanding and improvement of performance was the prime objective and was greatly aided by SHM activities. The paper also discusses the impact of design changes on performance. In response to observed issues in the field, design details and criteria can be changed to address that specific concern. A performance-oriented engineering approach would force us to consider the impact of these design changes not just on the specific issue of concern but on all performance issue for the structure. These issues are discussed with specific reference to the field case study. In one scenario it is shown that changing the design to improve one aspect of performance (i.e. cracking) may actually reduce another (i.e. fatigue life). In the second scenario, the danger of combining new technologies with full investigation is highlighted. It is shown that two distinct technologies that have been designed to address the same general performance issue of corrosion of steel-reinforced bridge decks may actual create new performance issues if combined. Both scenarios are meant to illustrate the importance of developing a performance-oriented rather than design-oriented engineering approach to technology and infrastructure improvement.

4.3 Changes to undergraduate curriculum The paper also presents some specific suggestions for changes to the undergraduate curriculum will be discussed. Perhaps the most important need is to re-emphasize the importance of physical observation in developing an understanding of structural engineering. With the maturation of numerical methods such as finite element modeling and the easy access to user-friendly commercial packages, many educators have focused on theoretical approaches to analysis. In addition, electronic representative of components and behaviour made convenient by graphics software has replaced physical models. Students have lost the inherent ‘feel’for the structure; but, more importantly do not fully understand 38 that the numerical modeling, albeit sophisticated, is still just a theoretical representation of actual behaviour based on simplified assumptions. It is the actual response they need to understand not simply the theory behind the models. The linkage with SHM activities is very strong on this point. Recorded field data contains all the complexities of response, including issues that are normally classified as secondary in design or ignored in analysis. Interpretation of field data therefore requires a higher level of understanding than is often found in design. In fact, if an engineer is attempting to design an SHM system with compensated gauges to eliminate unwanted effects such as temperature in the recorded response, then the level of understanding needs to be even higher. The authors propose therefore that performance-oriented education needs to include significant laboratory work and, where possible, field work on instrumented structures. To illustrate some possible specific activities as well as maintaining a connection with the case study of section 4, this lab work should include such activities such as measuring response a strain profile throughout the girder depth and then using that information to determine neutral axis location and composite behaviour. Ideally the experiment should be designed such that gauge readings include the effects of temperature and the girder response includes both torsion and flexural components. The latter may only be possible if access to field structures is provided. Challenging the students then designing a strain monitoring system to isolate or compensate for each phenomenon should be part of the assignment. In addition, students should be given an assignment to calculate the magnitude and configuration of the applied load from the recorded values. All these activities not only develop tools necessary for interpreting SHM data, but also reinforce and supplement the traditional education material in mechanics, analysis and design courses. Laboratory projects focusing on the effects of fatigue and the concept of cumulative damage should also be developed. Students should measure such parameters such as crack width, deflection and strain of a component and develop a curve of cumulative damage versus number of load cycles. In the time span of a regular undergraduate course, the stress range may have to be high to develop significant damage in a low number of cycles. Project groups can be formed where each group performs a different number of fatigue load cycles and then tests the component to failure. This will enable the students to make an assessment of whether the damage they observed (change in condition) affected the structural capacity (change in safety). Considering the two issues together rather than separate provides an understanding of performance. If possible the experiments should be designed in such a way that the initial defection or crack width meets prescribed design limits but after some level of cumulative damage, the prescribed limits are exceeded. This will reinforce serviceability performance as well as enable the students to understand that it is not just the initial design that must meet specific criteria but that a structure must continue to perform throughout its service life. The authors do not believe that structural engineering students need extensive training in fun- damental sensory and sensor system technology. In their education, they should be exposed to many sensor types but their undergraduate level understanding should be limited to basic physics principles, such as change in resistance or a change in wave length can be used to determine strain. The focus should remain on data interpretation and performance assessment. Within the limit of this paper, only a few example activities could be provided. But they do illustrate the nature of the activities required and the need for physical not virtual experience. The proposed changes are also consistent with recent education initiatives in project-based learning and in teaching outside the classroom environment.

5 CONCLUSION

The authors propose that a change in education to a performance-oriented engineering curriculum is necessary. An emphasis on learning in the physical environment of the lab and the field is preferred over the virtual environment of the classroom. In both instances, effective use of SHM will be required to properly understand and assess performance.

39

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Practical implementation of probability based assessment methods for bridges

I. Enevoldsen Ramboll, Virum, Denmark

ABSTRACT: This paper presents the practical implementation of probability based assessment methods to road and rail bridge structures. For the structures analysed, modelling of the critical limit states is described as are the statistical techniques employed in modelling the loads and resistances. The overall aim of the analysis was to achieve higher load ratings for the structures considered than those which resulted from traditional deterministic analysis. Ultimately the benefits to bridge owners/managers of performing probabilistic assessments are apparent from the results, which consistently provided (a) higher load ratings for the bridges considered and (b) considerable financial savings (both direct and indirect) through the avoidance of unnecessary repair/replacement of serviceable highway structures. Examples presented in the paper demonstrate direct monetary savings in excess of $30 ml (USD).

1 INTRODUCTION

A common problem among bridge owners/managers is the need to reduce spending whilst attempt- ing to operate and maintain an increasingly ageing bridge stock which is subject to a loading intensity for which, in many cases, it was not designed. The problem is compounded by the ever increasing trend in motorway traffic frequency which was seen to double in the decade 1992 – 2002 and by the debate regarding the need to increase legal weight limits for trucks and trains and/or to provide special routes/networks which they can use. By way of example, in Denmark, the Danish Roads Directorate (DRD) have established a so called Blue Road Network which comprises roads with no bridges having a class less than 100. A classification of 100 implies that the structures along the network are capable of simultaneously carrying a 100 tonne and a 50 tonne vehicle in normal passage (i.e. no restrictions on vehicle positions on structure, full dynamic factor applied to vehicles whose silhouettes are specified). As the blue road network includes all motorways and many other major roads it ties together transport routes throughout all of Denmark. The police and road/bridge administrators use a map of the blue road network when preparing special weight permits while haulage contractors use it when planning transports and selecting routes. The wider provision of such routes internationally would clearly simplify the process of facili- tating heavier trucks and trains to transport freight. This is a pressing issue which must clearly be addressed when considered within the context the European Unions (EU) own projections for a doubling in the volume of freight transported by road in the period 2000 – 2030. The problem with such projections of growth is how it will be accommodated on an already congested network. One partial solution could lie in permitting larger transports (i.e. longer and heavier trucks and trains) to operate on road/rail networks. However the problem with such a solution lies in the verification that structures along such routes have adequate capacity for these super transports. Given that in the vast majority of cases these structures were deterministically designed for loads far less than those which we would call upon them to carry in the future. Then it is clear that deterministic assessments would prescribe that extensive rehabilitation/strengthening projects would be needed for any pro- posed increase in allowable loads. However, the economic costs associated with such a continent 41 wide process would undoubtedly prove prohibitive and as such a departure from deterministic thinking/methods is required if the proposed partial solution is to be realized or even considered. Bearing these factors in mind the past decade has seen increased interest by bridge owners and managers in the use of probabilistic methods for the assessment of their bridges specifically within the context of documentation of higher load carrying capacities such that unnecessary repairs/rehabilitations can be avoided. Employed once a deterministic assessment of a structure has rendered a repair/rehabilitate/replace now scenario, the methods have been demonstrated to provide significant cost savings where the required safety of the structure at higher load levels can be demonstrated by probabilistic methods. This paper presents a number of case studies where probabilistic methods have been employed to validate the safety of structures for higher load carrying capacities than documented by deterministic assessment. The direct result of which has been the avoidance of unnecessary repair/rehabilitation/replacement of a serviceable structure and their inclusion on the network for heavy transports. The majority of cases are selected from Danish experience, as it is the Danish Roads Directorate (DRD), which leads the way amongst European agencies in the use of proba- bilistic methods as an assessment and management tool. This fact is evidenced by their publication in 2003 of a formal guideline for probability-based assessment of the load carrying capacity of highway structures (DRD 2003). However, the methods are receiving increased acceptance inter- nationally and in this regard the main paper also presents examples of the application of the methodology to a sample of bridges in Sweden.

2 PROBABILITY BASED ASSESSMENT

The basis of probability-based assessment of existing bridges lies in evaluation of the structures individualised safety. In this regard, the basic principles of the approach are no different to those employed in the derivation of design codes. The advantage however, lies in the fact that in adopting a structure specific approach the methods does not to the same degree suffer from the inherent conservatism of design codes which must necessarily generalise in order to be widely applicable. The methodology proposed for practical application of probability-based methods to safety assess- ment of existing bridges in Denmark, is taken from the guideline for probability-based assessment of the load carrying capacity of highway bridges which has been produced by the Danish Roads Directorate (DRD 2003) with Ramboll as consultants.

3 EXAMPLES OF PROBABILISTIC ASSESSMENT OF EXISTING BRIDGES

In the following, examples of the application of probability based assessment to existing bridges are presented. The cases presented represent both road and rail bridges. The examples are significant in that they show that (a) in many cases probability based methods can be employed to demonstrate higher structure load carrying capacities than deterministic methods thereby avoiding unnecessary repair/replacement of serviceable structures and that (b) in cases where sufficient capacity cannot be demonstrated that probability based methods can be employed as an efficient management tool to optimise rehabilitation.

3.1 Vilsund Bridge, Denmark The Vilsund Bridge, illustrated in Figure 1, is a 381 m long steel bridge from 1939 consisting of 5 ordinary 67.8 m spans and a 34 m bascule span. The two lane road bridge has a width of 8.6 m with a concrete slab deck supported by a steel girder system with cross girders for every 5.58 m. The Danish Road Directorate decided that it would be desirable to allow trucks with a total weight of 100 metric tons to pass the Vilsund Bridge instead of a maximum truck weight of 50 metric tons. (Trucks with a weight above 50 metric tons must take a 150 km long detour if the Vilsund Bridge 42 Figure 1. Vilsund Bridge, Denmark. 5 ordinary spans and 1 bascule span. cannot be passed). Analyses of the bridge were performed according to the general approach for classification of existing bridges in Denmark with Class 50 as a result, which corresponds to a situation where two Class 50 trucks with an approximate weight of 50 metric tons, are passing each other on the bridge in the most critical situation for the considered structural element. The results of the deterministic analyses showed that the critical structural members are the main cross girders supporting the concrete deck and some of the steel truss members in the ordinary spans. A rough estimate of strengthening cost of the Vilsund Bridge to Class 100 was found to be $4 million. Instead of performing the strengthening project, it was suggested to perform a bridge specific reliability analysis of the critical elements in order to reduce or eliminate the costs of the strengthening project. The reliability evaluations were split into probabilistic-based evaluation of (a) the main cross girders and (b) of the critical steel truss members in the main structure. The results of the probabilistic assessment showed that the structure could be classified in Class 100, and as such the proposed strengthening project was avoided without compromising on the safety of the bridge deck.

3.2 Storstroem Bridge, Denmark The 3.2 km long Storstroem Bridge connects the Danish Island of Zealand (on which Copenhagen is located) with the southern Danish islands of Falster and Lolland. The contract for the building of the bridge was given to the British company Dormann, Long & Co., who also fabricated the main steel structure. The contract was awarded to a British company as a political move to offset the significant trade deficit which had developed between the UK and Denmark at this time due to Danish pork exports. Until 1985 when the Faroe Bridge opened, Storstroem Bridge was the only fixed connection between Zealand and the southern Danish Islands. The Faroe Bridge carries only cars. Today the Storstroem Bridge carries only local traffic with an average annual daily traffic (AADT) of about 8000 vehicles. The main deck slab of the 3.2 km long Storstroem Bridge has suffered serious deterioration to both the concrete and reinforcement. Replacement of the bridge would be extremely costly especially when considered in connection with the possibility of the construction of the Femern Bridge at some point in the future. The bridge only carries local traffic, bicycles and pedestrians and trains. If the Femern Bridge is constructed, then the Storstroem Bridge may need to be extended from one to two train tracks. Thus, the DRD would like to postpone any decision on a strategy for the Storstroem Bridge until a decision about the Femern crossing is made. However, at the same 43 Figure 2. Storstroem bridge.

Table 1. Results of deterministic and probabilistic assessment; O’Connor et al (2004).

Load Combination Self Weight + KL10 Live Load

Deterministic plastic load carrying capacity 61% −13 Probabilistic Assessment: No deterioration pf = 2.94 x 10 β = 7.20 −7 Probabilistic Assessment: Stochastic modelling of pf = 6.92 x 10 β = 4.83 deterioration according to inspections results time the DRD must ensure that the structure has sufficient structural safety for both vehicles and pedestrians at all times. A deterministic assessment of the deck slab using plastic response models produced a maximum load factor of 0.61 for combined dead and live load, Table 1. This implied that the slab was incapable of sustaining the applied load. The recommendation would therefore involve costly rehabilitation of the structure. The result of a probabilistic assessment was a reliability index of 7.2 (greater than the required 4.75 (DRD 2003)) without taking deterioration into account thereby verifying that the applied load could be sustained. To obtain an accurate measure of the current deterioration a number of inspections were made in 2002 at critical locations, previously determined as the area close to the transverse joints. From the inspections a conservative deterioration model was made. Analyses using this model gave a reliability index of 4.8 in 2002. The results of the assessment of the structure performed by Ramboll are listed in Table 1, where β represents the safety index. Probabilistic assessment using a prediction of the future deterioration of the road slab identified that the safety of the road slab could not be maintained for heavy trucks unless the transverse joints with the most severe deterioration were continuously repaired. If this was done the remaining lifetime of the road slab was estimated to be at least 10 years. The repair of the transverse joints is continuing so that the most deteriorated joints are being repaired. The cost of these local repairs are insignificant to any major repair or indeed compared to replacement of the road slab.

3.3 Bergeforsen railway bridge, sweden Bergeforsen railway bridge is a single track bridge which was constructed in 1923. The bridge is sit- uated on Swedish rail’s Sundsvall-Härnösand line, approximately 350 km north of Stockholm. The superstructure of the bridge is composed of riveted trusses with spans of 42 + 84 + 42 = 168 m as illustrated in Figure 3. Simply supported side approach spans of 22.5 m and 11.6 m give a total bridge length of 202.1 m. The superstructure of 42 + 84 + 42 m is supported at 4 longitudinal locations, 1-fixed + 3-roller, which essentially results in the bridge working as a continuous beam over 3 spans. Deterministic assessment of the bridge was performed according to the Swedish Assessment code, BVH 583.11, using the train load model BV-3 (i.e. a trainload model with 25 tonne axles and 44 Figure 3. Bergeforsen railway bridge.

8 tonne/m line load). Structural analysis was performed in two phases using a 3d finite element (FE) beam model of the bridge constructed using the FE software LUSAS. The first phase of the analysis focused on the Serviceability (SLS) and Ultimate Limit States (ULS) while the second phase con- sidered the Fatigue Limit State (FLS) using Rain Flow counting employing actual train silhouettes. The results of the deterministic assessment demonstrated that while the structure had sufficient capacity with respect to the Serviceability and Fatigue Limit States, that at a number of locations the structure failed to demonstrate the necessary Ultimate Limit State (ULS) capacity and as such some form of strengthening was required. As this conclusion would prove extremely costly, rather than immediately accepting the consequences of a traditional decision making process, in which this failure in a deterministic assessment would lead to a requirement for strengthening/ rehabilitation of the bridge, Banverket’s management strategy focused on an alternative decision process which would permit a probabilistic assessment of the structure to be performed to see if the necessary capacity could be demonstrated. Where sufficient capacity could not be demonstrated by the probabilistic approach it was intended that a probability based decision process would be implemented which would provide a rigorous/robust decision methodology and consequently provide information on the optimal maintenance strategy.The legal justification for this process was obtained from BVH 583.11 where it is stated that Banverket permits probability-based assessments. In the majority of cases analyzed the probabilistic assessments were able to demonstrate sufficient capacity, i.e. β>4.8 for the elements and joints considered. However, in the case of 2 joints, probabilistic assessment could not succeed in avoiding some level of strengthening. Figure 4 illustrates one of the joints, circled, for which the calculated β = 4.51, which, as it was < 4.8 could not be deemed to satisfy the minimum safety requirements. Two sample proposals for strengthening the joint were considered, both of which involved removal of certain rivet groups from the joint and their replacement with high strength bolts. Option A, labelled in Figure 4(c) considered replacement of the central rivets, encircled in Fig- ure 4(c) and labelled A, with 27 mm diameter bolts, while option B, encircled in Figure 4(c) and labelled B, considered replacement of the bottom two rows of rivets with bolts. In both cases the probabilistic model which was established for safety analysis of the joint was used to recompute the revised structural safety on the basis of the proposed joint strengthening methodology. The results of these re-assessments were, Option A β = 6.05 > 4.8 and Option B β = 7.80 > 4.8. The significance of these results lies in the way in which they demonstrate the ability of the prob- abilistic model to assess the relative efficiency of proposed repair methods and as such to be used not only in safety assessment of structures to demonstrate increased load carrying capacity, relative 45 Figure 4. (a), (b) Connection for which β<4.8, (b) Possible alternative repair scenarios.

Table 2. Results of deterministic and probabilistic assessment; O’Connor et al (2004).

Phase 1 Deterministic Phase 2 Advanced Deterministic Phase 3 Probability Based Assessment ($USD) Assessment ($USD) Assessment ($USD)

Consultant Fee $0.1 ml $0.2 ml $0.28 ml Contractor Fee $3.2 ml $1.1 ml $0.47 ml Project Management $0.3 ml $0.2 ml $0.1 ml Total Cost $3.6 ml $1.5 ml $0.85 ml to deterministic assessment, but also to be used as a tool in optimising maintenance/rehabilitation planning for structures which are determined to require some level of repair/rehabilitation to satisfy specified minimum safety criteria. The results of the assessment represented a considerable eco- nomic saving for the bridge owner through the avoidance of unnecessary repairs and the optimisation of repairs where they were indeed required. The extent of this saving (in direct costs only) is evidenced in Table 2. Here associated costs for (a) Consultants Fees, (b) Contractors Work and (c) Project Management are presented for the results from 3 phases of the project (1) DeterministicAssessment, (2)Advanced DeterministicAssessment incorporating updated structure models and rain flow analysis for fatigue 46 Table 3. Examples of DRD savings from probability based assessments.

Result of Deterministic Probability-based Cost Saving Bridge Analysis assessment $(USD)

Vilsund Max W = 40 t Max W = 100 t 4,000,000 Skovdiget Lifetime ∼ 0 years Lifetime > 15 years 13,200,000 Storstroem Lifetime ∼ 0 years Lifetime > 10 years 2,500,000 Klovtofte Max W = 50 t Max W = 100 t 2,200,000 407-0028 Max W = 60 t Max W = 150 t 175,000 30-0124 Max W = 45 t Max W = 100 t 175,000 Norreso Max W = 50 t Max W = 100 t 600,000 Rødbyhavn Max W = 70 t Max W = 100 t 600,000 Åkalve Bro Max W = 80 t Max W = 100 t 1,750,000 Nystedvej Bro Max W = 80 t Max W = 100 t 2,200,000 Avdebo Bro Max W = 50 t Max W = 100 t 3,300,000 TOTAL >30,000,000 assessment and (3) Probability based assessment performed at critical locations as determined in Phase (2). It is noted that the Consultants Fee in Phase (2) includes that of Phase (1) and that that in Phase (3) includes those of Phases (1 + 2). It is also significant to point out that these are direct costs only and do not consider indirect costs associated with user- & freight delays which would have been associated with any major rehabilitation work to be performed on the structure.

3.4 Additional Cases from Denmark The Danish Roads Directorate pursues probability based assessment as a matter of course for all structures which have failed a deterministic assessment. As evidenced the results of this policy have provided significant savings in both the direct and indirect costs associated with bridge rehabilitation or replacement. In addition to the cases presented, Table 3 lists the direct monetary benefits, in excess of $30,000,000 (USD) accrued in some recent cases where probability based assessments have been employed. Photographs of some of the structures listed in Table 3 are presented in Figure 5. In Table 3 the documented load carrying capacities resulting from the deterministic and probabilistic methods are also presented. In this regard, and with reference to Figure 5(a) deterministic assessment yielded a capacity W = 50 t, which inferred that the bridge was capable of simultaneously carrying 2 No. standard 50 t trucks side by side in normal passage conditions (i.e. no restrictions on vehicle positions on structure, full dynamic factor applied to vehicles) whereas probabilistic assessment W = 100 t documented that the structure could safely carry a 100 t and a 50 t truck side by side in normal passage conditions. The wide applicability of the probabilistic methodology is apparent from Figure 5, where a wide range of structural forms/geometries to which the method has been successfully applied is presented.

4 CONCLUSIONS

This paper presents the practical implementation of probability based methods to the assessment of bridge structures with the aim of demonstrating higher load carrying capacities. In all cases presented deterministic assessments concluded that strengthening or rehabilitation of the structures was required if the bridges were to be kept in service. However, more thorough probabilistic analysis of the structures at their critical limit states demonstrated, in most cases, that they did indeed have the load carrying capacity to remain in service without strengthening works. In the case where probabilistic assessment could not document adequate capacity it was demonstrated how the methodology could be employed to optimise rehabilitation of the structure. In all cases the results represented significant savings for the bridge owners both in terms of the direct replacement 47 Figure 5. Practical Examples of Probability Based Assessment of DRD Structures. cost and with respect to indirect costs, i.e. user delay costs, which would have been incurred during rehabilitation of the structures. It is important to stress that at no stage was the safety of the structures compromised, rather a more realistic and bridge specific safety assessment was performed which was shown to be free from the generalisations of deterministic codes.

REFERENCES

BVH 583.11, (2000), Bärighetsberäkning av järnvägsbroar, Handbok BVH 583.11 Banverket, 2000 (In Swedish “Assessment of railway bridges”) Danish Roads Directorate, DRD, (2003). Guideline for reliability based classification of the load carrying capacity of existing bridges. Available from Danish Roads Directorate’s website www.vd.dk 48 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Bridge monitoring in Japan: The needs and strategies

Y. Fujino & D.M. Siringoringo Department of Civil Engineering, The University of Tokyo, Tokyo, Japan

1 INTRODUCTION

This paper gives an overview of development of bridge monitoring in Japan. The paper consists of three parts: 1) backgrounds of bridge monitoring development in Japan, 2) concept of monitoring, and 3) strategies implemented in bridge monitoring.

2 BACKGROUNDS OF BRIDGE MONITORING IN JAPAN

The need for monitoring in Japan was originally influenced by its geographical conditions. Due to the facts that Japan is prone to natural disaster and has severe environment for deterioration, monitoring of environment and loading conditions with regard to natural hazard have been realized for several decades. Natural disasters such as earthquakes and typhoon are some of the major concerns for civil engineering construction and maintenance. In its report (Japan MLIT 2005), the Japan Ministry of Land, Infrastructure andTransport (MLIT) has listed natural disaster as top factors that are threatening the sustainability of infrastructure. From 1970 to 2004, the total infrastructure loss due to natural disaster is approximately US$ 165 billion, or 15% of the world’s total infrastructure losses caused by natural disasters. In 1995, Kobe earthquake alone killed more than 6000 people and heavily damaged major infrastructure systems. The earthquake caused extensive damage to highway bridges in the national highway and local expressways network. To anticipate the similar scale of earthquake in the future, bridge seismic design code was later revised. Constructions of new bridges are subjected to the new code, while the existing bridges are undergoing seismic assessment and retrofit program to meet the demand required in the new code. In the context of a bridge retrofit and seismic assessment program, structural health mon- itoring (SHM) plays an important role. Many bridges in Japan especially the long-span ones are instrumented with permanent sensors. This instrumentation provides seismic or/and ambi- ent responses that are essential to obtain insight into the real behavior of bridges during different loading conditions. Another geographical aspect that influences Japan infrastructure, and bridges in particular, is the country’s topography. As an archipelago country, Japan consists of mountainous islands with pop- ulation mainly concentrated near the seashore. Therefore, transportation infrastructures including many bridges are built near the coast line, and some are crossing the channel, such as the world longest Akashi-Kaikyo Bridge. Due to this condition, some bridges are situated at severe environ- mental condition and subjected to deterioration caused by high chloride intrusion and humidity. Consequently, maintenance and monitoring system must be employed to control the environmental effects and to prevent further deterioration. In the last decade, bridge monitoring has extended its function as an instrument for an efficient stock management. Amid the accumulation of bridge stock and the concentrated construction years, many bridges in Japan are expected to have serious deterioration problems within the next decade. The rush to expand infrastructure during the rapid economy growth has led to cost-cutting by sometimes using minimal materials needed for the structures’ expected performance. Bridges 49 provide an excellent example of the problem. Deterioration is an issue for many highway bridges in Japan, which were mainly built around 1970s. Awareness of the current condition of bridges has emerged recently, triggered by the findings of steel member fractures in the Kisogawa Ohashi Bridge, Mie Prefecture, in June 2007 and in the Honjo Ohashi Bridge, Akita Prefecture, in August 2007. The sudden collapse of the I-35W interstate bridge in Minnesota, United States that followed later in August 2007 has only intensified the concern, considering similarity of the two bridges in Japan with the one that collapse. Many believe these two incidents are the tip of the iceberg of the problem. Adding to the problem of aging bridges is the drastic increase of traffic density.The high intensity and frequency of loading generated by this high traffic volume generates many problems in bridges. The most typical problem is fatigue damage in concrete slabs and steel welded girders. The decrease in maintenance expenditure is expected to continue in the following years due to the budget constraints. Japanese bridge experts now begin to call government’s attention to preventive maintenance of national bridge network. This includes the reevaluation and overhaul of the existing bridge inspection system.

3 CONCEPT OF MONITORING FOR RISK REDUCTION

The second part of the paper describes a concept of bridge monitoring as an essential part of risk reduction. To improve bridge safety, monitoring technologies for risk and vulnerability are implemented. Risk is defined as the function of two quantities: hazard and structure vulnerability. Hazard, defined as the probability of occurrence of unexpected events that may endanger the structure, is spatial and temporal dependent. Conventional practice assumes that the structural vulnerability is time-invariant, and a function of structural configuration such as system redundancy, ductility, materials, and quality of construction, all of which are designed at the construction phase. In reality, however, structural vulnerability is also temporal dependent in that it may differ from time to time due to deterioration, aging or changing in environmental and loading conditions. The methodologies for performing hazard analysis and monitoring have been well-developed and are well-established. On the other hand, approaches to quantify a time-variant structural vul- nerability are still in under research and development. An example of conventional approach to structural vulnerability assessment is the fragility curves. While useful, the empirical fragility curve is sometimes inadequate due to its limited events and observed damages. Therefore, the analytical- based fragility curve is developed for a more general use of vulnerability estimation. This approach also has some significant limitations. Since the curve is computed based on a normal distribution assumption, one can never be certain about the exact position of the studied structure in the given distribution. To minimize uncertainties, field monitoring is the feasible option. Using monitoring the actual structure condition can be assessed and therefore its vulnerability can be quantified. To ensure the benefits of monitoring to infrastructure’s stake holders, rationality behind invest- ment made on structural monitoring and its subsequent effects must be clearly presented. A recent study involving the Japan Railway Corporation shows that investment in monitoring for disaster prevention has increased the safety of infrastructure system. It is shown that the benefits of disaster prevention investment in infrastructure do not come immediately but last for longer time. This goes to show that negative impacts may not come immediately even if we do not invest in any disaster prevention measures. However, in the long run, we may suffer from greater consequences. Rapid advances in sensor technology, communication system, and information technology have potential benefits in managing and maintaining an infrastructure system. In recent years, innova- tions have led to the development of high-tech-based systems that are based on the principle of sens- ing the structural condition or loading, transferring the measured quantities through communication system to support a knowledgeable decision making. With such advanced systems, large amount of infrastructure can be managed more effectively and potential losses or failures can be minimized. The use of advanced sensor technologies for monitoring is a very important issue for Japan. Considering the increasing number of aged population and decline in the birth rate, it is expected 50 that Japan will face a serious labor shortage in the near future. This problem will also affect the strategy in infrastructure management. In the future, programs that depend extensively on manpower should be avoided, and instead shifted to those that are based on more advanced technologies. There are two fronts of research in monitoring that are currently active, namely sensing technolo- gies, and structural diagnostic/prognostic. The former typically deals with sensor types and systems, data acquisition, data processing, communication, management and storage. The latter deals with data analysis and interpretation, system identification, local and global diagnostic, defect/damage detection and remaining life prediction, all of which are represented by performance indicators. There are two aspects need to be considered when choosing sensor for monitoring civil infrastructure:

• Scalability. The civil infrastructure scale is considered on the mesoscale level. Because of the spatial diversity and sheer numbers of structures, the use of a too-expensive and sophisti- cated microsensing system should be avoided. At the same time, however, the uniqueness of information of each structure system should be retained. • Durability. Civil infrastructures are designed to last for a long time, typically over 50 years. During the service life, hazardous events may not occur very frequently, but nevertheless the monitoring system must be always up and running. This indicates the need for not only reliable but also durable sensors.

On the structural diagnostic/prognostic front, the methodologies can be categorized according to the scope and type of analysis. The scope of analysis consists of: (i) microscopic monitoring, where damage detection and localization are of main interest; and (ii) macroscopic monitoring, where holistic structural integrity and its comparison are the main focus. While the microscopic monitoring is the conventional mainstream of structural health monitoring and advancing steadily, macroscopic monitoring is recently attracting interest especially from practical point of view to connect health monitoring and existing inspection methodology.

4 BRIDGE MONITORING: STRATEGIES AND EXAMPLES

The third part of the paper outlines strategies implemented for bridge monitoring in Japan. They are categorized into three main groups according to the purpose of monitoring that is for: natural hazard and environment condition, effective stock management, and failure prevention. Examples of bridge monitoring systems that implement these strategies and the lessons learned from monitoring experiences are presented.

4.1 Monitoring for natural disasters: Seismic and wind monitoring Monitoring for environment and natural disasters prevention are described with example of seismic monitoring and wind monitoring. In seismic monitoring, examples are the instrumentation of three long-span cable-supported bridges in Tokyo bay area. The bridges are: Yokohama-Bay Bridge, Rainbow Bridge, and Tsurumi-Fairway Bridge. The instrumentation system and sensor network are presented. Dense seismic measurement systems installed on these bridges are very useful to capture real behavior of the bridge during various levels of earthquake. The study shows that by employing a system identification technique, performance of dynamic characteristics of the bridge can be evaluated. Monitoring for wind-induced vibration is also presented with example of monitoring system employed at Hakucho Bridge. The bridge is situated at a windy area, so that wind load and structural behavior under various wind conditions are of main concern. Owing to detailed measurement and dense sensor deployment, modal characteristics of the bridge can be identified from ambient vibration data. 51 Even though bridge monitoring systems for seismic and wind hazard were initially intended to measure loading, studies presented in this paper indicate the usefulness of these measurements for evaluation of structural integrity.

4.2 Monitoring for stock management In order to realize an efficient stock management, monitoring should be focused on performance evaluation of the condition of the stock. Performance evaluation of a bridge is essential to assess functionality, predict the deterioration and the possible failure mode, update the performance prediction, and decide the future inspection of retrofit plan. For these purposes, new monitoring systems using advanced sensor technologies are expected to improve the conventional inspection procedures. In this section, examples of monitoring for stock management are briefly discussed. They are: a) continuous monitoring of a workhorse bridge; b) Non-contact monitoring system of Shinkansen viaduct by laser Doppler vibrometer; and c) routine inspections of railway and highway by intelligent monitoring system.

4.3 Monitoring for failure prevention Monitoring can also be directed to prevent structural failure. For this purpose, possible failure locations and circumstances under which the failure may occur should be identified a priori. When the failure mode of interest is identified or specified, monitoring system is implemented by selecting the appropriate sensor. Some of failure modes are localized, and so are the locations of sensor. Several such systems are developed and implemented for safety of railway operation. Two examples of monitoring to prevent bridge failure are presented that is prevention of failure due to unseating and the scour-induced collapse detection. Both systems works in such a way that when the failure indicators reach the maintenance limits an alarm is sent to the operator, so that further check and countermeasures can be taken.

52 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Bridging capacity innovations on cable-supported bridges

Y.J. Ge & H.F. Xiang Tongji University, Shanghai, China

ABSTRACT: After technical review of the longest cable-supported bridges, the current bridging capacity is summarized as single main span up to 2 km and 1 km together with total deck width about 40 m for suspension bridge and cable-stayed bridge, respectively. Bridging capacity innovations is then introduced with two double main span suspension bridges and one twin parallel deck cable- stayed bridge. The special emphasis for double main spans is placed on longitudinal stiffness design of central pylon, and either inversed Y-shapedpylon, in Jiangsu Taizhou Bridge, or I-shaped pylon with fixed connection of deck and central pylon, inAnhui Maanshan Bridge, can provide with proper longitudinal stiffness for central pylon. Two types of pylon and three kinds of deck section have been compared in earlier design stage for twin parallel deck cable-stayed bridge, NingboYongjiang Bridge, and the composite cross section with two diamond pylon has been finally selected with considering structural durability.

1 INTRODUCTION

Traditionally, bridging capacity is referring to longitudinal bridge span with the latest records of Akashi Kaikyo suspension bridge and Jiangsu Sutong cable-stayed bridge, and transversal deck width in the current practice about 40 m. Bridging capacity innovation is not invention in making span longer or deck wider, but a further or new development of traditional concepts to make things better in proper way. The concept of continuous multiple main span scheme for crossing longer water body is innovative, and the idea of several parallel bridge deck configuration for providing wider passage way is innovative too. Two suspension bridges with double main spans over 1 km have been launched across Yangtze River inTaizhou, Jiangsu Province, and Maanshan, Anhui Province. Among several under designed cable-stayed bridges with twin parallel decks, Ningbo Yongjiang Bridge has the longest central span of 468 m. Taking as these three examples, special structural characteristics and practical implications related to double main spans or twin parallel decks are revealed and discussed in the following sections.

2 DEVELOPMENT OF BRIDGING CAPACITY

2.1 Steel-chain suspension bridges Although ancient suspension bridges were built in iron chain in China long before the history of steel application in the 19th century, the construction of modern steel-chain suspension bridges around the world has experienced a considerable development for more than a century. It took about 54 years for the span length of suspension bridges to grow from 483 m of Brooklyn Bridge in 1883 to G. Washington Bridge and to 1,280 m of Golden Gate Bridge in 1937, and had an increase by a great factor of about 2.7. Although the further increase in the next 44 years from Golden Gate Bridge to Verrazano Bridge and to Humber Bridge of 1410 m in 1981 was only 10% or by a factor of 1.1, another factor of about 1.4 was realized in Akashi Kaikyo Bridge with a 1,991 m main span 53 Figure 1. 1080 m + 1080 m Double main-span suspension bridge (Anhui Maanshan Bridge). within 17 years in 1998. It is probably correct to conclude that modern suspension bridge practice has provided with a traditional three-span bridging configuration with single main span up to 2 km and total deck width of 40 m for six traffic lanes.

2.2 Modern cable-stayed bridges Cable-stayed bridges can be traced back to the 18th century, and many early suspension bridges were of hybrid suspension and cable-stayed construction, for example, Brooklyn Bridge in 1883. One of the first modern cable-stayed bridge is a concrete-decked cable-stayed bridge built in 1952 over the Donzere-Mondragon Canal in France, but it had little influence on later development. The steel-decked bridge, Stromsund Bridge in Sweden by Franz Dischinger in 1955, is therefore more often cited as the first modern cable-stayed bridge with a main span of 183m. It took about 31 years for the span length of cable-stayed bridges to increase to 465 m in Annacis Bridge in Canada in 1986, but in the last decade of the past century, the span length grew very rapidly, from 520 m of Skamsund Bridge in 1991, to 602m of Yangpu Bridge in 1993, then to 856 m of Normandy Bridge in 1995 and finally to 890 m of Tatara Bridge in 1999. Another big jump with about two hundred meters in span length has been realized in Jiangsu Sutong Bridge with the 1088 m length of main span in this year. Based on the bridging capacity figures, we may dare say that recent development of cable-stayed bridges has ensured the bridging capacity of a 1 km long main span and 40 m wide bridge deck for six traffic lanes.

3 DOUBLE MAIN SPAN SUSPENSION BRIDGES

3.1 Conceptual comparison of single and double main-spans In order to make the conceptual comparison of a double main-span suspension bridge and the corre- sponding single main-span structure, Anhui Maanshan Bridge has been taking as a typical model of a double main-span suspension bridge, with the span arrangement of 360 + 1080 + 1080 + 360 m and the sag to span ratio of main cable of 1/9, shown in Fig. 1. The corresponding comparison model is a traditional three-span suspension bridge of 720 + 2160 + 720 m, and has the same sag to span ratio of 1/9. With the same steel box deck, these two bridge models have been investigated with different schemes of pylon stiffness as follows: Scheme A-1: Single main-span structure with infinite stiffness pylons Scheme A-2: Single main-span structure with proper stiffness pylons Scheme B-1: Double main-span structure with infinite stiffness pylons Scheme B-2: Double main-span structure with proper stiffness side pylons and infinite stiffness center pylon Scheme B-3: Double main-span structure with proper stiffness pylons Having performed a dynamic finite-element analysis, the first and second natural frequencies of the structures have been extracted and compared for these five schemes in Table 1. The first and second frequencies of the double main-span scheme (B-1) are tremendously enhanced comparing with the single main-span scheme (A-1) under the condition with infinite stiffness pylons, and the increase factors are 2.51 and 1.27 in the first and second lateral bending frequencies, 1.18 and 1.08 in the first and second vertical frequencies, and 2.11 and 1.88 in the first and second torsional 54 Table 1. The first and second natural frequencies.

Lateral bending (Hz) Vertical bending (Hz) Torsional vibration (Hz)

Scheme First Second First Second First Second

A-1 0.0367 0.0768 0.1016 0.1466 0.1789 0.2045 A-2 0.0363 0.0765 0.0960 0.1283 0.1788 0.1998 B-1 0.0923 0.0973 0.1195 0.1590 0.3780 0.3836 B-2 0.0922 0.0971 0.1194 0.1553 0.3347 0.3610 B-3 0.0922 0.0969 0.0828 0.1196 0.2634 0.3355 frequencies. This is the most important reason why a double main-span suspension bridge is better in structural dynamic performance than the corresponding single main-span structure.

3.2 Longitudinal stiffness optimization of central pylon 3.2.1 Decisive factors of longitudinal stiffness selection As concluded in the previous section, the most important structural characteristic is probably the longitudinal bending stiffness of a central pylon, Rp, defined as

where T1 and T2 = cable forces at the central pylon top; α1 and α2 = the cable angles at the central pylon top; and δp = longitudinal displacement of the central pylon top. Under the most unfavorable load condition, only one main span being loaded, the longitudinal bending stiffness of a central pylon dominantly controls bridge structure performance including displacements and stresses of central pylon and deck as well as sliding resistance between main cable and saddle pad. Accordingly, the selection of the longitudinal stiffness of central pylon should carefully check with four decisive factors as follows.

(1) δd = Vertical displacement of the mid-span deck, which steadily decreases with the increase of the longitudinal stiffness of central pylon, Rp. (2) δp = Longitudinal displacement of the central pylon top, which also decreases with the increase of the longitudinal stiffness of central pylon, Rp. (3) σm = Maximum or minimum working stresses in central pylon, which increases with the increase of the longitudinal stiffness of central pylon, Rp. (4) Ks = Safety factor of sliding resistance between main cable and saddle pad, which decreases with the increase of the longitudinal stiffness of central pylon, Rp, and can be defined as,

where µ = friction factor between main cable and saddle pad, and µ = 0.2 based on various exper- iments; and θ = angle of saddle arc. According to the current design code for highway suspension bridges in China, the value of Ks should be greater than 2.

3.2.2 Limit types of central pylon for longitudinal stiffness With the aspect of longitudinal bending stiffness, longitudinal shapes of central pylons can be broadly divided into two limit types, A-shaped central pylon with the greatest stiffness described in Fig. 2a and I-shaped central pylon with the smallest stiffness described in Fig. 2b. With adopting a A-shaped central pylon (Fig. 2a), by which the longitudinal bending stiffness of a central pylon is relatively large, the vertical displacement at the mid-span deck and the longitudinal displacement at the central pylon top will be relatively small, but the safety factor of sliding resistance and the 55 Figure 2. Longitudinal shapes of central pylon.

Figure 3. General layout of Jiangsu Taizhou Bridge. maximum and minimum stresses in a centre pylon will be relatively unfavorable. Otherwise, with using I-shaped central pylon (Fig. 2b), by which the longitudinal stiffness is relatively small, the deck and pylon displacement will be relatively large, and the sliding safety factor and the pylon’s stresses will be relatively favorable. Taking as an example, Jiangsu Taizhou Bridge, 2940 m in total length, consists of four spans including double 1,080 m main spans and two side spans of 390 m shown in Fig. 3. In order to select an appropriate stiffness of a central pylon, both A-shaped steel pylon and I-shaped concrete pylon with various stiffness values had been firstly compared and contrasted through structural analysis. ∼ Based on the critical value of Ks = 2.00, the longitudinal bending stiffness of central pylon, Rp, should be kept at about 25 to 26 MN/m. It can be found that the working stresses are very high (σmax = 296 MPa and σmin =−200 MPa) in A-shaped steel pylon, and the cross sections are very large (6 × 6 m at the top and 12 × 10 m at the bottom) in I-shaped concrete pylon.

3.2.3 Optimal longitudinal stiffness and related characteristics Combined with the advantages of both types ofA-shaped steel pylon and I-shaped concrete pylon, an optimal shape of central pylon, inversedY-shapedsteel pylon shown in Fig. 2c, was then proposed. The longitudinal stiffness of this type of central pylon is mainly depended upon the spacing w and the height h1 of two legs, and the calculation with different combinations of w and h1 has also been performed. The longitudinal stiffness is about 26 MN/m, which is realized by providing the spacing w = 36 m and the height h1 = 72 m of two legs, and the maximum and minimum working stresses are equal to 245 MPa and −134 MPa, which are 37% and 61% reduced from those of A-shaped pylon, respectively. Therefore, the inversed Y-shaped steel pylon was chosen as the final scheme of the central pylon. With the final scheme of central and side pylons, the static analysis with the finite element model of Jiangsu Taizhou Bridge has been conducted in various load combinations, and the main calculation results of the most critical combination, dead load plus live load applied in one main span, are shown in Table 2. 56 Table 2. Calculation results under dead load plus live load applied in one main span in Taizhou Bridge.

Longitudinal displacement Side pylon loaded Central pylon Side pylon unloaded at the top (m) 0.194 1.876 0.022 Applied force at the central Loaded-side cable force Unloaded-side force Longitudinal shear force pylon top (MN) 183.9 170.8 9.84 Maximum vertical displacement Upward Downward Summation of deck (m) 3.200 4.352 <7.552 Main parameters of sliding Frication coefficient Safety factor Maximum hanger angle ◦ resistance µ = 0.2 Ks = 2.36 9.75 Working stresses in steel central Single column section Two leg section pylon +234 to −125 MPa +236 to −135 MPa

Table 3. Comparison of the first and second natural frequencies.

Lateral bending (Hz) Vertical bending (Hz) Torsional vibration(Hz)

Deck and Pylon First Second First Second First Second

Hinged connection 0.0696 0.0934 0.0639 0.0868 0.2620 0.3457 Fixed connection 0.0911 0.0934 0.0788 0.1136 0.2624 0.3457 Taizhou Bridge 0.0723 0.0971 0.0799 0.1026 0.2732

3.3 Fixed connection of deck and I-shaped central pylon 3.3.1 Stiffness comparison of fixed and hinged connection With the experience gained from Jiangsu Taizhou Bridge, neither A-shaped scheme nor I-shaped scheme can meet with the requirement of longitudinal stiffness of central pylon in double main-span suspension bridges. However, if we can make a double main-span suspension bridge with A-shaped central pylon more flexible or with I-shaped central pylon more rigid, the same objective can also be touched. This is truly possible based on the following experience of Anhui Maanshan Bridge, a double main-span suspension bridge with I-shaped central pylon. As mentioned before, Anhui Maanshan Bridge is another double main-span suspension bridge with the span arrangement of 360 + 2 × 1080 + 360 m, 2880 m in total length. Since the I-shaped central pylon has the smallest longitudinal stiffness among three pylon types, very large cross section was chosen, for example, the cross section at the bottom being 17 m longitudinal and 10 m transverse (APHDI et al. 2008). In order to further improve the longitudinal stiffness of central pylon and the vertical stiffness of bridge deck, the fixed connection between central pylon and steel box deck was attempted while the typical connection between deck and pylon is hinged. The I-shaped pylon was finally designed as a hybrid structure of concrete box section at the bottom and steel box section at the top to be connected with steel box girder and to reduce dead weight. In order to compare the stiffness of two structures with fixed connection and hinged connection between deck and central pylon, a dynamic finite-element analysis was carried out, and the first and second natural frequencies of the structures have been identified and compared for these two connections and with Jiangsu Taizhou Bridge in Table 3 with the following conclusions (Ge et al. 2008). The first and second vertical bending frequencies with fixed connection of deck and central pylon are 23% and 31% greater than those with hinged connection, and 1.4% smaller and 11% greater than those of Jiangsu Taizhou Bridge, respectively. With this fact, the objective of stiffness improvement was clearly realized through an innovative concept with replacing traditional hinged connection to fixed connection.

3.3.2 Computational prediction of static structural performance With fixed connection of deck and central pylon, the computational prediction of static structural performance of Anhui Maanshan Bridge has been carried under various load combinations. Table 4 57 Table 4. Calculation results under dead load plus live load applied in one main span in Maanshan Bridge.

Longitudinal displacement Side pylon loaded Central pylon Side pylon unloaded at the top (m) 0.189 1.790 0.020 Applied force at the central Loaded-side cable force Unloaded-side force Longitudinal shear force pylon top (MN) 184.8 171.9 9.65 Maximum vertical Upward Downward Summation displacement of deck (m) 3.142 4.214 <7.356 Main parameters of sliding Frication coefficient Safety factor Covered angle at pad ◦ resistance µ = 0.2 Ks = 2.32 47.92 Working stresses in hybrid Steel section Concrete section central pylon +232 to −142 MPa +17.4 to −0.16 MPa lists the main evaluation results under the most critical load combination, dead load plus live load applied in one main span (APHDI et al. 2008).

3.3.3 Experimental evaluation of aerodynamic stability As a long-span suspension bridge, one of the most challenging problems inAnhui Maanshan Bridge is aerodynamic stability after evaluation of static structural performance. Based on the dynamic structural analysis results, including the first vertical bending and torsional vibration frequencies and the equivalent mass and mass moment of inertia, the sectional model of Maanshan Bridge was designed and manufactured with using CFRP plates to increase model stiffness and reduce model weight (Ge et al. 2008). With the emphasis on aerodynamic stability, the sectional model wind tunnel experiment was carried out in the TJ-1 Boundary Layer Wind Tunnel with the working section of the 1.8 m width, the 1.8 m height and the 15 m length. It was found that the minimum critical flutter speed is 60 m/s, occurred at the attack angle of 3˚ under smooth flow, which can meet with the aerodynamic stability requirement of 56.6 m/s (Ge et al. 2008). However, the final confirmation for aerodynamic stability will be carefully conducted with full aeroelastic model wind tunnel testing in the TJ-3 Boundary Layer Wind Tunnel with the world’s largest working section of the 15 m width, the 2 m height and 14 m length in the near future.

4 TWIN PARALLEL DECK CABLE-STAYED BRIDGES

Based on the information collected, the engineering practice of long-span cable-supported bridges has brought about a bridge deck width up to 40 m for six-lane traffic. With the ever-growing traffic demands, sometimes it is necessary to build a new bridge side by side to the existed one or originally to build two parallel bridges to provide wider bridge deck for more traffic lanes, for example, eight or ten lanes, which results in another bridging capacity innovation configuration, twin parallel deck bridge. Taking as an example in China, Ningbo Yongjiang Bridge with twin parallel deck spanned 468 m is introduced in conceptual design of pylon type and deck section.

4.1 Numerical comparison of pylon shapes and deck sections Ningbo Yongjiang Bridge was designed as a cable-stayed bridge with the span arrangement of 63 m + 132 m + 468 m + 132 m + 63 m shown in Fig. 4 due to the navigational requirement, and twin parallel decks of about 2 × 24 m wide in the feasibility study (ZPHPDI 2007). Aiming at low construction cost in China and high structural damping for dynamic performance, both pylons and bridge deck were proposed to be made by reinforced and prestressed concrete. During preliminary design stage, the further comparison of pylon types and deck sections should be made based on structural performance, in particular dynamic and aerodynamic characteristics. Two types of pylons, including the twin diamond shape in Fig. 5a and the twin H shape in Fig. 5b, and three 58 Figure 4. Span arrangement of Ningbo Yongjiang Bridge.

Figure 5. Alternative shapes of bridge pylons. Figure 6. Alternative cross sections of bridge deck.

Table 5. Comparisons of fundamental natural frequencies of six combinations.

Lateral frequency (Hz) Vertical frequency (Hz) Torsional frequency (Hz) Bridge Deck Cross-Section Diamond H-shaped Diamond H-shaped Diamond H-shaped

Closed box 0.248 0.279 0.310 0.302 0.937 0.896 Twin boxes 0.272 0.308 0.308 0.301 0.810 0.794 Twin side ribs 0.306 0.345 0.310 0.299 0.599 0.555

types of deck sections, including the closed box in Fig. 5a, the twin separated boxes in Fig. 6b and twin side ribs in Fig. 6c, have been numerically compared in structural dynamic characteristics, and the fundamental frequency values of six combinations are listed in Table 5 (Ge et al. 2007). With the comparison in total six alternatives, the fundamental frequencies of vertical vibrations are almost in same for different types of deck sections and pylon shapes. Comparing with two pylon types, although the fundamental lateral frequencies with the twin diamond shaped pylons are about 13% lower than those with the twin H-shaped pylons, the fundamental torsional frequencies with the twin diamond shaped pylons are about 2% to 8% higher than those with the twin H-shaped pylons. The twin diamond shaped pylons had been selected for further design with considering simpler force resistance in pylons, especially for the central column. Among three types of bridge deck sections, the fundamental lateral frequency of the twin side rib cross section is about 13% and 23% higher than that of the twin separated box and the closed box cross section, respectively, 59 Figure 7. Composite cross sections of bridge deck (Unit: cm). but the fundamental torsional frequency of the twin side rib section is about 26% and 36% lower than that of the twin box section and the single box section, respectively. In order to predict aerodynamic stability, two-dimensional flutter analysis was carried out for the twin side rib section as the worst case of deck flutter. It was found that the critical flutter speed at the attack angle of 3◦ is 83 m/s, which is greater than the required speed of 66 m/s (Ge et al. 2007). Therefore, the twin side rib cross section had been finally chosen as the most economical cross section for the further design.

4.2 Further step in durability consideration The story of deck cross section selection was not ended in the preliminary design stage, but con- tinued in the further design, the detailed design stage. At the evaluation meeting of the preliminary design, the expert evaluation committee did not confirmed the twin side rib section of bridge deck based on some unsuccessful applications in domestic cable-stayed bridges, which have some durability problems, for example, cracks on both the top and bottom of the deck as well as in transverse diaphragms, unpredictable vertical deformation of deck, irreplaceable deck as a whole cross section, and so on. Following the instructions made by the committee, the design team proposed a new deck section, a composite structure with 2.3 m deep steel box girders and 0.26 m thick prestressed concrete plate shown in Fig. 7. Of course, cracks on the bottom of the girder and in transverse diaphragms can be completely avoided with steel structures, and deck plates become replaceable members if necessary, by which structural durability is greatly improved. With this composite cross section, structural dynamic analysis results in the fundamental natural frequencies of 0.307 Hz in vertical bending, 0.334 Hz in lateral bending and 0.667 Hz in torsional vibration. Compared with the twin side rib section, the frequency values in lateral bending and torsional vibration are increased in the new section while the vertical bending frequency almost remains in same. Aerodynamic investigation was followed in sectional model wind tunnel testing manner, and the minimum critical flutter speed was found to be equal to 95.7 m/s, measured at the attack angle of 3◦ under smooth flow, which is much greater than the aerodynamic stability requirement of 66 m/s.

5 CONCLUSIONS

The current bridging capacity is technically reviewed as single main span up to 2 km and 1 km together with total deck width about 40 m for suspension bridge and cable-stayed bridge, respec- tively. Bridging longer water body and providing wider traffic passageway needs innovations in not only single longer span and sole wider deck but also multiple main spans and several parallel decks based on deep water environments and ever-growing traffic demands. After having made comparison and contrast of single and double main span suspension bridges, the special structural characteristic for double main spans is revealed as longitudinal stiffness of central pylon. Through structural design and analysis of static and dynamic performance, inversed 60 Y-shaped central pylon was innovatively chosen in Jiangsu Taizhou Bridge to provide with proper longitudinal stiffness. Anhui Maanshan Bridge was designed with I-shaped central pylon combined with fixed connection between deck and central pylon as another kind of innovative manner for realizing appropriate longitudinal stiffness. Two typical pylons for twin parallel deck cable-stayed bridges and three kinds of concrete deck section have been numerically compared in earlier design stage of Ningbo Yongjiang Bridge, and two diamond shape pylon and twin side rib section have been decided among six combinations. Under the pressure of structural durability, the concrete rib girder has been replaced by a composite cross section with steel boxes and prestressed concrete plate for reducing cracking and becoming replaceable. The work described in this paper is partially supported by the Natural Science Foundation of China under the Grant 50538050 and the Hi-Tech Research and Development Program of China under the Grant 2006AA11Z108. The dedication and efforts of our colleagues, Drs. X.H. Luo, Z.Y. Zhou, F.C. Cao and Z. Lin and Mrsses. T. Pan and Z.G. Wei, are highly appreciated.

61

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Overcoming technological challenges to create new values for bridges

S.P. Chang Department of Civil & Environmental Engineering, Seoul National University, Seoul, Korea

ABSTRACT: All along its evolution, bridge took fully place as entity of the life and also left behind footprints of the civilization as a remnant of its epoch. Regard to such aspect, the value of the bridge should not be limited to an engineering value that is a tool to cross natural obstacles like valleys, rivers or bays but should be indeed extended to economic, social and political, historical as well as cultural values. As a social infrastructure, bridge concretizes the lives of its contemporaries and occupies a place of importance among the social infrastructures. However, bridge construction technology has been set apart at the entry of the 21st century due to the strong public aspect and extremely long service life of bridge structures, which have to stand 10 to 1000 times longer than the products of other technologies, as well as the absolute requirement for structural safety during the lifespan of the bridge. Even though, the globalized world of 21st century being recognized as a world where boundary and national identity between scientific disciplines will not have sense anymore owing to the emergence of knowledge information technology, bridge technology will not survive if it remains secluded in its traditional Newton’s physics-based classics. If improvement of the quality of life and sustainable development are set as the vision of the future at the entrance of the 21st century, it appears today that the values of a bridge should go beyond its technical value featuring safety, constructability, economy and, maintenance by at first embracing the concept of blended values like its national and regional economic value involving global and local develop- ment, its cultural value related to aesthetics, history, culture and, tourism, and its political value. The overcoming of the limits in bridge technology regard to new environmental changes and social needs or the overcoming of limits or challenges that have never been thought by the mankind to date are the keys for the creation of new values of bridges. This paper compares the histories of bridge engineering with other scientific fields of engineering, and attempts to discuss the possibil- ities for bridge engineers to create new values through fusion technology with other engineering disciplines.

1 INTRODUCTION: THE 21ST CENTURY, ERA OF FUSION

Bridge structures evolved since ancient times together with the life of people. Bridge has unques- tionably widened the living space of humanity and allowed culture to spread freely from natural obstacles. All along its evolution, bridge took fully place as entity of the life and also left behind footprints of the civilization as a remnant of its epoch. Regard to such aspect, the value of the bridge should not be limited to an engineering value that is a tool to cross natural obstacles like valleys, rivers or bays but should be indeed extended to economic, social and political, historical as well as cultural values. As a social infrastructure, bridge concretizes the lives of its contemporaries and occupies a place of importance among the social infrastructures. Questioning people about their impression about a bridge will generally lead to answers mention- ing “beauty” and “symbolism”. Indeed, the Golden Gate Bridge will be straightforwardly associated with San Francisco and vice versa. Occasionally, the Golden Gate Bridge is also associated to the symbol of USA. However, despite of such symbolism, bridge engineers attached essentially the 63 value of bridge with its functional and economical features such as function, structural safety, constructability and maintenance even after the mid of 20th century. The adoption of bridge forms as a major design concept can probably be traced to the series of lectures given by Fritz Leonhardt in 1970s (Leonhardt 1982) on the aesthetics of bridges. However, bridge engineers are still apprehending bridge as an infrastructure of which operation intends to promote the development of the surrounding industrial zones and which generates unexpected sup- plemental societal and political outcomes. These supplemental outcomes can be tourism, historical and cultural values created by the bridge itself, and economical values at the regional and national levels generated by the natural inflow of population to the region at hand resulting in the increase of employment and boosting of buying power. Moreover, it seems also that the bridge engineer is often depreciating aesthetics as a task reserved to the supervision of the government or belonging to the domain of politicians. Though, today as we just entered in the 21st century, time is for bridge engineers to change their perception of bridge from its industrial societal value in the 20th century to a new and forwarding context. The 21st century is recognized as the era of fusion. The progresses realized in information and telecommunication technologies made them grow as new core technologies, of which fusion is becoming a requisite for not only the development of scientific and engineering technologies but also more extensively of cultural sciences and society, policy and economy, and culture and history. This paper compares the histories of bridge engineering with other scientific fields of engineer- ing, and attempts to discuss the possibilities for bridge engineers to create new values through fusion technology with other engineering disciplines, and particularly through cable supported long-span bridge technology.

2 THE 21ST CENTURY: CHANGE OF SCIENTIFIC AND BRIDGE TECHNOLOGY PARADIGM

The origin of bridge technology can be traced back to the beginning of civilization. Civilization developed technologies through learning and using tools and instruments for survival since the primitive ages when people discovered fire. Even if bridge technology is the result of a massage of technologies, it portrayed its contemporary societal trends at each epoch of the history. In other words, the progress of the societal public technologies being determinant in the development of cities takes a decisive role at the national level since it maintains the functions of the nation through the establishment of efficient transportation and communication systems. For example, the admirable roman road system and bridges were the indispensable instruments necessary to maintain the Roman Empire and sustain its prosperity (Figure 1).

Figure 1. Road network of the Roman Empire and Pont du Gard built between AD 40 and 60 to transport fresh water across the Gard river. 64 Despite of its importance, one has to notice the absence of scientific and systematic educational system related to the initiation of bridge technology. Similarly to other technologies, bridge tech- nology was transmitted individually through apprenticeship under a master. As a matter of fact, this process was extremely closed and exclusive. After having being elevated to the title of master, the master had to develop his techniques and skill by himself little by little through experiences. Since technical development was achieved by means of corroborative experiences, the ancient bridges we can still admire today are extremely robust. Technology started to be systematized thanks to the works of the Roman technicians, who suc- ceeded in giving hierarchical shape to these simple techniques. These technicians being skilled specialists in each field were socially recognized. They continued to prosper, maintain and develop their skills until the medieval ages without the support of sciences. In order to satisfy and respond to the restlessly changing aspirations of the time, bridge technology experienced extremely slow development thanks to a very few ingenious technicians. Throughout centuries, these ingenious bridge technicians transmitted their skills from generation to generation to achieve improved effi- ciency and convenience of the technology. Entering in the Renaissance, the era of these technicians or masters reached a turning point. The scientific way of thinking of Copernicus adopted and expanded by Galileo and Descartes resulted in the Scientific Revolution of the 16th and 17th centuries (Figure 2). Newton began his works combining science and technique. The Scientific Revolution continued until the Enlightenment of the 18th century and was the leitmotiv of the “Machine Revolution” or Mechanization in the 19th century (Figure 3). Meanwhile, it was also the motive of the Industrial Revolution (1776), which finally transformed gradually the living envi- ronment of mankind from the farming-based society onto the industrial society that is spinning

Figure 2. Works of Copernicus (1473–1543), Galileo (1564–1642), Descartes (1596–1650) and Newton (1642–1727), bases of the Scientific Revolution of the 16th and 17th centuries.

Figure 3. The Industrial Revolution (18th–19th centuries) and Firth of Forth Bridge (built 1882–1890).

65 Figure 4. Albert Einstein (1879–1955): 1905’sseminal paper on relativity and related NewYorkTimes articles published in 1919.

industry (1800s), railway/steel (1850s), automobile/electricity/petroleum chemicals (1900s) and electronics (1960s). The entrance of mankind in the industrial society gave rise to active circula- tion of manufactured goods and raw materials. This led to societal needs for wider and extensive transportation infrastructures like roads and railways. The development of technologies enabling the construction of bridge was an absolute requisite to meet the objectives of the Firth of Forth Bridge (Figure 3) that were longer, larger and stronger. Even if bridge technology recognized development owing to the scientific achievements realized throughout the Industrial Revolution, bridge technology was still featuring the level corresponding to that before the Renaissance. That is, bridge technology seemed to remain limited to the framework of the master system until 1960s which preserved traditional sciences based on Newton’s scientific paradigm and focused on becoming the best in one item at once. This trend continued until after 1960s when the technology-based civilized society started to turn into a science-based society through the “Electronic Revolution” with the successive publications of the special and general theories of relativity by Einstein and the raise of quantum mechanics and electromagnetic (Figure 4). In reality, the progress of bridge technology did not exhibit significant difference with those of scientific technologies until then. Thereafter, since the 1960s, the traditional sciences started to be coupled with the electronic information processing and electronic communication technologies. And, late in the 20th century, the Digital Revolution known with its new scientific paradigm revolutionized changed our civi- lization onto a knowledge information-based society. Until the mid of the 20th century, a linear development model pioneering the scientific technology from the discovery of knowledge to its commercialization was timely efficient. However, in the ubiquitous 21st century, the faster circula- tion of information is making such linear model obsolete. Since innovation shall occur directly on field, the feedback channels between each development stage of scientific technology comprising bridge will play an increasingly important role. The generation and spreading processes of information are fastening. The connection between demand and supply of scientific technology is becoming more intense. Globalization of the market is today a reality where the stronger is playing an overwhelming role and is omnipotent. Moreover, the fusion of culture, science and economy has become a must since not only products but also culture are today thrown into the market. The specificity of the late 20th century was the combination of scientific technologies, which was nonlinear, complementary and systemized. The combination of optical science with electronics created the optoelectronics and gave birth to the optical communication system through optical fibers. The blending of mechanics and electronics resulted in the mechatronics revolution and opened widely the doors for the entrance of automation in the manufacturing activities. 66 Figure 5. Development project of Incheon Songdo free economic zone.

Except for very particular cases, construction technology including bridge technology has been relatively set apart at the entry of the 21st century. Such disdain can be attributed in priority to the too tardy attempt of bridge technology to fuse with other technologies compared to other fields after 1960s and especially after 1990s. This late recognition is due to the strong public aspect and extremely long service life of bridge structures, which have to stand 10 to 1000 times longer than the products of other technologies, as well as the absolute requirement for structural safety during the lifespan of the bridge. These inherent features led to the accepted practice of refusing the application of technologies that have not preliminary been tested and proved on field. Even though, the globalized world of 21st century being recognized as a world where bound- ary and national identity between scientific disciplines will not have sense anymore owing to the emergence of knowledge information technology, bridge technology will not survive if it remains secluded in its traditional Newton’s physics-based classics. Without full participation in the glob- alization of scientific technologies, bridge industry will likely decline to become a subordinate subcontractor of other fields such as finances or telecommunication technology. Such trend can already be perceived in Songdo’s development project at Incheon International City, Korea, where information and telecommunication conglomerates are leading the project while construction industry is playing simply the role of supplier of structures required by the projectors (Figure 5). However, rather than the passive participation in the urban planning shown by the technicians of other fields, it is advisable that, owing to their specialized knowledge and skills, construction engineers should participate in the urban planning or land development planning since the beginning. Another difference of construction engineering compared to other engineering fields is the creation of comprehensive and cultural values. Accordingly, bridge engineers who are in charge of social infrastructures should also be able to provide the future living environment of our citizens, realize the infra satisfying the intended quality of life and convince citizens about the values of the bridge. Besides, bridge engineers should identify the insufficiencies of the current technology and implement as soon as possible R&D through the support of the government in order to complement our lacks. Since the quality of life in the future is relying on the level of the contemporary scientific technologies, the development process should involve newly developed sciences like robotics, ubiquitous computing, bioinformatics, and nanotechnology through merged education and R&D in order to plan the future social infrastructures rather than secluding bridge technology into a specific field.

3 THE SIGNIFICANCE AND VALUE OF BRIDGE

3.1 Concept of social infrastructure and value of bridge in the national stock Bridge bears inherently value as a national asset that is indisputable. However, in order to understand more thoroughly its value as national infrastructure, need is to remind and arrange the concept of SOC (Social Overhead Capital). 67 Figure 6. Comparative overview of the evolution of sciences and bridge technology.

3.1.1 Concept of SOC Any government shall continuously invest and finance SOC including transportation facilities to boost economy and reduce freightage fees. In Korea, the special accounts for transportation facilities introduced since 1994 made it possible to expand substantially investments for SOC. As a result, the inauguration of the first section of the Gyeongbu high-speed railway and the electrification of the Honam railway in 2004 launched a new era for the peninsula by enabling any place in the country to be reached within a half day and contributed significantly to the balanced economical development, national competitiveness and economic growth of Korea through the improvement of the domestic and international circulation of people and goods. However, investment for SOC decreased gradually after 2003 because of: (1) the sufficient expansion of the SOC stock realized to date; (2) overlapping investments for regional roads or railways; and, (3) the inevitable readjustment of the limited financial support to other branches like welfare, education and defense due to the reduction of financial incomes according to the entrance of Korea into the era of low economic growth. This resulted in the curtailment of financial support for transportation facilities and, subsequently, to various socio-economical problems like delays in construction. Such delays provoked the augmentation of traffic congestion and increase of indirect costs like field management costs by the construction companies. Especially, the continuous increase of freightage costs and traffic congestion costs will affect negatively the future national competitiveness. Currently, controversies are engaged regard to the supply level of SOC in Korea. The pros advocate the extension of SOC investments while the contras plead for their reduction (Ha et al. 2000). The amount of bridges being determined according to the evaluation of SOC supply level, the adequate volume of SOC stock shall be analyzed objectively and synthetically considering not only the future demand of traffic depending on the current and future economic growth of the country and, regional balanced development and globalization but also the cultural level of the country or city at hand. Diversified definitions are suggested for the concept of SOC with stress on functional aspect, which is used to classify SOC into industrial and welfare facilities. Industrial SOC as facilities necessary for the industrial production activity is closely related to the daily life. Representative examples of welfare SOC are roads, harbors, etc. as well as electricity, gas and communication. Welfare SOC is divided into several branches that are residential environment like water and sewage services, education, health and welfare, culture and recreational facilities. In Korea, the National Wealth Survey (Korea National Statistical Office 1999) proposes the following definition of SOC: “Overhead capitals with public purposes like transportation, telecommunication and electric power 68 supplying the bases necessary for the economic activities of the whole country but which are not participating directly in the production and consumption activities of each economic party.”

3.1.2 Interdependency between national competitiveness and SOC The value of bridge as SOC is closely related to the national competitiveness. Therefore, it seems interesting to reexamine the concept of national competitiveness and its relation with SOC. Clear definition of the national competitiveness has not been established to date. But, an indirect statement giving the meaning of national competitiveness can be expressed as: “the capacity of a country to provide efficient social structure, system and policy enabling companies and industries to compete successfully against foreign companies in the international market.” Securing SOC facilities like road, railway, harbor, communication, electricity and gas constitutes an imperative policy to sustain the industrial competitiveness. As a physical basis contributing directly to the economic growth and promotion of welfare, SOC is an asset supplying basic services necessary for any production activities, which promotes productivity, reduces transaction costs, increases hiring and revenues and augments exportation through the creation of external economic values. Drop down of the efficiency of circulation because of the insufficiency of freightage facilities like road, railway, harbor and airport and because of the lack of connections between transportation means, will as a matter of fact bring inconvenience in the lives of the population and degrade the competitiveness of the industry by generating tremendous socio-economic costs. The responsibility of the supply of SOC facilities falls basically to the government, but the burden of cost increase induced by the lack of SOC facilities falls to the private sector that is producers. This loop will finally result in the loss of international competitiveness of our national enterprises. Accordingly, the SOC stock of a country can be assessed as the index representing its capacity to supply public services that is the source of the national competitiveness. The annual reports on the national competitiveness ranking published by the International Institute for Management Development (IMD 2007) and World Economic Forum (WEF 2007) are also exploiting the SOC stock level as an index for the evaluation of the national competitiveness. SOC investment has naturally societal and economic impacts. Traditionally, SOC investment is known to contribute more sensitively on the economic development and activation of the market as much as the country is underdeveloped. In particular, investment on transportation facilities like road and railway is imperative for the balanced regional development by promoting smooth circula- tion between regions. Since the economic effect of SOC investment involves mostly external effects exhibiting non-economical nature, measuring such effect itself is practically impossible. However, the promotion of industrial and commercial activities through the extension of SOC in the domain of transportation has already been proved by means of the analysis of evidences related to the role fulfilled by public works in the industrialization process of countries all over the world (Yang1994). Referring to the ratio of direct production activities to SOC, Hirschman (1958) stated that if that this ratio reaches a critical level, direct increase of productivity would not occur without extension of SOC. This ratio is called as the industrial support effect of SOC. Moreover, the impact on the increase of revenues stands also as an important effect of SOC investment. Even if the correlation between transportation facilities and the national income per capita differs according to the conditions of SOC, a value exceeding 0.5 is generally observed during development process. In Korea, the correlation between the GRP (Gross Regional Product) and road stock reaches approximately 0.85. The socio-economic impact brought by SOC investment in the transportation domain can be classified into direct and indirect effects. Direct effects cope with the contribution to the national competitiveness and betterment of the citizens’ accommodation such as the shortening of distance or remoteness between regions through the construction of roads and railways, the reduction of circulation costs through the decrease of labor and oil expenses brought by the improvement of traffic conditions, the shortening of transit time and the improvement of traffic convenience. Indirect effects deal with balanced regional development and activation of the economy such as 69 Figure 7. Effect of the expansion of transportation SOC on the regional accessibility (remoteness) and regional economic development in Korea (Kim et al. 2007).

Figure 8. Topography and road network of the Korean peninsula and, bridge stock by decades in Korea. the optimization of land use over the whole country through the improved accessibility between regions, extension of factories location and boosting of regional economy (Figure 7). SOC investment in the transportation domain has effects not only in the SOC industry itself but also generates extensively and significantly the productivity in other domains. The production inducement coefficient by industrial sector stands for the volume of production induced directly and indirectly in the whole industry when the final demand for goods and services increases by one unit in each industrial sector. A survey of the production inducement coefficient by industrial sector in Korea reveals that, in 2003, construction reached 1.980 which is larger than the 1.972 averaged by the manufacturing industry and remarkably larger than other industries. In addition, the employment inducement coefficient of the construction industry was 14.2515 meaning that an investment of 1 billion KRW creates 14 new employments. This value is the highest behind the service industry, which demonstrates that SOC construction investment in the transportation sector is providential for relieving unemployment.

3.1.3 Bridge value in the national stock Bridges occupy an immutable place as national SOC for the economic growth of a country. Besides, being anchoring the joys and sorrows of people who crossed it, bridge becomes history as time elapses, and becomes culture as poured with the souvenirs and lives of each epoch. In our modern times, bridges are also constructed as a tool for creating the culture of a city, and politicians are framing bridge construction policies for political purposes as a solution to settle social dissensions between regions. In countries like Korea of which about 70% of the territory is composed of mountainous areas and surrounded in three sides by saw-tooth coastlines, the key facilities of the transportation network are naturally bridges and tunnels (Figure 8). Therefore, securing bridge technology is imperatively required for the national economic growth. Taking example of the bridge stock in Korea (Figure 8), statistics are giving a bridge of 87 m every 3.4 km on the mean in the 102,209 km of the road network, a bridge of 86 m every 2.9 km on the mean in the 7,381 km of the railway network and, a 70 Figure 9. Cultural value of bridges. bridge of 98.8 m every 6.1 km in the 508 km of the high-speed railway. Here, even if the volume of the bridge stock as a national asset occupies a minor portion of the whole transportation facilities stock, bridges constitute the most sensitive critical pass among the transportation facilities. This means that accidents occurring in a bridge whatever small it is in a road or railway section will deprived the corresponding section from its functions during the duration of repair or rehabilitation. Accordingly, despite of the relatively modest size of the bridge stock, bridges have a key role and value enabling the functions of road and railway to be realized in linking mountainous areas, islands or islands to the continent.

3.2 Historical and cultural values All over the world, any citizen wants to find out a monument or feature symbolizing the city he is living in. If a Parisian tells proudly that the Eiffel Tower is the symbol of Paris, no one should deny him. Similarly, no one in the world would contest that the Golden Gate Bridge is the emblem of San Francisco. If you are a British national and asked about your impression about the Firth of Forth Bridge, you will probably mention its historical value in your answer, and if you are asked about the London Bridge, there is great chance that you will talk that it is the idol of London. That is a fact; emblematic bridges as well as historical bridges are always bearing cultural value. Accordingly, most of the bridges having historical meaning are managed by the Cultural Heritage Bureau. This is probably due to the fact that culture is reflecting the life of its epoch, and that a bridge is concretizing and reflecting at the most the best engineering technology of its contemporaries in order to enhance their quality of life. For example, Brooklyn Bridge in USA, the first structure applying nonlinear theory, is an amazing achievement of the civil works in the 19th century with its main span of 500 m that any tourist visiting New York has to cross. The view of Manhattan offered from the bridge is a superb tourist brand. Especially, the silhouettes of the skyscrapers that can be admired from the bridge may be qualified without exaggeration as the highlight of New York. This bridge has also been the motive of numerous movies like “Once upon a time in America”, the Hong Kong made romantic drama “Autumn’sTale”, and also “A view from the bridge”, the opera of William Bolcom. The word “bridge” has also been adopted in the title of several movies such as “The Bridge on the River Kwai” featured by William Holden or “The Bridges of Madison County” of Clint Eastwood (Figure 9). In Korea, Supyo Bridge can be cited as a representative historic bridge (Figure 10). Supyo Bridge is a bridge constructed to measure the level of water flowing in Cheonggye Stream, dredged from the streamlet during the transfer of the capital to Hanyang in the Chosun Dynasty. During the reign of the 21st king of Chosun dynasty, Yeongjo, a dredging office was let in the eastern side of the bridge to report any change of the water level to the official clerk of the court. Another example of 71 Figure 10. Views of Supyo Bridge and Gwangtong Bridge before covering and after rehabilitation of Cheonggye Stream. bridge bearing historic and cultural values is Gwangtong Bridge in Seoul (Figure 10). The original name of the bridge was Great Gwangtong Bridge meaning the great bridge located in Gwangtong. At the time, Seoul was composed of mountainsides and valleys requiring numerous bridges for the roads to let passage to people and horses, and Gwangtong Bridge was the largest one among these bridges. A wood-framed bridge compacted with earth was erected at first. The original bridge was swept away by floods in 1410 (10th year of the reign of King Taejo, first king of Chosun dynasty) and reconstructed using the 12 carved stones decorating the tomb of a consort of King Taejo. The bridge disappeared during the covering of Cheonggye Stream in 1958 to 1961 but has been restored in July, 2003 through the rehabilitation of the stream. The designer of Brooklyn Bridge, John A. Roebling, did surely not imagine that his achievement would become the best attraction of New York100 years after its construction. However, if we rec- ognize today the cultural value of Brooklyn Bridge, we should certainly not limit the value of the bridges we are designing to their economical aspects but extend our approach to the planning of a new tourist brand or plan the bridge as an important cultural and historic component remodeling the existing downtown. In fact, numerous advanced countries have already implemented such approach like the Millennium Bridge in UK. Developing countries like Korea are now starting to recognize seriously the values of such bridges at the entrance of the 21st century. A point to be stressed in the design of a bridge is to conceive a symbol of the city at hand so that the surrounding buildings har- monize with it rather than letting the bridge be merely a sub-component subordinated to the urban planning. To realize such consideration in the design, apart from their engineering skills and knowl- edge about regional history and culture, bridge engineers should also have political competences that will enable them to achieve simultaneously their own ideas and initiatives (Seely 1996). Attempt to recreate or remodel a city as an international city of culture by reevaluating the values of bridges is actually conducted in Seoul. The master plan for the “Urban Renaissance for Seoul” has been launched by Seoul Metropolitan Government in 2007 to renew and regenerate the city as a design city recognized all over the world until 2010. The project divides the old downtown of Seoul through 4 cultural axes partitioning Cheonggye Stream flowing from East to West into 4 equal sections to be redeveloped so as to remodel Seoul as a masterpiece city. These 4 cultural axes are: (1) history and culture corridor; (2) tourism corridor; (3) green corridor; and, (4) creative corridor (Figure 11).Among them, the history and culture corridor linking Gyeongbuk Palace, where the suc- cessive kings reigned over the peninsula during the 600 years of the Chosun dynasty, to Seoul Station of which building was erected during the Japanese occupation, and passing through Gwanghwamun Palace, Cheonggye Stream and Namdaemun Market is an axis embracing Seoul’s past, present and future history and culture. However, we should not forget that the future Seoul has to be drawn from today. In order to realize this future concept, Seoul City seems to pay particular attention on the reshuffling of Seoul Station’s Plaza (Figure 12) and the repair of Namdaemun Market. The reconstruction of the viaduct of Seoul Station is involved as the pivot of the reshuffling of Seoul Station’s Plaza. The values required by Seoul City for the viaduct are SOC value in its traditional sense together with an emblematic value picturing Seoul’s transfiguration through a design identifying Seoul as capital of the future (Figure 13). Such request formulates the desire of Seoul City to create cultural and historical values forecasting the future quality of life of its 72 Figure 11. Urban Renaissance for Seoul where tomorrow harmonizes with past.

Figure 12. Current view of the surroundings of Seoul station with its viaduct.

Figure 13. The new Seoul station’s viaduct proposed by Seoul National University students: Chang Sung Feel, Chameroad, Neo-Spiral. citizens by means of today’sbridge technology. Accordingly, bridge engineers have now to design a bridge expressing symbolically the life contents of the future Seoulites by merging all the existing sciences and technologies as well as those to be developed in the 21st century.

3.3 Political value During its short 60 years old history, bridge technology in Korea fulfilled conscientiously its role as a tool for the successful execution of the industrial development policy of the government until the end of the 20th century. Korea grew from a developing country to join the OECD, and takes today much interest on environmental and welfare concerns. The desire of the citizens seems to be oriented toward the improvement of the quality of life through leisure and cultural activities rather than the national development. Korea is characterized by the concentration of political, social, financial, cultural and educa- tional headquarters in the capital area leading to restless issues on regional balanced development. 73 Figure 14. Olympic Bridge and location of the bridges built in the southwestern cost of the peninsula.

Figure 15. Pentaport project of Incheon Area between both Koreas and bridge project to link Incheon international Airport and Ganghwa Island.

The national unity in the South and the division with the North are also remaining critical political problems to solve. Political resolution exploiting technological tools may constitute a solution to settle these issues. Accordingly, Korea began to recognize the necessity to involve technicians as policymakers since 2000 whileAmerican engineers already participated directly in the national pol- icy relative to the land development since 1900 (Seely 1996). An example of the direct contribution of the American construction technicians to the execution of governmental policy can be seen in the additional development planning of Washington, DC for the 21st century (Griffith 1996). Before 2000, bridges in Korea were planned and constructed under the pressure of politicians rather than being the fruit of the engineering project. In spite of the poor economic benefits brought by the development of backward regions, a number of bridges have been constructed along the southwest- ern coast of the peninsula for political arguments (Figure 14). Another typical example in Seoul is the Olympic Bridge constructed in commemoration of the 1988 Seoul Olympics (Figure 14). For the Korean technicians to influence directly the setup of policy, engineers should suggest the construction of inter-Korean transportation channels to the government so as to enable direct overseas exportation of the goods produced in the Industrial Area of Gaesong in the North through Incheon International Airport. This can be achieved by the erection of bridges linking both Koreas, which will strengthen the inter-Korean economic collaboration (Figure 15).

4 COPING WITH TECHNOLOGICAL CHALLENGES TO REALIZE BRIDGE VALUE

4.1 Current limits of bridge technology The modern bridge technology in Korea started lately at least 150 years after Europe and 60 years after Japan. The Korean bridge engineers can be proud of their contribution in the conversion of 74 Figure 16. Namhae Bridge, the first cable-supported bridge in Korea, and Yongjong Bridge, the first three-dimensional self-anchored suspension bridge in the world.

Korea from an agricultural country to an industrial country during the last 50 years. However, optimists emphasize the dynamicity of the current bridge technology in Korea together with its specificity and advanced skills, whereas pessimists point out the low to moderate-priced image of its quality. The first suspension bridge in Korea, Namhae Bridge, was constructed in 1973 with the help of Japanese technology (Figure 16). This experience was the beginning of our entrance in the construc- tion of cable supported bridges. Thereafter and until a recent past, most of the long-span bridges constructed in the peninsula adopted the technologies of advanced countries. During this process, the attitude of the owner was extremely conservative and cautious in designing bridges in harmony with the surrounding because of the irrationality of the laws and systems established by the succes- sive governments, especially regard to budget and accounting law. Korean owners limited simply their recognition of the bridge as an instrument to solve the transportation and freightage problems. The turning point in the evolution of bridge technology in Korea happened in early 1990s following the successive collapses of several bridges crossing Han River in Seoul. Most of the bridges were designed and constructed from late of 1960s to late of 1980s as the final tool to achieve the industrial development policy. The governmental bureaucrats were essentially focusing on minimum safety considerations without concerns on the aesthetics and values of the bridges. These collapses ignited public consensus on the safety of bridges since people recognized that bridges had potential risk to collapse and could lead to tremendous social, political and psychological effects. Therefore, the after-Sungsu started the introduction of a supervision system as well as R&D related to maintenance involving periodical inspections. Especially, Seoul National University pioneered systematic development of bridge monitoring system techniques to evaluate the health of the bridges, which were then applied on field. Today, most of the long-span bridges in Korea are instrumented with bridge monitoring systems enabling automatic inspection. Until the end of the 20th century, the Korean bridge engineers believed that the bridge technology of a country could be measured by the maximum realizable main span length within safety require- ments. In such optic, Japan can be seen to remain the world leading country in bridge technology. Accordingly, Korea tried to compete with the Japanese technology of which results are the design and construction of two landmark bridges: a cable-stayed bridge with main span of 800 m linking Incheon International Airport to Songdo, and a suspension bridge with suspended span of 1545 m that is currently the third longest bridge in the world and to be completed in 2012 (Figure 17). However, the opinion of the Korean bridge engineers began to change gradually after the entrance in the 21st century taking notice of the assertion of Richard Rogers, who participated to the design of the Centre Pompidou in 1971 (Figure 18). Rogers asserted that “the determinant element in making a landmark is not the length or height of the bridge or building, but should be found in the superiority of the conceptual design.” As everyone knows, the Pompidou Centre in Paris is only 60 m wide with 6-storeys and 1 underground floor but the ducts and conveyance systems are disposed outside making it another landmark symbolizing Paris as the EiffelTower. The Millennium 75 Figure 17. Drawings of Incheon Bridge and 1545 Bridge.

Figure 18. Pompidou Centre in Paris and Millennium Dome in London.

Figure 19. Viaduc de Millau in France and Rion-Antirion Bridge in Greece.

Dome in UK is also framing a new landmark in London with its particular external appearance supported by 12 columns (Figure 18). The pertinence of such assertion can be easily demonstrated through two bridges that are the Viaduc de Millau in France and Rion-Antirion Bridge in Greece (Figure 19). The Millau Viaduct is a 7-continuous span cable-stayed bridge with tower height of 343 m and length of 2,460 m, which opened to traffic in 2004. The originality and ingenuity developed for the bridge can be observed through its challenging scene, remarkable conceptual design and perfectly automated construction method, which gave birth to a bridge bearing timely value at the entrance of the 21st century. Moreover, Rion-Antirion Bridge in Greece overcame the technological limits of the current seismic engineering with the perfection of the aseismic design of its foundations, from which we can experience new technological and cultural values. 76 The technological limits to be surmounted by the bridge engineer are multifold. These are the change of transportation and freight system through the connection of continents according to the globalization of the international economy, and the environmental changes of bridge construction according to the warming of the earth, the corresponding climactic changes which produce stronger and stronger typhoons, and the uncertainties of seismic activities. The wind technology required for Messina Bridge gives us an example and opens possibilities to create new scientific and cultural values through R&D for the minimization of the effects of wind for super-long bridges. Even if unfortunately the construction of Messina Bridge is encountering delays, the application of the developed technologies has been and is being already attempted in several bridges like Stonecutters Bridge in Hong Kong. The overcoming of the limits in bridge technology regard to the extreme climate we are experi- encing today or the overcoming of limits or challenges that have never been thought by the mankind to date are the keys for the creation of new values of bridges.

4.2 Multidisciplinary technology fusion 4.2.1 Fusion of scientific technologies Bridge technology to date is the result of the combination of the knowledge of practically all the scientific fields like road and transportation engineering, hydrology and hydraulics, soil and rock engineering, metal and non-metallic material engineering, structural engineering, mechanical engi- neering (construction equipment), maintenance technology (non-destructive inspection) as well as aerodynamics (wind engineering) and earthquake and aseismic engineering. Despite of such combi- nation, bridge engineering experienced a very slow development compared to the other industrial fields after 1960s, which can be attributed to the accelerated fusion of these fields with fore- front industries and the sudden development of information and telecommunication technologies, nanotechnology and biotechnology. In a technical point of view, the objectives of bridge engineering are primarily safety and eco- nomic efficiency. The realization of higher, longer while robust bridges depends firstly on the development of new high performance materials. For example, longer main span in cable sup- ported bridges shall be achieved through composite materials developed in collaboration with fiber and chemical engineering. Durability of bridge requires the development of high strength and high performance concrete conjointly with chemical engineering and high performance steel through the fusion with material engineering. The advancement of design shall pass by the automa- tion of the whole process from the planning and construction to the maintenance. The automation of construction needs robotics technology to replace human as well as the fusion with IT. Especially, the automation of maintenance shall be achieved through the development of robots based on IT together with unmanned inspection instruments for the detection of eventual anomalies or dam- ages in the members of the bridge. In addition, corrosion of the members constitutes a problem of primary importance to be solved where nanotechnology and biotechnology will play a helpful role.

4.2.2 Fusion of bridge technology with human and social sciences The collapse of Sungsu Bridge at the crack of dawn on October 21, 1994, made Korean people realize acutely and bitterly the importance of the development of bridge technology in constructing a healthy and safe society. Evidence of defective welding at the connection of members was diagnosed during the analysis of the causes of the collapse of this bridge. However, a more fundamental but indirect cause was the impediment of the development of construction technologies due to the series of regulations preventing all types of frauds and corruptions that may happen during the ordering process in the bridge industry together with the impetuous industrial policy established without consideration of the level of the domestic bridge technology at the time. In Korea, the regulations relative to construction, the regulations relative to market admission, the partitioning of the industrial sectors, the regulations relative to prices, the regulations relative to production system and the regulations for the protection of socially unfavorized people are 77 essentially containing features minimizing the scope of authority in the framework of corruption control and, formal transparency and impartiality through objectivity. However, whatever rigorous the law and system are, no further development of bridge technology should be expected if the level of the officials in charge of the management of the law and system or the consciousness of the industry parties providing the causes of corruption are far from being on the average themselves. The development of bridge technology of a country cannot be realized without improvement of the level of the construction culture of its citizens. The most straightforward way to advance the construction culture and bridge technology is the global standardization of the laws and systems relative to bridges. Such standardization shall adopt as fundamental principle of advancement the pursuit of the efficiency of the bridge industry, the pursuit of the best values, the contribution to the competitiveness, the securing of transparency and publicity, the consideration of the role of the owner, and the recognition of the authority. Moreover, the active support of R&D by the government remains a requisite for the development of technologies agreeing with the global standard. Therefore, decision of priorities should be taken under public consensus through the opinion of civic organizations.

4.2.3 Fusion of bridge technology with history and cultural fields Since a bridge is a SOC providing the means for freightage and transportation, the construction of this bridge embraces not only its contemporary engineering technology but also the lifestyle of its epoch. Following, bridges that are constructed today will become the barometers of the technological civilization and cultural levels of our epoch to our future posterity. Therefore, the bridge technology we are pursuing is not simply chasing for engineering safety and economic efficiency, but shall go beyond by embracing contemporary advanced and leading technologies based on aesthetics in harmony with the surrounding environment so as to symbolize and be the model of the city we are living in or more extensively of the history and culture of the country. To that goal, bridge technology should not remain solely in an engineering scope but should target a multidisciplinary level combining extensively the history, culture and even the philosophy of the construction site. Especially, a point to be stressed should be the prevention of the destruction of the past and cultural values of a country by bridge engineers through the construction of roads and bridges disregarding the existence of the national cultural heritage because relying excessively on the economy development policy. Such destruction can be avoided by making bridge technology creating new historic and cultural values.

4.2.4 Participation of bridge engineers in policy and education of experts A survey of the changes in the scientific policy in major advanced countries after the World’sSecond War reveals that its objectives focused on politics and security from 1950 to 1975, shifted toward economics and industry in the period between 1975 and 1995, reoriented to social welfare since 2000 (Chung 2000) to be currently axed on searching for solutions to solve global problems going beyond the national sphere. Especially, the environmental preservation is today leading resolutely the change and orientation of the market structure merged with advanced technologies to solve the corresponding societal problems. Similarly to the development flow of the whole scientific technology, the fundamental objec- tive of R&D for the future bridge technology shall be social welfare. Therefore, the determinant elements should naturally be employment and improvement of the quality of life. In addition, together with the development of information and telecommunication, its geographical scope shall be extended to the globe. Differently from 1960s, the R&D process will be centered to the system with nonlinear interdependencies between the scientific technologies. R&D will focus primarily on domains developing composite technologies through multidisciplinary and industry-academia- research institute collaboration. The executors of the policy shall achieve harmony and cooperation between the governmental administrative offices without being limited to specific fields. The fruits of R&D shall be able to provide solutions to societal problems like the warming of the globe and environmental changes. R&D priorities should break with the top-down model derived from a unique policy like politics/sciences or economics/industry but should apply a bottom-down 78 Figure 20. The 3 aspects of engineering as the art of science and production (engineering science, design and development, management and organization; Needs of bridge technology to overcoming challenges. model relying on society/politics. Even if the scientific technologies of the 21st century could be formulated based on the needs of organizations or typical parties of the society, public works including bridges require the arbitration of the government to conciliate the conflicting interests of the relevant parties. In other words, the level of bridge technology depends on the level of the officials in charge of construction affairs. Consequently, civil engineering should be learned in undergraduate college while graduate college should focus on legislation so as to educate policy experts in the field of bridge technology able to take active part of the establishment of governmental policies relevant to construction including bridge industry.

5 CONCLUDING REMARKS

The industrial society of the 20th century strived for independent development of each industrial sector, which was not a problem at the time. However, at the entrance of the knowledge information society of the 21st century, the independent survival of a particular sector has become unviable. The individual survival of what was called the chimney industry of the 20th century like the machine industry, the shipbuilding industry and the chemical industry has become today unfeasible, and continuous development can only be achieved through the fusion with advanced and leading technologies such as IT, BT and NT. Such situation is not limited to the machine and chemical domains but extends also to bridge industry as an unavoidable result of our times. Today, bridge technology cannot survive as the property of a country as much as the construction of bridge is not limited within the frontiers of a country. Similarly to the international vulgarization of information, bridge technology is not belonging anymore exclusively to a particular country. Bridge technology should be continuously updated and advanced through the merging of new and innovative technologies in other fields so as to be applied straightforwardly to the bridge construction site. However, differences with other industries shall be clearly identified since the construction sector and particularly the bridge construction industry being SOC is enrolled as a national asset. SOC and specially transportation facilities are not the properties of individuals but are pertaining to the public, which means that the bridge construction authority is the public that is the government. In other words, the users of transportation facilities like bridge structures are the citizens, and their planning, design, construction and maintenance shall be executed in compliance with the desire of the citizens. Though, until the end of the 20th century, bridge construction projects were driven according to the decision of officials in charge of construction affairs rather than through direct hearing of the citizens’ opinions. Especially, in developing countries, construction projects 79 are often executed unilaterally in the name of political policies favoring public purposes at the sacrifice of the privates under the despotism of the nation without consideration of the intention of individual citizens. In such case, the argument of economic development being the priority, construction should be cheap, rapid and numerous, which leads naturally to the destruction of the natural environment and residential environment. Under the immunity given by the national economic development, our construction industry destructed the environment and was marked as the principal cause impeding sustainable development. However, if improvement of the quality of life and sustainable development are set as the vision of the future at the entrance of the 21st century, it appears today that the values of a bridge should go beyond its technical value featuring safety, constructability, economy and, maintenance by at first embracing the concept of blended values like its national and regional economic value involving global and local development, its cultural value related to aesthetics, history, culture and, tourism, and its political value. To achieve such values of the bridge, bridge engineers should not limit their role as passive executors of the projects proposed by the government, but extend it as active participants and decision-makers in the establishment of land and city functions and corresponding development planning for the sustainable development of the society. Finally, for our opinion and voice to be reflected in the national policy, this process should go along with the fusion with engineering fields of other industrial sectors as well as social and human sciences.

ACKNOWLEDGEMENTS

This work is a part of a research of the Korea Bridge Design & Engineering Research Center (KBRC) at Seoul National University. KBRC is supported by the Korea Ministry of Construction & Transportation through the Korea Institute of Construction &TransportationTechnology Evaluation and Planning (KICTTEP).

REFERENCES

Chung, S.C. 2000. Trends of R&D policy in EU. Report of Science & Technology Policy Institute (STEPI). Griffith, R.W. 1996. Extending the legacy: Planning America’s capital for the 21st century. In Jerry Rogers et al. (ed.) Civil engineering history: Engineers make history; Proc. first national symp. on civil engineering history, Washington DC, 10–14 November 1996. New York:ASCE: 25–33. Ha, H.K. & Cho, D.H. 2000. Estimation of regional gross capital stock in transport sector of Korea. Policy Study 2001–06. The Korea Transport Institute (in Korean). Hirschman, A.O. 1958. The strategy of economic development. New Haven: Yale University Press. International Institute for Management Development (IMD). 2007. World competitiveness yearbook. Kim, H.J. & Chung, S.Y. 2007. Necessity of new frontier policy for the construction of expressways. Planning and Policy Brief No. 136, 2007.04.16, Korea Research Institute for Human Settlements (KHRIS). Korea National Statistical Office. 1999. National Wealth Survey. Office of Economic and Statistical Research. Leonhardt, F. 1982 (1994). Brücken / Bridges. Stuttgart: Deutsche Verlag-Anstalt. Seely, B.E. 1996. State engineering as policymakers: Apolitical experts in federalist system. In Jerry Rogers et al. (ed.) Civil engineering history: Engineers make history; Proc. first national symp. on civil engineering history, Washington DC, 10–14 November 1996. New York:ASCE: 123–135. World Economic Forum. 2007. The global competitiveness report 2007–2008. Yang, J.C. 1994. Social Overhead Capital. Seoul Press (in Korean).

80 Technical Contributions Advanced and high performance materials

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Durability of bridges made of advanced composite materials

J.R. Correia, F.A. Branco & J.G. Ferreira Instituto Superior Técnico/ICIST, Technical University of Lisbon, Lisbon, Portugal S. Cabral-Fonseca, M.I. Eusébio & M.P. Rodrigues Laboratório Nacional de Engenharia Civil, Lisbon, Portugal

Fibre Reinforced Polymers (FRP) have great potential for bridge construction, presenting several advantages when compared with traditional materials, namely the higher strength to weight ratio, the lower self weight, the electromagnetic transparency, the possibility of being produced with any cross-section geometry, the easier installation, the lower maintenance requirements and the improved durability under aggressive environments. Comparing with traditional materials there is practical evidence of improved performance of FRP composites when submitted to aggressive environments, which can be perceived by the long history of their use in marine vessels, piping, storage tanks and in several corrosive applications, such as the oil, chemical and water treatment industries. Paradoxically, one of the factors that is delaying the widespread acceptance of FRP composites in bridge construction as load-carrying structural elements is the lack of comprehensive and validated data on durability, which creates an obstacle for the construction agents (including owners, designers and contractors), as bridges service life is generally expected to exceed 50 years. For certain FRP materials, such as Carbon Fibre Reinforced Polymer (CFRP) strips/sheets or Glass Fibre Reinforced Polymer (GFRP) bars for concrete strengthening and reinforcement, design codes or guidelines are already available (e.g. American ACI 440.2R-02 and ACI 440.1R-06; and Canadian CAN/CSA S806-02), though a current debate is presently going on about the reliability of the adopted strength reduction factors. For other FRP materials, such as GFRP pultruded profiles, no design codes or guidelines are yet available and durability has been recently identified by several authors as the most critical gap between perceived need of information and available information, regarding future research (Kharbari et al. 2003, Harries et al. 2003). This paper presents results of an experimental research into the physical, chemical, mechan- ical, and aesthetical changes suffered by GFRP pultruded profiles used in bridge infrastructure under accelerated exposure to moisture, temperature and ultraviolet (UV) radiation (Correia et al. 2006). The profiles were submitted to the influence of four different exposure environments: (i) in an immersion chamber at 20◦C, (ii) in a condensation chamber at 60◦C, (iii) in a QUV accelerated weathering apparatus, and (iv) in a xenon-arc accelerated weathering apparatus. The results obtained were analyzed regarding the changes in their weight; sorption ability; tensile and flexural strength characteristics; colour and gloss; and chemical changes (investigated by means of FTIR). Considerable chromatic changes were observed, especially owing to the UV radiation. Although some reduction in the mechanical properties was observed, particularly in the immersion and condensation chambers, the durability tests proved a generally good behaviour of this material under the aggressive conditions considered.

REFERENCES

Kharbari, V.M., Chin, J.W., Hunston, D., Benmokrane, B., Juska, T., Morgan, R., Lesko, J.J., Sorathia, U., and Reynaud, D. 2003. Durability Gap Analysis for Fiber-Reinforced Polymer Composites in Civil Infraestructure. Journal of Composites for Construction 7(3): 238–247.

83 Harries, K.A, Porter, M.L., and Busel, J.P. 2003. FRP materials and concrete – Research needs. Concrete International: 69–74. Correia, J.R., Cabral-Fonseca, S., Branco, F.A., Ferreira, J.G., Eusébio, M.I., and Rodrigues, M.P. 2006. Durability of pultruded glass-fiber-reinforced polyester profiles for structural applications. Mechanics of Composite Materials 42(4): 325–338.

84 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Analytical study on the performance of reinforced high-strength concrete bridge columns

D.J. Seong, H.M. Lee & H.M. Shin Department of Civil and Environmental Engineering, Sungkyunkwan University, Suwon, Kyonggi, Korea J.H. Choi Department of Civil Engineering, Hankyong National University, Ansung, Kyonggi, Korea M.S. Oh Department of Structure, Seoyeong Engineering Co., Ltd., Seoul, Korea

The objective of this study is to evaluate the performance of High-Strength Concrete (HSC) bridge columns analytically. The reliabilities of several stress-strain confinement models for HSC are compared with experimental data reported in the literature. In this paper, an analytical model for inelastic behavior and ductility capacity of confined HSC columns under reversal cyclic loading are presented. A general purpose finite element analysis program, implementing the reinforced concrete plane stress element is used for this purpose. The lateral confinement effect and the stress-strain relation of HSC are accounted for in the proposed analytical model. Based on the experimental data in the literature, nonlinear analysis results of HSC and Normal-Strength Concrete (NSC) columns subjected to reversal cyclic loading are evaluated. HSC offers many advantages such as higher material strength, less creep, and improved durabil- ity. The use of HSC leads to reduction in dead weight, economic benefits from reduced section size. However, HSC tend to behave more brittle than conventional NSC. This brittle response

Figure 1. Geometric details of reinforced concrete column specimens. Figure 2. Loading schedule for lateral load.

85 and relatively scarce experimental data available for HSC have apparently set some limitations on its widespread use. The capacity of NSC in compression can be calculated using an equiva- lent rectangular stress block (ERSB) model, which is not appropriate for the stress-strain relation of HSC, literatures have reported that the use of current ACI stress block parameters results in overestimations for HSC. The brittle response of HSC column can be improved by adequate lateral reinforcement, however, it has been recognized by many researchers that confinement effectiveness and ultimate strain at post-peak of columns decrease with increase in concrete strength. A widely accepted reliable and practical model for confined HSC columns is still needed. A method for numerically predicting the post-peak behavior and ductility capacity of reinforced concrete columns of HSC under reversal cyclic load is proposed.

86 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Use of steel fibre concrete to eliminate shear reinforcement in pretensioned concrete beams

P. De Pauw & L. Taerwe Magnel Laboratory for Concrete Research, Department of Structural Engineering, Faculty of Engineering, Ghent University, Ghent, Belgium N. Van den Buverie & W. Moerman Willy Naessens Industriebouw nv., Wortegem-Petegem, Belgium

ABSTRACT: The precast concrete industry shows an increased interest in the structural possibil- ities of steel fibre concrete. Diffused steel fibre reinforcement could be used to reduce conventional shear reinforcement like stirrups in structural concrete elements and may result in a reduction of labour costs. In an experimental test program, the behaviour of two precast pretensioned concrete beams made with steel fibre concrete and without conventional shear reinforcement is compared with the behaviour of a standard beam made with concrete without fibres but with stirrups as shear reinforcement. Furthermore, a beam made of plain concrete without shear reinforcement is tested to investigate the effect of the shear reinforcement. The beams are designed according to Eurocode 2 and, especially for the beams made with steel fibre concrete, according to the guidelines of the “σ − ε − design method” mentioned in the recommendation from the RILEM TC 162-TDF technical committee and according to information found in literature. Beside the four beams, several small scale test specimens are used to determine the compressive strength, the secant modulus of elasticity, the shrinkage and creep behaviour and the cracking and post-cracking behaviour of the concrete. Special attention is given to the concrete mix design, the mixing operation and the casting procedure for the steel fibre concrete. Both fibre concrete types, one with 40 kg fibres and one with 60 kg fibres per cubic meter of concrete, have to pas through the small web area and in between the closely spaced prestressing strands in the bottom flange of the beams. The use of poke vibrators is not allowed because they might influence the orientation of the steel fibres in the concrete. The correct filling of the formwork and the adequate compaction of the concrete depend on the workability of the concrete and on the action of the formwork vibrators. The small scale tests show that different bending tests lead to different values for the flexural tensile strength at first cracking. The loading tests on the four beams show that traditional shear reinforcement can be eliminated by using steel fibre reinforced concrete. The beams with the steel fibre concrete show similar shear failure loads as the reference beam with the stirrups. The addition of 60 kg of fibres instead of 40 kg of fibres on the other hand does not give an increase in shear capacity. This might be due to a lesser compaction of the concrete with the higher volume fraction of fibres. The conclusion of the tests is that ordinary shear reinforcement can be replaced by steel fibre concrete in precast pretensioned beams if the necessary attention is given to the mix procedure and casting procedure of the steel fibre concrete.

87 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Durability of Structural Composite Lumber in bridge applications

Nur Yazdani Department of Civil & Environmental Engineering, University of Texas at Arlington, Arlington, USA Eric C. Johnson Bridge Engineer, Figg Bridge Engineers, Tallahassee, Florida, USA Sheila Duwadi Team Leader, Bridge Safety, Reliability and Security, Federal Highway Administration, McLean, Virginia, USA

ABSTRACT: Structural Composite Lumber (SCL) is a family of newly engineered wood products finding increasing use in highway bridge applications. The advantages of SCL are high strength, flexibility of sizes and shapes, stiffness and excellent treatability with preservatives. One of the main concerns in SCL bridge applications is the effect of moisture, exposure to ultraviolet light and varying temperature/humidity effects on the long-term durability of the bridge members. In order to document these effects, monitoring of full scale SCL T-beam bridge members was performed under ambient conditions and accelerated aging process. A total of 16 beams were monitored under exposed weather conditions with frequent wetting and drying. Variables in the experiment were: lumber type (Douglas Fir and Southern Yellow Pine), SCL type (LVL and PSL), and preservative type (CCA and Penta). The moisture content of the SCL, ambient temperature, humidity and the general condition of the beams were monitored. The experiment concluded with the determination of an applicable SCL member for bridge applications. Although most of the members were found to be adequate, a Douglas Fir CCA treated LVL member was found to be most suitable.

Keywords: Structural Composite Lumber (SCL), lumber type, SCL type, preservative type

88 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Development of Self-Consolidating Concrete for bridge construction and repairs

P. Paczkowski, A.S. Nowak, G. Morcous & M. Kaszynska University of Nebraska, Lincoln, Lincoln, USA

For cast-in-place applications, the Self-Consolidating Concrete (SCC) has to satisfy the basic requirements such as flowing ability, passing ability and resistance to segregation, in addition to pumpability, for delivery times up to 90 minutes. These properties depend on the type of materials used, mix design (e.g. water/cement ratio and admixture content), placing temperature and delivery time. This paper presents the development of SCC mix recipe and testing procedures to confirm ability for on-site bridge applications. Materials include 1PF cement mixed with fly ash type F, 47B sand and gravel, with the maximum aggregate size of 12.5 mm (0.5 in), and admixtures. The reduced aggregate size helped to increase resistance to segregation and eliminated problems with blocking. It was found that for the on-site assessment of quality for the mix, it is sufficient to perform only J-ring test and Slump Flow test with the Visual Stability Index. The performance of High Range Water Reducers (HRWR)was investigated for different delivery times and impact of the Viscosity Modifying Admixture (VMA) on mortar flowability. It was found that the addition of VMA has to be compensated with an extra dosage of the HRWR. Trials on mortars also showed that high dosage of VMA can significantly reduce the workability time. Slump flow retention curves were obtained for the selected SCC mixes. Two different types of retarding admixtures were checked for their ability to extend the workability time. The effect of Air Entraining Admixture (AEA) on air content and compressive strength was investigated. The analysis of delivery time effect on SCC properties showed that for on-site applications a retarder should be used and, if needed, an additional dosage of the HRWR. However, the amount of retarder should be carefully selected because an excessive amount does not increase workability time and it only affects the setting time. The selected SCC recipe was tested on four walls (1. m high), each one with the mix poured at a different delivery time. This test also checked the passing ability of the mix as the minimum spacing between bars was 1.9 cm. In another test, the mix was pumped for over 120 m to verify the pumpability and its impact on the fresh concrete properties. Air content in concrete is currently the subject of intensive research. Comparison of the results obtained using the ASTM 231 Pressure Method and the Modified Point-Count Method ASTM C457 shows a good agreement. Therefore, the Pressure Method is recommended on-site for a fast assessment of the air content because of simplicity. The plastic air content can cause a reduction of compressive strength of concrete, and therefore, practical quality control procedures are needed. Laboratory tests showed that it is possible for the mix with admixtures to maintain SCC properties up to 70 minutes. If after that time the spread decreases below acceptable limit, the flowability can be recovered by an additional dosage of HRWR. On-site pilot tests showed that the SCC mix remains pumpable even in highly elevated temperatures.

89 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

On the use of duplex stainless steels in bridge construction

O. Hechler ArcelorMittal Commercial Sections S.A., Technical Advisory, Luxembourg former RWTH Aachen University, Lehrstuhl für Stahlbau und Leichtmetallbau, Gemany P. Collin Lulea University of Technology, Division Structural Engineering, Sweden

ABSTRACT: Maintenance and the related costs in the life time of existing structures are an ever growing problem and bridges are no exception to this rule. Much of the maintenance costs are related to corrosion of steel members. Especially when steel and composite bridges are built in aggressive environments (e.g. close to seawater or in industrial atmospheres) these costs become of major importance. However also concrete bridges are concerned as the re-bars in concrete decks may be subjected to corrosion caused by chloride ingress from de-icing salt or seawater. Consequently the use of a material providing high mechanical properties with a sufficient cor- rosion resistance for the application in these aggressive environments would lead to economic and durable construction. Further the maintenance costs become predictable, which would increase the quality of the bridge management. Duplex stainless steel is this material with sufficient yield strength, stiffness and toughness as well as an excellent durability in aggressive environments. With a targeted material choice, innovative construction solution may be achieved. However, high material costs, gaps in common knowledge, e.g. on fatigue, as well as fabrication aspects are preventing a wider application of these steels in bridge construction. In this paper an introduction on the requirements on construction in aggressive environment is shortly given. In reference to them alternative approaches with the use of stainless steel reinforcement or stainless steel plates are presented.

Further design guidance for the fatigue design of critical bridge details made in duplex stainless steel are given in the paper. These results close a gap in current standardization as up to now it is not possible to design a bridge using duplex stainless steels according to the Eurocodes. 90 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Self-consolidating lightweight concrete – Excellent material for bridge applications

M. Kaszynska Szczecin University of Technology, Poland

The latest trend in concrete technology is the development of Self–Consolidating Lightweight High- Performance Concrete. This concrete combines the favorable properties of lightweight concrete (LWC) with those of Self-Consolidating Concrete (SCC). The development of light-weight self-consolidating concrete is very difficult in spite of a consid- erable experience related to self-consolidating concrete. Comparison of these two types of concrete indicates that light-weight self-consolidating concrete has characteristic features resulting from use of light-weight aggregates. High variation of absorption properties of the light-weight aggregates due to its porous structure is a significant problem in the mix design. Water absorption by the light-weight aggregates pores can lead to a loss of self-consolidating properties of concrete. Due to a considerable difference between volumetric densities of the light-weight aggregates and the surrounding cement matrix, course aggregates has a tendency to surface if the cement paste does not have an adequate viscosity, and furthermore, these concretes have a stronger tendency to seg- regation of the ingredients compared to ordinary self-consolidating concrete. This is one of the major problems in the development of the required properties. There are various techniques available to avoid negative consequences of absorption of water needed for hydration by the light-weight aggregates. The most common method is soaking in water. It is particularly effective in case of high-performance light-weight concretes, because due to low water-cement ratio there is a possibility of autogenous shrinkage. Another method is application of a thin protective layer (film) of cement paste on the aggregates to close the pores in the aggregates and block the access of water while maintaining the bond between the aggregates and cement matrix. The objective of the presented research is to develop a light-weight self consolidating con- crete using local light-weight aggregates. Various self-consolidating mixtures were considered by replacing a part of fine and coarse aggregate with light-weight aggregates Pollytag. Pollytag is a light-weight aggregate obtained by sintering fly ash in temperature of 1000◦C −1350◦C. In the tests, Pollytag size of 0–4 mm and 6–8 mm were used. Three aggregate conditions were considered: dry, pre-wetted or coated in a cement paste. The best workability properties were obtained for the mixes with aggregate coated in a cement paste and the worst for the mixes with dry aggregate. How- ever, the technology of coating the aggregates in cement paste is time-consuming and expensive, therefore, the it is recommended to use pre-wetted aggregate. The obtained light-weight concrete is featured with good self-consolidating properties, the den- sity lower than 2000 kg/m3 and strength from 55 to 75 MPa. The tests showed that initial preparation procedures of light-weight aggregate do not have significant influence on 28 days compressive strength of concrete.

91

Assessment and control of bridge vibrations

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Detection of bridge damages by recognition of non-linear dynamic effects

H. Wenzel VCE Holding GmbH, Vienna, Austria

ABSTRACT: Damages of structures are expressed by a non-linear behavior that can be recognized from monitoring records. The well established procedures of ambient vibration monitoring are able to produce the necessary data sets. Nevertheless it has been experienced that these effects are well hidden in the data. Several tests with permanent monitoring stations have shown that they appear clearly under certain circumstances, which are rare events. This suggests that a very accurate damage assessment goes along with a sufficient length of monitoring. The necessary compensation of environmental factors and external load influences also depend on the knowledge of their values and their development over time. This contribution shows examples where damages have been successfully detected through per- manent ambient vibration monitoring campaigns. It comprises steel, concrete, composite bridges as well as interesting structures like high-rise chimneys or wind power generator shafts.

1 INTRODUCTION

Previous measurements at the Europabrücke matched very well with the comparative analytical calculations, but they also exhibited the remarkable level of intensity of the loading impact. Cur- rently the bridge is stressed by more than 30000 motor vehicles per day (approximately 20% freight traffic). The superstructure is represented by a steel box girder (width = 10 m; variable height along the bridge-length 4,70–7,70 m) and an orthotropic deck and bottom plate. This motorway bridge with six spans differing in their length (longest span 198 m, supported by piers with an elevation

Figure 1. Europabrücke – overview.

95 Figure 2. The permanent monitoring system and its several measurement sections. of 190 m) and a total length of 657 m comprises six lanes, three for each direction distributed on a width of almost 25 meters. To reach the already defined goals, a permanent monitoring system has been developed in a stepwise manner (Fig. 2). It consists of 24 measuring channels (sampling rate 100 Hz) representing the main span’s, the pier’s and the cantilever’s accelerations, the abutment’s dilatation, wind speed and direction, and temperatures at several locations.

2 METHODOLOGY

The developed methodology delivers a standardized kind of comparison. The utilization of consid- ering the varying length of the torsional bracings in the uphill and the downhill driving direction – automatically ensures that each expected value of the structural member’s dynamic response includes the same uncertainty of modelling. The present approach intends to provide a horizontal progression of the calculated ratio between measured and expected values for eigenfrequencies under regular conditions. Probable deviations from this trend line are much better recognizable than from the progression of the measured values itself. This fundamental assumption of the methodology is confirmed when the progression of the calculated ratio between each of the three functions for expected values is compared to the other ones. Analogous to the comparison of the individual functions of expected values to each other, the comparison between the measured and the expected values of the bracing’s effective stiffness was done.

96 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Experimental dynamic analysis of steel concrete composite railway bridges: The Sesia viaduct on the high speed line Turin-Milan

G. Chellini, L. Nardini & W. Salvatore Department of Structural Engineering, University of Pisa, Italy G. De Roeck, K. Liu & E. Reynders Department Civil Engineering, K.U. Leuven, Belgium B. Peeters LMS International, Leuven, Belgium M. Tisalvi & G. Sorrentino RFI S.p.A, Roma, Italy

In the development of new high speed railway lines across Europe, steel-concrete composite bridges have become an important alternative to concrete structures due to considerable advantages regarding the design, construction time, durability and costs. The recently launched research project “DETAILS” (funded by European Commission) aims to deeply analyze dynamic effects and interaction phenomena, fatigue loadings, structural modelling, fatigue life and damage assessment of these types of bridges, removing actual uncertainties that still concern such structures. As part of the experimental work in the project, the Sesia Viaduct, located on the new Italian high speed line between Turin and Milan near Novara, was extensively tested. The viaduct is a box girder steel-concrete composite bridge formed by seven simply supported spans each about 46 m long for a total length of 322 m (Figure 1). In the experimental campaign, three spans of the bridge were instrumented by accelerometers covering 103 sensor locations and by optic fibers, in order to obtain detailed mode shape information for correlation and updating of Finite Element models and to check the real vibration level under train passages. The preliminary results of modal identification performed by Operational Modal Analysis techniques (namely Stochastic Subspace Identification and PolyMAX) revealed that the bridge presents a certain level of coupling between the spans (“statically” separated) showing modes whose geometry is repeated simmetrically or anti-simmetricaly along the spans. Nevertheless, under train passages only some of the modes participated to bridge vibrations. The paper discusses the acquired data under ambient and train excitation conditions and presents the Operational Modal Analysis results.

Figure 1. Panoramic view of the Sesia viaduct.

97 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Excitation of pedestrian structures during marathon events

C. Sahnaci & M. Kasperski Department of Civil Engineering, Ruhr-University Bochum, Germany

Mass events are an effective method for advertising a location or a city. In recent years, especially marathon events have become more and more popular. Both, the number of events and the number of participants in the single events are clearly increasing. The courses are often designed to pass attractive places like parks, monuments and impressive buildings and structures. If the course is intended to cross a bridge, the question arises, whether the large number of runners may lead to vibrations which may lead to disturbances for both, the runners and the spectators. For both, the passive and the active users, reasonable threshold values have to be specified to avoid serviceability problems during the event. These threshold values have to be compared with predicted vibration amplitudes, which can be obtained by simulating the induced actions due to the random flow of runners and calculating the structural responses. As input, two sub-sets of information are required. The first sub-set describes the arrival rate and the running velocities during the different phases of the race. These values can be obtained from electronic measurement systems used during marathon events. Based on a transponder chip the competition times for each individual runner can be tracked automatically for several sub-sections of the course. The paper demonstrates in detail the statistical analysis of the chip data leading to the conclusion that the probability density of the velocities shows a considerable skewness. As appropriate model, the Weibull distribution is recommended. Since the mean running velocities for male and female runners are considerable different, separate statistics are performed. The second sub-set provides the information on the basic running parameters step frequency and step length and their interdependency. Based on experiments, the interdependency of the velocity and the step frequency for the individual runner can be reproduced by a simple linear model with the describing parameters slope and intersection. These two parameters have to be understood as random parameters, i.e. they differ between the individuals. The paper presents a first tentative model for the interdependency of slope and intersection based on 14 individuals. While the loads induced by walking have been intensively studied in recent years, sound and reliable data for the loads induced by running are scarce. Therefore, at this stage, only a very simple approach can be used, which is based on the recommendations in ISO 10137. The load amplitude is obtained as a random value following a normal distribution with a variation coefficient of 15%. The set of parameters are applied in a simulation for the responses of a bridge with a span of 100 m. The structure is assumed to be a steel-concrete composite structure. The static system of the bridge is a simple beam, the natural frequency of the bridge is assumed as 2.8 Hz. The bridge is thought to occur in a marathon course at different positions from 5 km to 40 km. For each runner, the gender-dependent individual parameters body weight, velocity and step frequency are simulated. The mean value, the 95%- and the 99%-fractile values of the experienced vibrations are calculated. These values are compared to the largest bridge accelerations and the serviceability criterion for walking given in ISO 10137. The paper shows that the number of persons on the bridge is not a direct indicator for the vibration amplitudes. The simulations demonstrate that vibration problems can be reduced by positioning the bridge rather at the end of the course.

98 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

A novel active mass damper for vibration control of bridges

U. Starossek & J. Scheller Structural Analysis and Steel Structures Institute, Hamburg University of Technology, Germany

Slender structures such as high-rise buildings and long-span bridges are susceptible to dynamic loads. In addition to earthquakes, wind is a major source of such loads. The effects of the resulting vibrations may be disastrous. An important example of such a phenomenon is bridge deck flutter. It is crucial to take into account this aeroelastic instability during the design stage, and the installa- tion of stiffer girders may be required. As an alternative to changing the structural system, passive or active control has proven to effectively suppress structural vibrations. Although both are imple- mented in some high-rise buildings, to date only passive control has been used in bridge structures. The reason for the rare application of active control is the difficulty in reliably generating adequate control forces with low power demands. Furthermore, the additional weight resulting from the active device needs to be kept small. In accordance with these requirements, a novel active mass damper for vibration control of structures was developed. The damper can utilize centrifugal as well as tangential forces induced by rotating unbalanced masses. In this study, the new damper and the possibilities of force generation are presented. The main advantage of the new active mass damper lies in the low power demand. Active mass dampers for bridge vibration control described in the literature generate the control forces by means of accelerating auxiliary masses. For example, the generation of a twisting moment by variable eccentric weight or by changing the rotational speed of a mass is proposed. Another possibility is to make use of the gyroscopic effect of a high speed rotating mass. In this case, the spin axis of a gyroscopic wheel is tilted and accelerated by an actuator; due to the gyroscopic effect, control moments are induced. For the acceleration of the masses, energy is consumed. In contrast to the above-mentioned systems, in the preferred configuration, the control forces in this novel damper are generated with a constant motion, with no acceleration being needed for their generation. Thus, a very low power demand is achieved as demonstrated using a numerical example for bridge deck flutter control.

99 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Analysis of fuzzy active control for traffic-induced vibration of highway girder bridge

Mitsuo Kawatani, Yasutoshi Nomura, Chul-Woo Kim & Yusuke Otsubo Graduate school of Engineering, Kobe University, Japan

ABSTRACT: According to demands on cost-effective bridge systems causing light and simple bridge structures and the increase of weight and traffic volume of heavy vehicles in recent years, undesirable vibrations, so-called environmental vibration such as the low frequency sound and the ground vibration due to traffic-induce vibration of bridges, have been one of technical issues in major cities of Japan. Those environmental vibrations usually link with dynamic serviceability of bridges. Therefore it is extremely important to reduce the undesirable vibration attributable to moving vehicles in order to improve the serviceability of bridges. Dynamic responses of highway bridges due to moving vehicles are a kind of transient vibration. It is generally difficult to control this kind of traffic-induced vibration of bridges, especially for simply supported bridges, by a passive control such as TMD (Tuned Mass Damper), even though there are some examples of applying TMD for traffic-induced vibration of bridges. This study is an attempt to reduce the traffic-induced vibration of a simply supported highway bridge such as girder vibration and dynamic reaction force which causes ground vibrations by active control through a numerical analysis. Fuzzy theory is adopted as a control rule for the active control. The implementation of fuzzy controllers makes use of linguistic synthesis, and therefore they are not affected by selection of a specific mathematical model. This study presents a feasibility investigation of fuzzy active control for the traffic-induced vibration of a highway bridge through a numerical analysis. Effectiveness of the fuzzy active control is examined by comparing with the control effect by the Tuned Mass Damper (TMD). The vertical acceleration response, Fourier spectrum and a dynamic response of the reaction force at the pier base of the bridge under a single moving vehicle with 30 km/h are compared with those of without control as shown in Figures 1 and 2. In addition, the peak response, RMS (Root Mean Square) value, and reduction effect estimated from ratio of controlled acceleration responses to uncontrolled ones are summarized in Tables 1 and 2. It is observed that vertical acceleration responses under the fuzzy active control are reduced effectively in comparison with those of without control. As for RMS values, the fuzzy active control provides about more than 50% reducing effect. Fourier spectrum shows that the peak at 2.32 Hz is decreased. TMD of this study can hardly reduce the peak and RMS values of the acceleration response. Fourier spectrum, however, shows that the Fourier amplitude at 2.32 Hz is also decreased by theTMD. It demonstrates that theTMD is effective to a specific mode (frequency). It is also observed that a dynamic response of the reaction force of the pier base of the bridge under the fuzzy control is reduced effectively in comparison with those of without control as shown in Figure 1, because inertia force by the vibration of the bridge decreased by the fuzzy control. On the other hand the dynamic reaction force does not change due to TMD as shown in Figure 2. Observations throughout this study show that the passive type control TMD is not adaptable to reduce traffic-induced vibration of the simply supported bridge, which usually is dominated by transient vibration due to traffics, considered in this study. On the other hand, the study demonstrates feasibility of the fuzzy active control for reducing traffic-induced vibration of the bridge.It is also observed that a dynamic response of the reaction force of the pier base of the bridge under the

100 Figure 1. Dynamic response with and without fuzzy active control.

Figure 2. Dynamic response with and without TMD.

Table 1. Summary of acceleration responses and vibration reduction effect according to control methods.

Reduction effect (control/w/o control)

PEAK (m/s2) RMS (m/s2) PEAK RMS

w/o control 0.291 0.073 active control 0.124 0.026 0.43 0.36 TMD 0.279 0.066 0.96 0.90

101 Table 2. Summary of reaction force and reduction effect according to control methods.

RMS (m/s2) Reduction effect (control / w/o control)

Inertia Dynamic Dynamic Inertia Dynamic Dynamic force wheel load reaction force force wheel load reaction force

w/o control 0.741 0.285 0.767 active control 0.380 0.297 0.473 0.513 1.04 0.62 TMD 0.648 0.284 0.669 0.874 0.996 0.87

fuzzy control is reduced effectively, because inertia force by the vibration of the bridge decreases by fuzzy control.

102 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Sensitivity-based optimal design of damper connecting system for vibration control of parallel bridges under wind excitations

D.-S. Kim Seoul National University, Seoul, Korea S.-Y. Ok University of Illinois at Urbana Champaign, USA K.-S. Park Dongguk University, Seoul, Korea H.-M. Koh & C.-Y. Choi Seoul National University, Seoul, Korea

To prevent pounding between two adjacent structures, the method of connecting adjacent structures with dampers was first proposed by Klein et al. (1972). Subsequently, extensive studies on dissi- pative links, active and semi-active devices have been carried out for mitigating wind-induced or seismically-excited responses of the neighboring structures. However, most of the previous studies mainly focused on the building structures. For bridge structure, it has been widely recognized that longitudinal and vertical vibrations are more menacing to structural safety. Accordingly, a lot of studies concentrated on the vibration control problem in the longitudinal and vertical directions. However, as the span length of a bridge becomes longer, the transverse-mode behavior could be dominant which possibly causes serious vibrations of the bridge in the transverse direction. Of particular interest in this study is the vibration control of two adjacent long-span bridges in the transverse direction. For the neighboring bridges, damper-interconnected approach is expected to account for the transverse vibration in more efficient manner (Ok et al. 2006). In order to decide optimal distribution of the connecting dampers, we first introduce a sensitivity-based measure to represent the contribution of the damper capacity to modal damping ratio. This sensitivity measure thus accounts for the effectiveness of the dampers to the increase in the modal damping ratio. Under the assumption that the distribution of the damper capacities is proportional to the sensitivity measure, then, the optimization process is performed by minimizing the peak value of the frequency responses of the damper-connected bridge system. To demonstrate the proposed optimal design method, two adjacent cable-stayed bridges in Korea (Jindo grand bridge I and II) are considered and the commonly-used linear viscous damper is employed as inter-connecting control device. Numerical analysis is performed to evaluate the per- formance of the optimal damper system designed by the proposed method and its performance is compared with those of uniformly distributed damping system and uncontrolled system. Compar- ative results show that the sensitivity-based optimal damping system distributes the dampers in an efficient manner so that the total amount of the dampers can be considerably saved without any significant loss of response control performance. In order to validate the performance of the pro- posed system, numerous time-history analyses are carried out using a set of wind velocity records simulated by spectral representation. The illustrative example shows that the sensitivity-based damper distribution system can be both efficient and economical in controlling the wind-induced vibration of two adjacent bridges.

Keywords: damper connecting system, sensitivity-based optimal design, wind excitation, parallel bridges. 103

Bridge 200 toward durable bridge

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Research project “Bridge 200”

B.S. Kim, S.Y. Kim, K. Cho, J.R. Cho & S.T. Kim Korea Institute of Construction Technology, Goyang, Korea

The “Report of Bridge Stock in 2006” published by the Korea Ministry of Construction and Trans- portation reported that the bridge stock in Korea increased continuously since mid of 1990s to reach today 23,805 bridges for a total developed length 2,139 km, as of December 21, 2006. The enact- ment of a special law related to the maintenance of facilities after the collapse of Sungsu Bridge in 1994 made it possible to establish systematic planning of safety management and management for bridge structures. However, the service life of bridges in Korea ranging between 30 to 50 years remains relatively shorter than the 50 to 75 years re-ported in advanced countries that are 50 years in Japan, 70 years in UK and USA, 75 years in Finland and 76 years in Sweden. If bridges are reconstructed at the frequency adopted today, tremendous costs to be poured directly for their construction together with degradation of the quality of life can be easily forecasted. Moreover, considering that the reconstruction of existing bridges, differently from the construction of new bridges, requires the installation of deviation roads as well as social economic losses brought by the blocking of traffic, the shortened lifespan of bridges is leading to astronomical national expenses. Being critical infrastructures in the road network, bridges are used by citizens. Any problem occur- ring in bridge will thus have direct impact on the life of the population and will increase goods transportation costs. Therefore, advanced countries are investing huge budgets for bridges as a national asset strengthening the national competitiveness. This appears in the form of large gov- ernmental investment in R&D related to the extension of the lifespan of bridges and the reduction of maintenance costs. Since the mid of 1990s, these countries implemented actively researches on innovative bridge technology extending significantly the lifespan of existing and new bridges by adopting advanced materials like composite materials or high-performance materials enhancing the properties of former construction materials that are steel and concrete together with reduced maintenance costs, which result finally in reduced lifecycle cost (LCC). Accordingly, considering the short lifespan exhibited by the bridges in Korea, implementing R&D devoted to the extension of the service lifetime of bridges and minimization of maintenance constitutes a necessity and priority regard to our national competitiveness as well as the economy of the national budget. The research project “Bridge 200” as a core R&D was launched in 2002 at the Korea Institute of Con- struction Technology and was managed during 5 years to secure practicable technology extending the lifespan of bridge up to 200 years if required.

107 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Development of FRP bridge decks in Korea

Ki-Tae Park & Young-Ho Lee Korea Institute of Construction Technology, Korea Jinwoo Jeong Construction Technology International Inc, Korea Yoon-Koog Hwang Korea Institute of Construction Technology, Korea

Due to the vehicle loading and environmental attacks, deterioration of the bridge decks is usually accelerated. Bridge deck often requires repairs and deteriorated one will eventually be rehabilitated or replaced. For these reasons, service life of the conventional bridge deck is many times shorter than those of other bridge components. Therefore, the use of high-strength and high-durability bridge decks in the highway bridge construction becomes very crucial to minimize the maintenance during the service and to increase the service life of bridges. Since the conventional reinforced concrete deck is heavy and requires long erection period, bridge engineers seek new materials to decrease the overall dead load of the bridge and to shorten the erection period. The FRPs are relatively new materials in bridge construction. With high strength to weight ratios, excellent durability, and low life-cycle costs of FRPs, FRP bridge decks can offer low dead load, minimum maintenance, and long service life. Due to the lightweight of FRPs, if the existing concrete deck can be replaced with FRP deck, the load carrying capacity of the superstructure can be increased without strengthening the girders. In this paper, the brief summary of the developed FRP bridge deck in Korea are discussed. To verify the developed FRP bridge deck, fiber direction flexure test, transverse direction test, and the buckling test were performed to examine the basic performance of the developed FRP module for decks. In the case of fiber direction flexure test, it was realized that the test values of the failure load are 1.3∼1.4 times larger than the analysis value. It was also observed that the buckling safety factor was above 5 and the failure load of FRP deck to be three times safer than axial load of design truck load DB-24 as specified in the specification. Based on the test results which are performed in this paper, it was realized that it was safe to design using the modeling technique applied to this analysis because generally, the test values were larger than the analysis value, and our optimum design algorithm is valid. Also, there were a variety of tests conducted to evaluate the durability of composite materials of FRP bridge deck. It can be concluded that, when the defects such as a decrease in flexural strength caused by chemicals can be complemented, bridge deck members made with FRP have a tendency to spread.

108 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Development of FRP-concrete composite bridge deck system

S.Y. Park, K. Cho, J.R. Cho, S.T. Kim & B.S. Kim Korea Institute of Construction Technology, Goyang, Korea

The bridge deck is the most potential member for damages, because it is against direct actions of wheel loading, freezing-thawing and drying-wetting. Therefore, the average life span of the deck is known as only about 25 years, which is quite shorter than that of the bridges (50 ∼ 70 years). Live load from the overweighted vehicles causes the severe fatigue damage to the deck. In addition to damage susceptibility, most of the current decks are constructed using in-situ concrete with the scaffolding, which delays the construction of the deck. To overcome these problems, Korea Institute of Construction Technology (KICT) had strived huge effort to develop new bridge deck systems increasing life span of bridge deck. The Fiber Reinforced Polymer (FRP)-concrete composite deck is one of them. This deck system is a new-type FRP-concrete composite deck, which is made by pouring concrete onto FRP panel fabricated in the factory. The FRP panel plays roles of a formwork under construction and a main tensile member in service life. The deck on developing has the advantage of corrosion free characteristics and is more durable than conventional decks. The use of hollow FRP section makes it lighter, and the construction of it can be more easily done because there is no need of additional formworks. Compared to FRP deck, deflection design criterion can be easily satisfied because of high stiffness of it, and there is no problem in local buckling. The structural system of the FRP-concrete composite deck (FCCD) was developed, and then its design criteria were established. Sand coating and perfobond shear connectors were applied for the composition of FRP and concrete. Details of the interface between FRP and concrete were decided based on the results of several tests, and the local bond-slip model was obtained from the test results related to sand coating. The level of stress occurring actually at the FRP-concrete interface in FCCD under service load was evaluated. From the results, the resistance strength appeared to be larger by 2.33 and 2.73 times compared to the bond strength required respectively in the transverse and longitudinal directions under service load state. Shear performance tests were performed to verify the intensity of the composition of FRP and concrete. As a result, the mean shear stress at maximum loading state was 2.075 MPa, which is more than 3 times larger than required. Full-scale tests on deck specimens were carried out to evaluate performance needed as a bridge deck. The ultimate strength of the deck specimen reached 800 kN, which was not much different from that of the reinforced concrete deck specimen with similar test conditions. The deflection under service load showed 1.86 mm, which satisfied sufficiently the serviceability. Moving wheel loading tests revealed that the fatigue strength of this new bridge deck guarantees a minimum of 4,200,000 loading cycles from wheel loads, which also satisfied sufficiently the fatigue limit of 2,000,000 loading cycles. Consequently, this innovative FRP-concrete composite deck system developed by KICT is seen to offer improved constructability and durability compared to former decks, and to present sufficient practicability as a bridge deck.

109 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Development of Ultra High Performance Cementitious Composites (UHPCC) in Korea

S.W. Kim, J.J. Park, S.T. Kang, G.S. Ryo & K.T. Koh Korea Institute of Construction Technology, Goyang, Korea

Ultra High Performance Cementitious Composites (UHPCC) is an ultra-high strength, high dura- bility and high toughness concrete. Most types of UHPCC are composed of a powder containing cementitious materials and very fine sand with a maximum particle size of 600 µm. UHPCC contains no coarse aggregate and the entire material of UHPCC is composed of very fine parti- cles. Various types of UHPCC, as illustrated by examples of RPC (Reactive Powder Concrete), ECC (Engineered Cementitious Composite), and CRC (Compact Reinforced Composite) have already been developed by other research groups. However, application of UHPCC as construction materials has not been widespread. Local materials for UHPCC are sourced from areas near the construction site and have some different properties. It is, therefore, necessary to modify the mix proportions of UHPCC according to local materials and to take into account the effect of this on the mechanical properties of UHPCC. In order to develop an UHPCC suited to the Korean conditions, several mechanical tests were carried out. The material parameters included types and quantity of filling powder, amount of poz- zolanic material, quantity of superplasticizer and fiber volumetric ratio. Compressive and bending strength were tested to find the optimal mix proportions of UHPCC. This material is a structural material exhibiting very remarkable mechanical performances with compressive strength, tensile strength and flexural strength rising up to 200 MPa, 10 MPa and 35 MPa, respectively. In addition, this material presents exceptional durability regard to the very low diffusion and penetration speeds of noxious substances like chloride ions. However, its shrinkage being relatively larger than ordinary concrete, careful attention should be paid when used as cast-in-place material. This study performed tests to assess quantitatively the autogenous and drying shrinkage characteristics of UHPCC, a steel fiber reinforced cementitous composite developed in KICT and using domestic materials except for silica fume, and to evaluate the shrinkage reduction characteristics of UHPCC according to the introduction of expansive additive and shrinkage reducing agent. The experimental results revealed that, even if the cementitous composite without steel fiber exhibits larger shrinkage than ordinary concrete, shrinkage reduction larger than ordinary concrete is achieved through the adoption of expansive additive and shrinkage reducing agent. Accordingly, if adequate shrinkage reduction measures are taken, the use of UHPCC as structural member will not produce problem regard to shrinkage deformation. Besides, negligible creep was seen to develop after completion of steam curing. These 200 MPa strength concretes have been effectively adopted for the construction of bridges like Sherbrooke Bridge in Canada in 1997, Sunyu Pedestrian Bridge in Korea in 2002, Meata Bridge in Japan in 2003, Sheperds Guelly Creek Bridge, the first ultra-high strength concrete highway bridge in Australia in 2004 and, more recently in 2005, Mars Hill highway bridge in USA in 2005. The construction of structures using ultra-high strength concrete is a worldwide development trend of concrete technology for the construction of advanced facilities in the 21st century. Even if Korea presents today an equivalent level in the domain of ultra-high strength technology compared to advanced countries, now a day relevant R&D shall be imperatively conducted in order to develop structures by using of UHPCC.

110 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Flexural strengthening of RC structures with externally unbonded prestressed CFRP plates

Young Hwan Park, Jong Sup Park & Woo Tai Jung Korea Institute of Construction Technology, Goyang, Korea

Strengthening techniques for concrete structures have been developed and applied in various forms during decades, but most of them are proceeding by adding reinforcements in existing structures or by increasing the sectional strength through enlargement of the section. Of these, Fiber Reinforced Polymer (FRP) sheets or plates being composed of a material offering high strength-to-weight ratio, corrosion resistance and remarkable constructability have recently seen large applications for the strengthening of RC structures. Traditionally, most of the strengthening methods using FRP sheets or plates make use of adhesives like epoxy to bond FRP on the surface of the existing structure. Even if the FRP bond technique is effective in increasing the ultimate strength of the structure, poor improvement of the crack load or yield load is achieved at service load level and practically no control effect on the crack width and deformation is expected due to the principle underlying this technique. In addition, early debonding failure occurs in FRP bond-strengthened structure due to the concentration of stress at the bond interface, which impedes full exploitation of the material efficiency. Such early debonding failure is now identified as the most critical issue in the research field related to FRP bond strengthening techniques, and numerous researchers are striving efforts to find the causes of early failure and provide solutions through experimental and analytical studies. CFRP (Carbon FRP) plate or sheet prestressing system has been proposed to overcome the drawbacks of such FRP bond strengthening technique. Normally, only a portion of the strength of CFRP plates is effective in non-prestressed strengthening applications. By introducing a prestress into the plates, they may be used more efficiently, since the greater portion of their ultimate strength is used. Prestressed CFRP plate strengthening system is regarded as the most advanced technique in the FRP strengthening field enabling not only to exploit sufficiently the material merits of CFRP but also to supplement the problems encountered in the previous FRP bonding techniques through the advantages of the prestressing method. This study realizes the problems of FRP bonding methods by means of experimental and ana- lytical studies using diversified variables and proposes a prestressed FRP strengthening technique as an alternative to improve such problems together with the verification of its efficiency. The con- clusions derived from the strengthening verification tests performed in this study can be arranged as follows. • Bonded FRP or NSM strengthening methods experienced early failure due to debonding failure, which degraded the efficiency of FRP.These methods require strength calculation considering debonding failure during design and need to be improved to increase the efficiency of FRP. • Parametric analysis and static/dynamic test results on the wedge-type anchorage developed for unbonded prestressed CFRP plate strengthening technique revealed that the wedge-type anchorage develops sufficient anchoring performance until tensile strength of CFRP. • The unbonded prestressed CFRP plate strengthening technique appeared to exploit effectively the FRP reinforcement and to be a remarkable method increasing the serviceability and strength of the structure.

111 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Bridge scour countermeasures to minimize bridge failures during floods

Juhyung Lee, Jaehyun Park, Moonkyung Chung & Kiseok Kwak Geotechnical Engineering Research Department, Korea Institute of Construction Technology, Kyonggi-do, Korea

In recent, various efforts have been concentrated to develop a new bridge scour countermeasure for minimizing bridge failures during flood since bridge scour has been the leading cause of the bridge failures. Researches have been conducted (1) to introduce a bridge scour analysis and design system which allows to consider bridge scour characteristics in Korea, which is based on the study of the scour mechanism of bridge foundations and of soil erosion properties; (2) to develop the efficient bridge scour monitoring system and the advanced technology for bridge scour inspection; and (3) finally to suggest the comprehensive countermeasures which can augment the safety of bridges during floods. As a result, a new guide line for bridge scour design and inspection which can consider soil types and erosion properties of soil is suggested. The SRICOS method was verified against full- scale scour measurements through a series of case studies and adopted as one of the scour analysis methods for fine-grained soils. This method suggests that the scour process is highly dependent on the shear stress τ imposed by the flowing water at the soil-water interface. Through tests of soil samples from bridge sites using an Erosion Function Apparatus (EFA), the scour rate z˙ versus the shear stress τ is obtained. Using this relationship and the maximum scour depth zmax equation, a hyperbolic function describing the scour depth z versus time t curve can be developed. A new portable scour monitoring device is proposed to inspect both of bridge scour and foun- dations. This system consists of three parts: installation deck, automatic operating equipment and measuring devices. Field applicability test was accomplished at the Daehwa Bridge to validate the performance of the prototype device and procedures under real world conditions. Through the test,

Figure 2. Scene of preparation for measur- Figure 1. Erosion Function Apparatus. ing pier scour.

112 Figure 3. Scour countermeasure using geobags. it is shown that this new portable scour monitoring system enables to perform the bridge scour inspection easily, quickly, and safely. A new bridge scour countermeasure using geobags and aggregates which is more stable and economical than existing methods is proposed, and its stability was verified through material tests and numerical studies. In addition, the field applicability of the geobag scour countermeasure was verified through the pilot construction and long-term monitoring at the Hwasang bridge located in Kwangwon-Do.

113

Bridge codes

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Distribution of demand in single-column-bent viaducts with irregular configuration in longitudinal direction

Reza Akbari & Shahrokh Maalek University College of Engineering, Department of Civil Engineering, University of Tehran, Tehran, Iran

ABSTRACT: Modern approaches to the seismic design of RC structures and particularly bridges are based on the capacity of the structures to dissipate the energy transmitted by the foundation through inelastic behavior. This capacity is generally taken into account in the design phase by means of a behavior factor (q in EC8 or R in others). When a design approach, based on dynamic linear analysis and a single R value is used for bridges, very high ductility demands are imposed on the stiffer piers with strong geometric irregularities, whilst the more flexible piers tend to remain in the elastic domain. The value of the R factor (and consequently the reduction in the elastic force and the expected ductility) specified in the codes is a function of many parameters, among which the geometric regularity of the bridge appears to be prominent. The regularity issue, though relevant in building design, is considered to be extremely important in bridge design. As a result, the applicability of the R-factor approach to such structures is sometimes questioned. Bridges are often quite irregular structures; with complex geometries, both in plan and elevation. The key issue with bridges is the high ductility demand for stiffer elements. The high ductility demand in those elements may be reduced by either designing them to be stronger, or by increasing the strength and stiffness of the more flexible elements such that the forces attracted by the stiffer elements are reduced. Distribution of gravity load and equivalent longitudinal seismic force in single-column-bent viaducts with irregular configuration is addressed parametrically in this paper through applying the elastic-static demand analysis. The particular RC single-column-bent highway bridge investigated herein has been used extensively as a Reference Bridge (Ispra Bridge), with regular and irregular configurations, in several research programs in the past The height and locations of the piers alter the bridge regularity. The results show that the key parameters influenced the distribution of demands and overall longitudinal stiffness of the viaducts includes: 1. Deck to pier cross-section moment of inertia. 2. Ratio of span length to pier height from span to span. 3. Location or arrangement of the unequal piers (or intensity of irregularity) and especially the stiffest pier. 4. Ratio of the height of the subsequent piers from span to span. 5. Viaduct longitudinal displacement. It was shown that because of the variation in overall stiffness and longitudinal displacement, distribution of demands in viaduct elements varies significantly when equivalent static loads push the structure longitudinally, especially in irregular viaducts. If the viaduct geometry is known, probable critical and plastic hinge locations is determined simply. Moreover, viaduct geometry with minimum demands of the piers is identified with considering the cyclic nature of earthquake action. At the end, it was shown that bending moment and shear force demands in the viaduct piers will be vary with increasing in longitudinal displacement even in completely regular viaducts.

117 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Effective length factor of X-bracing system

J. Moon & H.-E. Lee Korea University, Seoul, Korea K.-Y.Yoon Sunmoon University, Asan-si, Chungnam, Korea

In this study, the elastic out-of-plane buckling load and the effective length factor of X-bracing systems were studied. Points of the intersection of diagonals were modeled as a rigid connection or a pinned connection depending on the connecting method of diagonals. For each boundary con- dition of the point of the intersection, out-of-plane effective length factors of X-bracing systems were derived as a function of the length, forces, and flexural stiffness of the compression and tension diagonal. If the tension diagonal has a sufficient flexural stiffness, the tension diagonal resists completely the out-of-plane displacement. Then the compression diagonal would buckle in the second-mode. For identical diagonals, the second-mode buckling occurs when T/P ≥ 0.57 for the X-bracing system with rigid connection. Similarly, the second-mode buckling occurs when T/P ≥ 1 for the X-bracing system with a discontinuous tension diagonal and T/P ≥ 0.57 for X-bracing system with a discontinuous compression diagonal. Proposed out-of-plane effective length factors of X-bracing systems were compared with the results of previous researchers and those of the finite element analysis. Finally, proposed out-of-plane effective length factor of the X-bracing systems are successfully verified as shown in Figure 1–3.

Figure 1. Variation in k with T/P of X-bracing Figure 2. Variation in k with T/P of X-bracing system with rigid connection. system with discontinuous tension diagonal.

Figure 3. Variation in k with T/P of X-bracing system with discontinuous compression diagonal.

118 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Monitoring system of suspension bridges and the utilization of recorded data

Chihiro Kawatoh, Shigeki Kusuhara, Susumu Fukunaga & Kazuo Endo Honshu-Shikoku Bridge Expressway Company Limited, Kobe, Japan

ABSTRACT: Japan consists of four main islands, Honshu, Hokkaido, Kyushu and Shikou. The Honshu-Shikoku Bridges (herein after, referred as HSB) connects Honshu with Shikoku with three routes. The HSB consist of ten suspension bridges, five cable-stayed bridges, an arch bridge and a truss bridge, including theAkashi-Kaikyo Bridge. Japan is located in typhoon and earthquake prone area; therefore, wind resistant design and seismic design with high standard are required. Various researches had been conducted in designing HSB, and latest research developments were adopted to the design of the bridges. However, the design methods were based on some assumptions, and verification of the design methods is necessary. For the verification of design method and for the maintenance purposes, the monitoring systems to record the bridge behaviors are installed on the Akashi-Kaikyo Bridge. They are seismographs, anemometers, GPS stations, accelerometers, velocity gauges, hygrometers, etc. Out of 10 years data after the completion of the bridge, the following data are reported here. Utilizing the recorded data, deformation analyses of the stiffening girder were carried out and the structural responses were compared with the recorded response. From this analysis, it was found out that the bridge mean displacement well coincides with the analytical result, whereas the bridge dynamic responses are slightly smaller than the analytical results, whose reasons may be due to the safety margins for unknown characteristics of wind force and structural damping. The influence of long term temperature fluctuation on the bridge shape, which are recorded by the GPS stations on the girder. The center of the bridge girder went down in proportion to the temperature rise. The main cables of a suspension bridge are structural elements that play a pivotal role, but they are extremely difficult to replace for repair or maintenance reasons. Consequently, protecting these cables from corrosion is one of the most important tasks of proper suspension bridge maintenance. Dehumidification system has been developed on the basis that almost no corrosion when atmo- sphere around the cable has a relative humidity of less than 60%. And utilizing the data of cable dehumidification system for bridge health monitoring is reported. In March 2001, the Geiyo Earthquake hit the Kurushima Kaikyo Bridges and resulted in the breakage of the center stay rod of the 1st Kurushima Kaikyo Bridge. The Kurushima Kaikyo Bridges were analyzed afterwards using the recorded earthquake motion, which verified the fact that the rod of the 1st bridge was broken in a same manner as design calculation and the rods of the 2nd and the 3rd bridge did not suffer excessive stresses in a same manner as design calculation.

119 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Reliability analysis of composite girder under positive and negative flexure designed by LRFD method

D.K. Shin, J.S. Roh & E.Y. Cho Department of Civil & Environmental Engineering, Myongji University, Yong-in, Korea

The reliability analysis of single span and 3-span continuous composite plate girder and box girder bridges under flexure is performed by applying the Rackwitz-Fiessler technique. The bridges are designed based on theAASHTO-LRFD specification using a newly proposed design live load which is developed by analyzing the traffic statistics obtained from measurement of passing traffics on several Korean highways as well as local roads. Single span bridges considered in the reliability analysis have 20 m to 60 m span length while the longest span length for the 3-span continuous bridges with span arrangement ratios of 1:1.25:1 is assumed to be from 30 m to 70 m. The single span bridges are for the reliability analysis of maximum positive moment section whereas 3-span continuous bridges are selected to analyze the reliability of maximum negative moment section over interior pier. A performance function for flexural failure is expressed to be a function of such random variables as flexural resistance of composite section and design moments due to permanent loads and the live load. For the flexural resistance, the statistical parameters are obtained by analyzing over 16,000 samples of Korean structural steel products using the elastic-plastic material nonlinear analysis of various composite girder sections. Several different values of bias factor, in the range of 1.0–1.2, for the live load moment, are used. Due to the lack of available domestic measured data for the moment by permanent loads, the same statistical data used in the calibration of AASHTO-LRFD are applied. The reliability indices for the flexural failure of composite plate girder and box girder bridges with various span lengths are calculated applying the live load factor in the range of 1.5 to 2.0. Relatively uniform values of reliability index are obtained for different span lengths, positive and negative moment regions, and girder types as long as the same live load factor and statistical parameters for live load are used.

Keywords: flexural resistance, composite section, reliability index, LRFD method.

120 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Flexural design of prestressed high-strength concrete girders

W. Choi, H.C. Mertol, S. Rizkalla & P. Zia North Carolina State University, Raleigh, NC, USA A. Mirmiran Florida International University, Miami, FL, USA

This paper proposes stress block parameters for flexural design of high-strength prestressed con- crete bridge girders. The proposed parameters are intended to extend the current AASHTO LRFD Specifications to include concrete strength up to 124 MPa. Research findings are based on a com- prehensive experimental program that included twenty-one C-shape specimens with High-Strength Concrete (HSC) subjected to combined flexure and axial compression. The specimen configuration allowed triangular compressive strain distribution along the cross-section with maximum strain at one face and zero strain at the opposite face. The selected configuration simulates typical strain distribution in the compression zone of flexural members. Test results, combined with the available data in the literature, were used to develop new proposed rectangular stress block parameters for HSC and recommend revisions to extend the applicability of LRFD Specifications to HSC up to 124 MPa. A proposed simplified design procedure, using the proposed stress block parameters and calibrated by the results of full-scale prestressed HSC girders tested to failure, is presented. The girders were tested to validate the ultimate flexural strengths using the proposed rectangular stress block parameters, which provided conservative prediction of the flexural strength of prestressed HSC girders.

121 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Strength prediction on the stiffened plates in compression

Y.B. Kwon & D.W. Kang Department of Civil and Environmental Engineering, Yeungnam University, Gyongsan, Korea B.H. Choi & T.Y.Yoon Research institute of Industrial Science and Technology, Civil Engineering Research Team, Hwasung, Korea

The local buckling caused negative effects on the ultimate strength of the compression members which buckled and failed in a distortional mode. The strength curves for the stiffened steel plates in compression have been proposed to account for the interaction between local and distortional buckling, which was proposed byYangand Hancock(2005) and Modified by Kwon et al. (2007).The strength formulas by Kwon et al. (2007) for Direct Strength Method are expressed as Equations 1a-b.

√ where λdl = Fkh/Fcrl; Fnld : design compressive strength; Fcrl : elastic local buckling stress; Fkh : distortional buckling strength computed by the distortional buckling strength(Kwon and Hancock, 1992). The strength formulas are compared with the test results Figure 1. The comparison shows that the predictions by Equations 1a-b and test results agree well.

Figure 1. Comparison of tests and DSM.

122 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Strength of fillet welded splices of SM570-TMC, extra thick plates

Jae Byung Jo & Jin Woo Kim , Seoul, Korea

Recently, SM570-TMC steel plates were produced up to about 100 mm thickness and applied to bridges in Korea. As test results were not available for the strength of fillet welded splices fabricated from extra thick plates of steel grade SM570TMC, it was necessary to investigate the applicability of the strength equations of design codes. Fillet welded lap splice specimens were fabricated by flux cored arc welding from SM570-TMC steel plates of 60 mm and 82 mm thickness and 20 mm thick SM520 plates provides by POSCO. The sizes of fillet welds are 7–18 mm. The specimens have weld axes parallel (for type B) or perpendicular to the loading direction (for type C) or both of them (for type A). The specimens were axially tensioned to failure. The test strengths of type C are significantly higher than those of type A and B. The increment in strengths of fillet welds loaded transversely should be considered for economical design and also for the same level of safety in comparison with longitudinally loaded fillet welds. The influence of plate thickness on the strength of fillet welds is not recognized. The strength equations of different design codes are not all same for fillet welds whose axis are not parallel to the direction of loading, especially when fillet welds are loaded perpendicular to their axes. By using the von Mises’ condition and the results of FEM calculations, an engineering model is proposed to develop simple equations for the estimation of strengths of the tested specimens. The proposed strength equations yield results which are very similar to those calculated with the equations of AISC. Figure 1 shows ratios of test strengths (max. Load) to calculated values with the proposed strengths equations (Pref ) and also with the AISC equations (PAISC) by using the specified tensile strength of electrodes. In both cases, the ratios are distributed in the similar ranges for all types of specimens and all thickness of plates. For all specimens, the test results are higher than the calculated strengths. The average ratio for each type is 1.30–1.39, and for each thickness 1.32–1.37. The reason, why the calculated strengths are lower than the test results by about 30% on average, can be found in the value of tensile strength used for the calculations. The actual tensile strength of weld metal seems to be higher than the specified value for electrodes as it could be noticed from the hardness test results.

Figure 1. Ratios of test strengths to calculated strengths with the proposed equations.

123 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Safety of ductility demand based seismic design for bridge columns

Jae-Hoon Lee & Jin-Ho Choi Department of Civil Engineering., Yeungnam University, Gyeongsan City, Korea Jung-Kil Hwang Dong-Ho Co., Ltd, Korea Hyeok-Soo Son Seoyeong Engineering Co., Korea

The current seismic design criteria of the Korean Bridge Design Specifications [KBDS] have adopted the same seismic design concept and requirements as the AASHTO specifications. This design approach has been based on the force-based full ductility design concept which was essen- tially developed for strong earthquake regions. It regulates the use of constant values of response modification factor which is called behavior factor in Europe, i.e. R = 3 or 5 regardless of ductility demand, and requires providing a large amount of confining steel to bridge columns so that enough ductility can be guaranteed. However, adopting the full ductility design concept sometimes results in not economical design and construction problems due to reinforcement congestion, especially in moderate seismicity regions like Korea. It sometimes results in construction problems such as steel cage manufacturing and concrete placement due to reinforcement congestion as well as economic problems. Therefore, a design based on required ductility and proving required transverse steel determined by the demand might be a reasonable approach for the moderate earthquake regions. This paper introduces ductility demand based seismic design method and procedure for the bridge columns in low to moderate earthquake regions. The basic concept is that R factor should be used as variable and determined as required, which is the constant value by the current design specifications of KBDS and AASHTO Specifications. The required confining steel amount is determined by the required ductility demand of the columns. This approach may provide more economical and reasonable seismic design methodology. It may also prevent possible construction problems due to excessive use of transverse reinforcement. Design equations were developed to determine required confining steel amount based on ductility demand. The natural period of bridge, column aspect ratio, mechanical properties of concrete and reinforcement, longitudinal steel ratio, axial force ratio, and ductility demand are considered in the design equations. A total of 89 column test results by quasi-static test were selected to investigate the safety of the proposed ductility demand based seismic design method for the bridge columns. The column specimens had no lap splice of longitudinal reinforcement in plastic hinge region and confined by spiral or circular hoops. For the selected 89 columns, safety of displacement ductility is investigated by confining steel index, longitudinal steel index, axial force ratio, and aspect ratio. The range of safety factor for displacement ductility shows between 1.11 and 3.98, and the average is 1.90 that may be quite reasonable margin of safety for design purpose. It is believed that the proposed method is rational in concept and provides economical design especially for moderate earthquake regions.

124 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

An experimental investigation of the ultimate flexural behavior of steel tub girders with top lateral bracing

B.H. Choi & T.Y.Yoon Research Institute of Industrial Science & Technology (RIST), Hwaseong, Korea Y.S. Park Myongji University, Yongin, Korea

This paper presents a series of experimental test results on the flexural strength of a steel tub girder with top lateral bracing (Fig. 1). Three test girders were fabricated with high tensile strength SM570 TMC (nominal yield strength, Fy = 460MPa) steel plates. The pure bending test was conducted on the tub girder specimens using a universal test machine (Fig. 2). The flexural resistance capacity specified in the AASHTO LRFD design specifications was compared with the evaluated ultimate strengths, and then, a good correlation was found in between the tested re-sults and the design specifications.

Figure 1. Typical steel tub girders (non-composite).

Figure 2. Test setup for uniform bending (1000-ton UTM).

125 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Reliability based calibration of limit state bridge design code with material and member resistance factors

I.Y. Paik & D.J. Bang Kyungwon University, Songnam, Kyunggi-Do, Korea

This paper presents some of the features regarding reliability based calibration process for the ongoing calibration work for Korean bridge design code with limit state design concept. The partial safety factor format for the structural resistance of the proposed code consists of both material resistance factor and member force resistance factor. Calibration of the resistance factors is carried out for the proposed live load model and load factors. Statistical data for materials are collected and analyzed from domestic construction sites and are used in the calculation. Numerical results are presented in this paper for prestressed concrete girders. Girder sections are designed both for the stress limit state and strength limit state. Statistical characteristics for the ultimate limit state of the moment capacity and shear capacity of the designed sections are obtained and compared through simulations. Parametric studies with different sets of resistance factors and different types of formulations are carried out during the design of sections and the capacity simulation. Reliability analysis is performed for flexural moment and shear strength for structural member design. Reliability index is obtained following the first order reliability method of Rackwitz- Fiessler procedure. Results show that reliability index for flexural strength with member force resistance factor φf = 0.85 is greater than that with material resistance factor φc = 0.65, φs = 0.90 because reinforcing steel governs the reliability index for ductile flexural strength. Reliability index for shear with member force resistance factor φv = 0.80 and that with material resistance factor φc = 0.65, φs = 0.90 are smaller than those of flexural case. When additional member force resistance factor of φv = 0.95 is applied to material resistance factor system, the reliability index increases by about 0.3 in this example and yields about the same value of reliability index as flexural moment. Thus by introducing additional member resistance factor in combination with material resistance factor in shear design, the reliability for both flexural moment and shear reaches about the same level. The serviceability limit state is studied for the stress of the girders subjected to the proposed live load model. When the design is based on serviceability limit state, the required tendon area for the proposed load model is larger than that for the current load model whereas it was almost the same when the design is based on strength limit state for the span ranges considered in this study. Also, the design is controlled by allowable stress for most of the span ranges for the proposed model whereas the strength limit state controls design when span is shorter than 30 meters for the current load model. Examples of the reliability analysis is presented in this paper for the combination type of resis- tance factor format which accounts for the variability from different materials as well as the variability from model for different member actions effectively. The results of the numerical study for the reliability indexes of the designed sections with different sets of partial safety factors may serve important information for the decision of the partial safety factors and the target reliability index of the proposed design code.

126 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Bridge E-8 in the new railroad of high velocity to the northwest of Spain

C. Jurado Civil Engineer (PhD in course), Ingecal Ingenieros, S.L., Polytechnic University, Madrid, Spain

ABSTRACT: During the years 2001 and 2006 it has been constructed in Spain the new railway line of high speed, that will connect Madrid (in the centre of the country) with the north and northwest of Spain.

The present article exposes one of the more singular bridges in the area near to Madrid between the kilometre points 104 + 800 and 108 + 832 including the project and the construction of a bridge E-8 called in Spanish “in pergola”, over the existing railroad that connects Madrid with the city of Alcobendas and with Segovia in the north of Spain. The project of the bridge was realized by means of a three-dimensional finite element model realized with the program SAP2000N that includes all the elements of the structure such as: piles, girders to tie piles in the zone of the pergola, abutments, prestressed beams, the deck and the girders for the prestressed beams (see fig. n◦ 1 and 3). The project of the bridge was realized during the year 2004 and the construction during the years 2005 and 2006 (see figures 2 and 4) Aspects of the 3D finite element model and relevant circumstances of the construction of the bridge are included in this paper. The finite element model of the bridge was a three dimensional model with three thousand elements which included all responses of the bridge.

Figure 1. F.E.M. model with 1664 nodes, 1323 frame and 592 shell elements.

127 The construction was realized without stopping the traffic of the existing railroads from Madrid and Alcobendas and from Madrid to Segovia. The bridge was put in service the 23rd of December of 2007.

Figure 2. Construction of bridge E-8 “in pergola” Madrid-Valladolid in phase of construction. Sight from the abutment of the side of Madrid.

Figure 3. Flexural moments M33 (HIP. COMB. n◦ 31).

Figure 4. Construction of bridge E-8 “in pergola”. High velocity train Madrid-Valladolid in phase of construction. Sight from the abutment of the side of Colmenar.

128 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

A simple iterative method for determining the effective length of structural members in steel cable-stayed bridges

Dong-Ho Choi, Hoon Yoo, Deok-Soo Lee & Yong-Sik Kim Hanyang University, Seoul, Korea

The eigenvalue analysis based on the system buckling approach is an accessible tool to calculate effective length of girders and towers in a steel cable-stayed bridge. However, there is a problem that the conventional eigenvalue analysis may yield an unduly large effective length for the members having a small axial force. Based on the system buckling analysis, the present paper proposes a modified eigenvalue analysis in order to determine the effective length of the members in a steel cable-stayed bridge. An example of a bridge model used in this study demonstrates that the proposed method not only provides acceptable outcomes overcoming the problem of conventional eigenvalue analysis, but also can be a good substitute for obtaining the effective length of the structural members in the practical design of a steel cable-stayed bridge.

Figure 1. Numerical model of the example cable-stayed bridge.

Figure 2. Effective length of the example bridge ((a) Girder members, (b) Tower members).

129 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

The influence of friction/sliding behavior of rubber bearing to the seismic performance of highway bridges

K.Y. Liu National Center for Research on Earthquake Engineering, Taipei, Taiwan K.C. Chang & W.I. Chen National Taiwan University, Taipei, Taiwan J.S. Hwang National Taiwan University of Science and Technology, Taipei, Taiwan

Lessons learned from 1999 Taiwan Chi-Chi earthquake have demonstrated that the rubber bearing system which significantly affected the seismic behavior of existing simply-support PCI highway bridges was due to the construction practice of a non-bolt detail, leading to a fuse-like mechanism to prevent columns from yielding plastic hinge damage. Though it might be safer for substructures, however, evidences of friction/sliding movement and residual displacement of rubber bearings can also be found during the investigations. Therefore, to get better understandings of the influence of the friction/sliding phenomena to the seismic performance of the bridge structure, an experimental program of a single-span scale-down bridge model equipped with two types of rubber bearings was performed by both shaking table test and pseudo-dynamic test. Among the four test cases with three peak ground accelerations (0.2 g, 0.5 g and 0.7 g), major differences for the case with same input ground acceleration but different test methods is aroused from the coefficient of friction. Since coefficient of friction is velocity-sensitive, experimental results show that for reinforced elastomeric rubber bearing, the coefficient is 0.2 to 0.45 regarding to slow and fast speed; while same bearing paved with PTFE material obtains the value 0.1 in static stage and 0.25 in dynamic stage, respectively. Hence, the maximum displacement of girder obtained from shaking table is smaller than the one from pseudo dynamic test as well as the residual displacement. In addition, whether the friction/sliding of rubber bearing occurs or not, the measured base shear force of the column well satisfied the inertial force of the superstructure, not only from the test but also the analytical model, being transferred via bearing system. In comparison with the test results, the analytical results by proposed numerical model are good at predicting the maximum displacement demand on girder and shear force demand in the columns.

130 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Ultimate flexural strength of hybrid composite girders at sagging bending

S.G. Youn & Y.T. Kim Department of Civil Engineering, Seoul National University of Technology, Korea D.B. Bae Department of Civil & Environmental Engineering, Kookmin University, Korea

ABSTRACT: This paper presents experimental work conducted with the purpose of verifying the ultimate flexural strength of hybrid composite girders constructed using the high performance steel of HSB600. HSB600 is newly developed in Korea and its 0.2% proof stress and tensile strength are over 450 MPa and 600 MPa, respectively. Two composite beams included a hybrid beam were fabricated and tested. The test results show that if the plastic moment of hybrid composite beam is calculated with the 0.2% proof stress for yield strength of HSB600, the ultimate flexural strength will be over conservative. Based on the test results, the ductility requirement and the flexural strength for hybrid composite girder constructed using HPS600 are discussed.

1 INTRODUCTION

Strain-hardening strain of HSB600, which has no significant length of yield plateau, is almost coincided with yield strain. Therefore, the behaviors of hybrid composite girders constructed with HSB600 for the tensile flange will be significantly different from those of conventional composite girders. In this paper, an analysis of the theoretical flexural strength and ductility parameter of hybrid composite girders is performed initially and then a series of experimental tests is presented. Two test beams having a cross-sectional geometry with the difference of the material of tension flange are fabricated and tested. The behaviors of the test beams such as ultimate loads and deflections are summarized. Based on the comparison of test results and theoretical results, it is discussed about a simple way for estimating the ultimate flexural strength and the ductility of hybrid composite girders.

2 BACKGROUND

At sagging bending, the higher strength steel can be used in the tensile flange which contributes most of the bending strength to the composite girder. For the hybrid composite girders with higher strength steel, which has a yield strength and enough the length of yield plateau, the moment capacity can be determined using both of the plastic moment, Mp, calculated using a yield strength of the higher strength steel and the ductility parameter. For conventional steels, the tensile strains of 0.027 mm/mm and 0.0315 mm/mm are always larger than not only the yield strain but also the strain-hardening strain of tension flange. How- ever, the stresses corresponding to the strains are not much higher than the yield stress due to the length of yield plateau as shown in Figure 1. However, in case of HSB600, the stresses at the strains of 0.027 mm/mm and 0.0315 mm/mm are nearly 580 MPa and 590 MPa, respectively. These stresses are higher than the 0.2% proof stress of 450 MPa. The differences between the stresses are considered enough to change the ultimate moment capacity. 131 Figure 1. Stress-Strain Curves of Typical Conventional Steel and HSB600.

3 EXPERIMENTAL WORK

The ultimate behaviors show that the 0.2% proof stress, f0.2%,ps, is not acceptable to be the yield stress due to the lack of yield plateau. When the 0.2% proof stress, f0.2%,ps, is used as the yield stress, the plastic neutral axis and the plastic moment, Mp, are underestimated. Based on the tests and the theoretical analyses, it is proposed that the stress f2.0% at 2.0% strain is the yield stress in calculating the plastic moment, Mp, and the strain-hardening modulus, Esh, is 4,455 MPa in the moment-curvature analyses. Further, for the safety, the strain-hardening strain, εsh, is proposed as the elastic strain calculated at the 0.2% proof stress of 450 MPa.

4 CONCLUSIONS

The objective of this study is to propose the simplified formulation of the moment capacity of the hybrid composite beams constructed with the high performance steel of HSB600 developed in Korea. The test results of two composite beams are presented in the paper. The results show that the 0.2% proof stress is not acceptable to be the yield stress of HSB600 due to the lack of yield plateau. Based on the test results and the theoretical analyses, some variables are proposed for the plastic moment and the moment curvature analyses. A series of three hybrid composite beams have been tested as an on-going project and it will be used to calibrate the above proposals.

ACKNOWLEDGEMENTS

The financial support to the research activity of the authors provided by the Korea Bridge Design & Engineering Research Center (KBRC) and the assistant of Pohang Steel Corporation (POSCO) are highly acknowledged.

REFERENCES

American Association of State Highway Transportation Officials (2007). AASHTO LRFD Bridge Design Specifications, 4th Ed., Washington, D.C. Rotter, J.M., and Ansourian, P.(1979). “Cross-section behavior and ductility in composite beams.” Proc. Instn Civ. Engrs. Part 2, 67, June, 453–474. Wittry, D.M. (1993). “An Analysis Study of the Ductility of Steel-Concrete Composite Sections.” MS thesis, University of Texas-Austin, Austin, Tex. Eurocode 4(1997): Design of Composite Steel and Concrete Structures, Part 2. General rules and rules for bridges. ENV 1994-2:1997. Ansourian, P (1982). “Plastic Rotation of Composite Beam.” J. Struct. Div., ASCE, 108(3), 643–659. Mans, P.,Yakel, A.J., and Azizinamini, A. (2001). “Full Scale Testing of Composite Plate Girders Constructed Using 485-MPa High-Performance Steel.” J. Bridge Eng., 6(6), 598–604. 132 Bridge inspection and diagnostics

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Application of data fusion technology on scour and siltation monitoring in river bed

Z.J. Chen, W. Wang & S. Chen College of Civil Engineering, Hohai University, Nanjing, Jiangsu, China S.H. Cao Jiangsu Province Sutong Bridge Construction Commanding Department, Nantong, Jiangsu, China

The size of the pile groups of the main bridge is huge in Sutong Bridge. So it brings adverse influence on hydrological condition and scour and siltation situation. Especially on the soft and sub- consolidated bottom soil of riverbed, unpredictable local scour may occur which impacts the safe construction and operation of the working flat roof and the foundation. In order to monitor the scour and siltation situation of riverbed effectively and provide reference for engineering construction in time, multi-beam radar technology with large-scale is adopted in engineering operation to generally monitor river topography and protective layer periodically and macroscopically and handle the change of river topography in macrocosm. At the same time, the river topography inside the pile group foundation can not be monitored by multi-beam radar technology after the forming pile foundation cap. So, in the construction of pile group foundation, a monitoring network of river topography in key area, which can do continuous, real-time and quick monitor is composed by high accuracy water pressure sensor and tide level sensor to analyze the change of scour and siltation in key area of river surface detailedly. Because the installation method itself, working environment and the acquisition mode of monitoring instrument itself can make the observation data include noise which seriously interferes the true value of data and prevent further analysis and treatment. Therefore, to acquire accurate and comprehensive observation data, it is necessary to have the data de-noised. In order to obtain comprehensive information about scour and siltation, multi-scale data fusion technology based on wavelet is adopted to reconstruct and decompose the data of different scales, describe scales uniformly by fusion and get the comprehensive information which is fine and flexible more than ever. The model of scour and siltation can be made by using the fused information. And it can reflect the general condition in scour and siltation of river bed in large-scale and also can be used to query the location, shape and amount of scour and siltation in key area conveniently. The scour and siltation condition from qualitatively to quantitatively also can be studied by comparing and analyzing the scour and siltation model in different observational times. It states that the model of scour and siltation reflects well the evolutionary process of scour and siltation of the riverbed in the location of the bridge.

Keywords: Sutong Bridge, pile group foundation, river bed scour, water pressure sensors, multi- beam measure system, multi-scale monitoring, wavelet, information fusion technology.

135 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Special tests of two post-tensioned concrete viaducts

J. Ciesla, M. Lagoda & P. Olaszek Road and Bridge Research Institute, Warsaw, Poland

ABSTRACT: In the paper the results of field-tests of two post-tensioned, concrete bridge struc- tures, after their completion, have been presented. First structure is six span continuous viaduct. In cross-section the structure has been designed as box girder with two cantilevers. The structure is curved horizontally and vertically. The second structure has been designed similarly, as five span continuous box girder structure, but straight in plane and length profile. In both cases during erection they were some unforeseen accidents. In both cases they were made some improvement of design and reparation of structures, and after completion they have been decided to arrange a special enlarged program of field-test.

1 DESCRIPTION OF THE STRUCTURES

First structure, that’s six span continuous viaduct structure of total length of 187 m. Spans lengths are 21 + 29 + 34 + 40 + 36 + 27 [m]. After completion of two segments, which was about a half of total length of structure, it was some turn of structure with vertical displacement over one bearing on the abutment. In the situation, some rectification of all part of structure, by displacement of bearing and using additional prestressing vertical tendon over abutment, have been applied. After that, the viaduct has been completed. The second structure that’sfive span continuous viaduct structure of total length of 196 m. Spans lengths are 38 + 39 + 39 + 45 + 35 [m]. The structure was concreted and prestressed progressively segment by segment also. It was a local damage of concrete during prestressing one of over support cables because of lack of special reinforcement for carrying vertical pressure along curvature of cable.

2 SCOPE OF TESTS AND CONCLUSIONS

Described tests in case of both structures they were tests under load, we are obliged to apply for bigger bridge structures in Poland. Because of mentioned mistakes in design and execution applied tests had a special enlarged program. Main purpose of tests was to prove, that all improvement in design and execution of both structures were right. Under prove tests all span were tested. Special enlarged tests have been applied in case of chosen spans for checking its structural behaviour and level of prestressing. They were applied tests under static and dynamic loading. Presented tests in case of both structures conformed right way of structural behaviour under static and dynamic loads and correct way of improvement in design and execution. Both structures have been permitted for exploitation. In case of structure No 1, because of complex shape of structure and some incorrectness during execution, they were recommended some cyclic geodetic measurements of all supports and spans. They were no objections in case of structure No 2.

136 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Active lamb wave propagation-based damage detection and location for steel plate

Woon Jeong & Juwon Seo Hyundai Institute of Construction Technology, Hyundai Engineering & Construction Co., Korea Hyeungyun Kim Advanced Structure Monitoring, Inc., Cupertino, USA

Conventional NDE/NDT (non-destructive evaluation and testing) method such as C-scan, x-ray, eddy current, and coin tapping are time consuming and labor intensive for large structures, and impractical for an in-service condition, and difficult to perform for some parts that are inaccessible to the maintenance crew. For the in-situ or in-service SHM (Structural Health Monitoring) of critical members used in civil infrastructures, which include bridges and buildings, LW(Lamb Wave)s used in conventional testing techniques cannot be directly applied, because they usually require bulky instruments and human interference. Recently, many researchers have studied the technology of integrating piezoelectric actuators/sensors into the structures for the purpose of generating and collecting diagnostic LWs and thus realizing continuous monitoring of their structural integrity. In this study, DNP (Diagnostic Network Patch) system is implemented as an integrated solution for inspections in 5 mm thick steel plate. PZT ceramic disks are surface mounted on a steel plate acting as both actuators and sensors to generate and collect LW. A narrow-band, modulated and Hanning-windowed sinusoidal tone-burst of five peak cycles excites the actuator generating LWs, in order to overcome the dispersive nature of LW propagation. For a successful LW testing, the dispersion curves for the plate, known as Rayleigh-Lamb equations, are drawn by some sort of iterative root-finding algorithm such as Newton-Raphson, bisection, etc.. Group velocity may be calculated by dispersion curves. To select much less dispersive region where the interpretation of response signals becomes easier, an actuator is excited at narrow-band driving frequency 0.17 MHz. According to the dispersion curves, both S0 and A0 mode LWs are generated synchronously in that frequency region. And S0 mode LW is used for the diagnosis. LW testing technique, pitch-catch method, was used for interpretation of notch defects with depth of 50% of plate thickness and 7 mm width. Scattered LW signal from the damage, referred to as the ‘scatter’, can be obtained by subtracting the baseline sensor signal of the undamaged plate from the recorded sensor signal of the damaged plate. Damage-sensitive features of the scatter, such as delay in time of transit from difference in TOF (Time-Of-Flight)s, amplitude, frequency content, etc., can be extracted from the signal using some signal-processing algorithm. The CWT (Continuous Wavelet Transform) using Gabor function as a mother wavelet was adopted to extract the TOF from sensor signals. The group velocities calculated from the CWT agree well with the theory. Also, Donoho’s DWT (Discrete Wavelet Transform)-based thresholding denoising technique using Haar function as a mother wavelet was applied to minimize noise interference and increase the SNR (Signal-to-Noise Ratio). A new damage location practical algorithm, based on damage location triangle method, has been proposed and verified with good accuracy. The possible damage location can be estimated by the average on calculated location points and the damage extent by the standard deviation.

137 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Settlement prediction model for pile foundation based on field observation

Xiaoyan Li & Zhijian Chen College of Civil Engineering, Hohai University, Nanjing, Jiangsu Province, China Xuewu Dong Jiangsu Sutong Bridge Construction Commanding Department, Nantong, Jiangsu Province, China

ABSTRACT: The main of Sutong Bridge, which was built as the longest span cable stayed bridge in the world, is situated in the tideway of Yangtze River. The pile-group foundation of it has the characteristics of large scale, complex loading and structure, and bad geologic and hydrology condition. Prediction of settlement is important for construction control, design change and health monitoring of super-large pile foundation built on thick overburden layer. Because the maximum tidal range reaches to 4m on the location of Sutong Bridge, and the area of the pile cap is 5603 m2, the maximum variable load caused by tide range reaches to 224 MN. It acts on pile foundation circularly, and affects the settlement of the pile group foundation. General computing methods for settlement, such as equivalent pier method, elastic theory method, shear displacement method, load transfer method etc, is brief and easy, but it is difficult to consider tidal fluctuation and consolidation process. Biot’s consolidation equation by finite element method can be used to simulate actual three dimensional consolidation of soil and to analyze complex loading, but it can’t carry out real-time computation considering tidal fluctuation. It is necessary to set up real-time settlement prediction model based on Biot’s consolidation theory considering tidal fluctuation and consolidation process. The steps of modeling for settlement prediction are given as following. (1) To set up discrete finite element model. Based on symmetry of the structure and loading, the discrete model of 200 m × 200 m × 200 m is set up for finite element analysis according to the distribution of soil and the structure of half pile-group foundation. Goodman element is adopted to simulate the contact interface between piles and soil. (2) To obtain rational parameters for Biot’t consolidation equation. Intensity and permeability of foundation soil should change with the process of consolidation, so hyperbola constitutive model of Duncan-Chang is adopted to give elastic modulus and Poisson’s ratio of foundation soil in the equation, and nonlinear coupling model is adopted to give permeability coefficient. Settlement of the pile head had been observed during the concreting of pile cap. Parameters of intensity and permeability of foundation soil are back analyzed by least square method according to test data and observed settlement during the concreting of pile cap, and nonlinear parameters of intensity of contact interface are obtained at the same time. (3) To compute settlement by positive analysis. Settlement caused by load of superstructures and tide fluctuation can be computed respectively and progressively. (4) To fit the relationship of computed settlement and load of superstructure and that of settlement and tide level respectively, real time settlement prediction model of pile-group foundation can be set up. (5) In order to test the reasonability of the settlement prediction model and the reliability of modeling method, prediction model of settlement observation point between binder and bottom tower is set up. According to the prediction model, settlement caused by construction of tower is about 26.1 mm, and the observed one is 26.8 mm.The relative error is less than 3 per cent. It states that the settlement prediction model is effective and rational.

138 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Ambient vibration of stay cables used for damage detection in cable-stayed bridge

Chien-Chou Chen & Wen-HuaWu National Yunlin University of Science & Technology, Taiwan Jay Lin National Center for Research on Earthquake Engineering, Taiwan

ABSTRACT: The vibration-based method has been a common approach to detect the damage situation and diagnose the health condition of cable-stayed bridges. The methods developed are mainly based on the premise that modal parameters, such as natural vibration frequency and mode shape, would change as structural damage forms. However, from the measurement of the dynamical response of some bridge models with artificial damage it was shown that modal parame- ters are, in general, not sensitive to structural damage and temperature change can induce remarkable frequency shift. In addition, when structural damage forms the major change of modal parameters often takes place in higher modes, which, in general, needs very precise measurement of the girder vibration to quantify the change. Accordingly, it would be costly to arrange the measuring device. Thus, another approach, which is based on the ambient cable vibration, is proposed to recognize the structural damage and diagnose the health condition of Chi-Lu Cable-stayed Bridge. Roughly speaking, there are two reasons why the proposed concept was applied to construct the structural health monitoring system of the bridge. First, the geometry of cable enables its modeling as a simple one-dimensional structure, which significantly simplifies its dynamic response measure- ment and corresponding analysis. Second, the cable is the primary force-transmitting member of cable-stayed bridges and consequently plays an important role in reflecting the health condition of the whole structural system. In the initial phase of this study, the measurement on the cable system of Chi-Lu Cable-stayed Bridge was implemented to get its ambient vibration data. From the transverse and longitudinal measurement of cable vibration displayed in frequency domain, it is not difficult to identify the accurate vibration frequencies for the first several natural frequencies of the cables and the vibration frequencies of the longitudinal wave along the cables, which are the important information to monitor the internal force and axial stiffness of the cables. Then, a series of static analysis and dynamic analysis were performed to give the force variation of stay cables and the natural frequency change of the bridge to unusual loading conditions, such as differential settlement, unexpected static overloading, earthquake, strong wind and so on. The analyzed results indicated that the cable force was more sensitive to the applied loads than the natural frequencies of the bridge and the features of the force variation in the cable system were different and could be recognized to different type of unusual loadings. Finally, a neural network was trained to identify the type and degree of the unusual loading conditions subjected to the bridge from in-situ measured ambient cable vibration even with noisy. The architecture of the neural network for Chi-Lu Cable-Stayed Bridge include 4 layers, which are one input layer with 35 neurons, 2 hidden layers with 12 (adjustable numbers) of neurons, and one output layer with 6 neurons. The test results indicated that the neural network is feasible for the recognition. The structural health monitoring, which may take the proposed concept into account, is being constructed to make sure the structural safety of the bridge in service phase.

139 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Specials inspections and maintenances of prestressed concrete in Rio-Niterói Bridge

C.H. Siqueira Ponte S/A, Rio de Janeiro, Brazil

With its 14 km of extension, the Rio-Niterói Bridge is the largest work of road engineering in Brazil, is among the 10 largest bridges of the world, and represents the most important group of prestressing structures of the Americas, with 43.000 prestressed cables, equivalent to 140.000 km of wires. To guarantee the perfect acting in that cast’sservice, it is necessary to not only have the prestressing con- crete inspected and maintained, but also to obtain knowledge from its theoretical/practical behavior against eventual anomalies installed in the cables, that can alter the acting of the structural group. This paper shows the inspection and maintenance techniques, unprecedented in Brazil and maybe in the world, that are used to guarantee the structural safety of prestressed concrete of the Rio-Niterói Bridge.

140 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Non-destructive testing of suspender ropes with magnetostriction

Michael S. Higgins Pure Technologies, Columbia, MD, USA Oliver Tozser Pure Technologies, Calgary, AB, Canada

Suspender ropes are important structural members for suspension bridges and arches with sus- pended decks that must be periodically inspected and assessed. Research and testing has shown that visual inspection of the exterior of the rope does not necessarily correlate with the interior condition of the rope. The interior condition is important because corrosion is often initiated on the interior of the rope and then progresses outward. Thus corrosion may be most extensive on the interior of the rope. Recently, a Non-Destructive Testing (NDT) technology was used on suspender ropes to detect either broken wires or loss of cross section. The technology relies on the principles of magnetostric- tion to generate a guided acoustic pulse that travels up and down the rope. Where there is a defect in the rope resulting from corrosion or fatigue damage, a portion of the pulse is reflected back to the source. These reflections can be measured and quantified to determine the size and location of corrosion or fatigue damage on a suspender rope. This technique does NOT use the magnetic flux principles, which has been used on suspender ropes in the past with limited success. Magnetostriction is a characteristic of all ferrous materials to undergo a change in dimension when subjected to a magnetic field. By inducing a dynamic magnetic field around the rope in a controlled fashion, an acoustic pulse can be generated that travels the length of the rope. Since the reflections are returned to the source, the testing can be performed at one convenient location, typically from the bridge deck. This approach to testing suspender ropes offers a convenient method of providing detailed information on the ropes. This paper discusses the principles and case studies of projects using magnetostriction to test the condition of suspender ropes and the capabilities and limitations of such an approach.

141 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Judging suitability of arch bridges for higher axle loading by load testing

Rama Kant Gupta Ircon International Limited, District Center, Saket, New Delhi, India

ABSTRACT: Arch is the only structural configuration having safe load carrying capacity much more than its designed capacity. This is so, since many factors are there which enhance the strength potential of an arch, but not considered in design. Being a compressive member, life of the arch bridge is comparatively more than any other bridge structures. Many of the arch bridges are having heritage importance and system requirement is to safely retain them to the extent possible. Con- sidering all these factors clubbed with its age and enhance traffic load, arch bridges require more careful consideration than any other type of bridge structures. Every railway/road system of the world is having appreciable percentage of arch bridges. Out of total 119724 numbers of bridges on Indian Railways, 20967 bridges, i.e. approximately 17.5% are arch bridges. Majority of the arch bridges are more than 100 years old. In the beginning, axle load on Indian railways was much less. It was 7.5 tonnes in 1853, when railway network started in India. Gradually, the same has been increased to 25 tonnes at present and likely to become even more in the near future. It was the requirement of the system to increase the axle load to cater for the increased traffic demand. It is the practice on Indian Railways that before introduction of any change in loading patterns, bridges have to be analyzed so as to check its safety. Accordingly, all the arch bridges were also theoretically analyzed by the prevalent elastic method and it was found that overstressing to the tune of 500% was coming. Even with such an extent of overstressing, majority of the arch bridges were not showing any sign of distress. Rather, all those were performing satisfactorily. To assess the actual strength of the arch bridges, Research, Designs and Standard Organization (RDSO) of the Indian Railways had conducted lot of research work on arch bridges and found that there are so many factors, which are not included in the design, but add strength to the arch bridges. Some of those factors are hunches, parapet walls, filling on the arch bridges, continuity effect like railway track in case of rail bridges and road metal in case of road bridges provided on the arch bridges. Even extent of contribution of each of the aforesaid factors was arrived at after the thorough testing and research work. It was concluded that particularly in case of arch bridges, where strength potential is much more than its design capacity, real assessment about fitness should be based on the load testing. Based on the rigorous field experimentation as a part of research work, load-testing criteria for arch bridges was subsequently finalized. In this paper, the author will share about the methodology adopted for testing of the arch bridges, conclusions drawn from the numerous test results and how contribution of strength adding factors like hunches, parapet walls, filling on the arch bridges, continuity effect etc were quantified. All such exercise saved A lot in safely continuing the old arch bridges even designed for lighter axle load.

142 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Health assessment of pre-stressed concrete girder bridges by non-destructive testing

Rama Kant Gupta Ircon International Limited, District Center, Saket, New Delhi, India

ABSTRACT: Non-destructive testing is one of the very effective means in Decision Support System for ensuring economical, timely and effective maintenance of the bridges. Its appropriate utilization not only ensures safety of the structures, but simultaneously also boost up the confidence of the maintenance engineers by ascertaining them about actual condition of the bridges. Here is the case study of 7 number of PSC girder bridges on which, unnecessarily speed restriction continued for several years and on account no confidence and expenditure of INR 800 million was likely to be incurred. On Obra-Singarauli Section of Indian Railways, 7 bridges consisting of 27 numbers of spans varying from 60 ft. to 80 ft were constructed during 1965–66. In the beginning itself, it was noticed that fine cracks were there on the surfaces of those pre stressed concrete girders. Accordingly, speed restrictions were imposed on all those 7 bridges. Simultaneously, all those 7 bridges were kept under observations. On account of no further development of the cracks and no other adverse effect, those bridges were continued in service, but the maintenance engineers remained in dilemma about the actual health of all those 7 bridges. Expert opinions were taken from the experienced bridge engineers as well as from the professors. None of them were confident about the actual condition of the bridges. Their opinions were also differing. As such, situation remained inconclusive. Meanwhile, varieties of ND equipments came in the market. Their reliability and acceptability also increased in due course of time. Use of ND equipments particularly for strength of concrete and corrosion to the reinforcement revealed that the concrete is having adequate strength and corrosions to the reinforcements were not initiated. Crack width and depth measuring equipments revealed that the cracks are superficial, and not reached up to the reinforcement. Use of AET revealed that the cracks are dormant and non-structural. Load testing revealed that the structures are very much within elastic limit up to the severe most loading in operation on that section. Vibration Signature revealed that the performance is consistent and not deteriorating further even over a period of appreciable time. When it was concluded that that bridge structures are safe in all respect, speed restrictions imposed on those bridges were relaxed in phases and rebuilding works were postponed saving INR 800 millions. While working as Executive Director/Bridge & Structure at Research, Designs and Standard Organization of Indian Railways, author of this paper was actively involved in the aforesaid decision making process. In this paper, it has been tried to share the associated problems, remedial measures taken during that time and how final decision was taken to retain all those bridges saving huge amount of money by the way of postponement of rebuilding of those bridges.

143 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Optimal inspection and maintenance strategies for bridge network using supply and demand approach

A.D. Orcesi & C.F. Cremona Laboratoire Central des Ponts et Chaussées, Paris, France

ABSTRACT: The current approaches in Bridge Management Systems (BMS) for determining optimal maintenance strategies are mainly focused on structural condition. The aim is often to maintain bridges between their original condition state and a minimal acceptable level that is linked to security concepts. With such an individual strategy, it is not possible to determine optimal maintenance strategies at a network scale. Yet, in the context of limited budget resources, decision makers need to take into account parameters such as influence of bridges performance on the network connectivity and on the user satisfaction (in terms of traffic demand) to decide of the optimal allocation of available resources. The objective of the paper is to provide a methodology to be used to determine optimal mainte- nance planning for several bridges within a network. A supply and demand approach is combined with a probability-based formulation of the inspection and maintenance activities. On the one hand, the model that evaluates user costs takes into account congestion and assigns the traffic on the network by supposing that vehicles always try to minimize their travel cost. On the other hand, the model that determines the time-performance of the bridges uses a probabilistic approach to assess the performance of degraded structures. At an inspection time, different events can occur: no damages on the bridges, maintenance actions or rehabilitation of the bridges. All these events are uncertain and expressed in terms of probabilities. As maintenance actions are cost-dependent, the total expected cost of maintenance is calculated. Optimal maintenance instants for each bridge are then determined by using a Genetic Algorithm based procedure in such a way that total expected

Figure 1. Studied road network with pictures of bridges.

144 costs are minimized and that constraints are fulfilled. This optimization helps to find a good compromise between bridge reliability and users/agency costs. To illustrate the theoretical aspects, an application example is introduced. The studied network is a part of the French national road network (figure 1) managed by the Road Directorate of the Ministry of Ecology, Sustainable Development and Spatial Planning. The originality of this method is to keep a very general expression for the evolution of the bridges performance, which allows adjusting the methodology to a large number of failure modes. The objective is finally to provide to decision makers a good basis in the field of decision processes where multiple criteria are considered on a bridge network.

145 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Experimental investigations on the strength behavior of box beam and circular column connections

Y.P. Kim Yooshin Engineering Corporation, Seoul, Korea W.S. Hwang Inha University, Incheon, Korea

This paper presents the experimental results on the strength behavior and failure modes of box beam-to-circular column connections in steel piers. Recently, steel piers have been widely applied for pier structures of urban overpasses and elevated structures in East Asian countries due to their small space requirements, excellent earthquake resistance capacity, and fast construction period. At the T-type or framed beam-to-column connections of box-sectioned steel piers, it has been widely acknowledged that serious shear lag and stress concentrations may occur due to abrupt direction changes in member forces, so it is necessary to handle these problems properly in the design stage. Instead of box-sectioned steel piers, circular piers with box beam-to-circular column connections have been introduced recently due to a structural advantage of circular columns: they behave efficiently for a direction change of loading and they are a little affected by shear lag stress. Therefore, circular-sectioned piers are being recommended as a standard type of pier due to their external appearances. Previous researches except Okumura and Ishizawa (1968)’s, have been limited to box-sectioned connections only. Also, most researches do not consider that the behavior of the box beam-to- circular column connection is much more efficient than that of H-sectioned or box-sectioned connections in steel piers. For the box beam-to-circular beam connection in steel piers, Okumura and Ishizawa (1968) performed an unique research through theoretical approaches. They suggested that the shear lag stress of circular column could be negligible and introduced the concept of equivalent web depth d2 in the substitution of circular column. An equivalent web depth d2 is the function of the central angle α, which is used to calculate the shear lag stress and shear stress of the box beam. However, d2has the problem that the shear lag stress and shear stress of the box beam was overestimated with the variation of the central angle α. Their study ignored failure modes of connections. The theoretical yield strength equations of box beam-to-circular column connection have not been suggested. Also, the experimental study of welded box beam-to-circular column connections has not been investigated sufficiently. Therefore, in this research, monotonic loading tests are carried out for the eight beam-to-column connection models considering central angles, horizontal plate stiffeners, and equivalent web depth. From the results of out-of-plane deformation and strength of connection, failure modes and strength behavior are investigated. A more reasonable equivalent web depth is suggested, and the experimental yield strength of connection is compared with the theoretical one.

146 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Nondestructive evaluation of effective prestress using the core-drilling method

S. Pessiki Department of Civil and Environmental Engineering, Lehigh University, Bethlehem, PA, USA M.J. McGinnis Department of Civil Engineering, University of Texas at Tyler, Tyler, TX, USA

The core-drilling method is a stress-relief technique used to determine in-situ stresses in concrete, such as the effective prestress in a prestressed concrete bridge element. Information about effec- tive prestress is needed to evaluate the anticipated service performance of prestressed concrete bridges. In the core-drilling method, a small hole is drilled into the concrete of a structure and the displacements that occur as a result are measured. The displacements are converted into in-situ stresses using elasticity theory. Three factors that influence stresses determined with the core- drilling method have been identified: (1) swelling of concrete around the core hole due to exposure to water used as a lubricant during the drilling process; (2) changes in the measured deformations due to relief of differential shrinkage stresses; and (3) the presence of steel reinforcement in close proximity to a core hole. Recent research has addressed each of these factors by applying analytical and numerical techniques to adjust stresses calculated using the core-drilling method. This paper describes experiments that were performed to verify these approaches and to show that the core- drilling method can be used to determine accurate in-situ stresses in concrete structures. Concrete plates representative of webs of prestressed girders were loaded in compression and subjected to hole-drilling, and the resulting displacements were measured with 3D digital image correlation. Stresses calculated from measured displacements agreed with applied stresses to within 28%. When the calculated stresses were modified to account for the effects of the three influencing factors, the relative error in applied versus measured in-situ stresses in the experiments was less than 10%.

147 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Bridge safety management system by using Bridge Inspection Robot

Dong-jin Park & Hyun-guk Jung Robots and Design Co.,Ltd., SeongNam-Si, Korea Byeong-ju Lee & Won-tae Lee Highway Research Center, Hwaseong-si, Korea Jin-oh Kim Dept of Information and Control Eng, Kwangwoon University, Seoul, Korea

In this paper, we discuss a bridge safety management system by using Bridge Inspection Robot which enables acquiring images of the bridge condition for managing the safety of its structure. The purpose of this study is composed of two parts: 1) The image acquisition of a bridge structure with the application of a vision-based robot. 2) Development of software for assessing condition of the bridge, and measuring the size of any crack captured in the images. The robot arm is designed to operate on the underside of superstructure during inspections, and equipped with vision devices (digital camera, lens, and lights). The domain of a superstructure has been divided into appropriate segments to facilitate the image acquiring task by making its total domain irrelevant to the procedure. The system’ssoftware detects, measures the width and length of a crack semi-automatically, and also composes a flaw map of the whole area of the bridge from the crack’s data for a user’s convenience in determining the status of bridge. Through field experiments, the application of this inspection system with specialty software has proven to be much faster, safer, and reliable than the inspections carried out by the naked eyes in managing safety of the bridges.

148 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Matrix based cable-stay bridge cable force and deck elevation adjustments and FEM updating

A. Turer Civil Engineering Department., Middle East Technical University, Ankara, Turkey

ABSTRACT: Cables of cable-stay bridges commonly have an adjustment mechanism used to change the locking length of the cable according to the design values. Jacking force that will be applied on each cable is commonly determined by nonlinear structural analysis utilizing staged- construction and sag related nonlinear calculations. However, the final cable forces and vertical deck elevations might differ from those calculated by the analysis. Readjusting the cable forces and/or deck elevations is a complex problem since length or force alteration in one of the cable would affect the other cable forces and deck elevations. This paper discusses a novel approach to adjust the cable forces and deck elevations to desired values by using experimentally generated response matrices. The relationships between cable forces among themselves or between cable shortening displacement and deck elevation differences are constructed in a matrix format by generating a unit disturbance in all cables, one at a time. The generated experimental matrix contains the actual behavior of the bridge and used to solve for the unknown amounts of disturbance that need to be applied to each cable in order to reach at the target values. In this way, for example, the forces of symmetric cables on each side of a bridge can be adjusted to be equal to each other, or the camber of a bridge can be adjusted to the desired profile. The matrix based modification approach may also be used for analytical model calibration. Condition evaluation or load rating studies on existing cable-stay bridges must be conducted on calibrated analytical models which reflect the existing cable forces under dead load. The relationship between analytically obtained cable forces of a complicated analytical model can also be simplified to a response matrix by analytically post- tensioning each cable at a time and calculating cable force changes in other cables to analytically duplicate the existing (measured) cable forces during model calibration.

149 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

A regularization scheme for displacement reconstruction using measured structural acceleration data

Y.H. Hong Department of Civil and Environmental Engineering, Seoul National University, Seoul, Korea H.W. Park Department of Civil Engineering, Dong-A University, Busan, Korea H.S. Lee Department of Civil and Environmental Engineering, Seoul National University, Seoul, Korea

Time integration schemes based on time-marching algorithms such as the Newmark-β method are probably the most straightforward and easiest way to obtain displacement from measured acceleration. However, the time marching algorithms yield erroneous displacement. First of all, initial conditions on velocity and displacement required in the time marching algorithms are usually unavailable or inaccurate in real situations. Moreover, random noise in measured acceleration data causes physically inadmissible errors in the reconstructed displacement. Particularly, low- frequency spectral components in random noise are amplified during time marching procedures, which severely deteriorate the accuracy of the reconstructed displacement. This undesirable effect becomes a critical issue in the displacement reconstruction for large-scale civil infrastructures, which usually exhibit very low fundamental frequencies. This paper formulates a new class of the displacement reconstruction scheme as a boundary value problem rather than an initial value problem using measured acceleration without any infor- mation on initial conditions. In case measured accelerations are given over a finite time interval referred to as a time window, the relation between the measured acceleration and the defini- tion of acceleration forms a boundary value problem. As the second-order time derivative of displacement is acceleration, the displacement is reconstructed through the minimization of the least squared errors between measured acceleration and the second-order time derivative of dis- placement in a time window. The second-order time derivative is approximated by the central finite difference. As the reconstruction problem of displacement is defined as a boundary value problem in a time window, boundary conditions at both ends of the domain should be specified to solve the mini- mization problem, but neither displacement nor velocity is known at the boundaries. Therefore, the minimization problem for the reconstruction of displacement becomes ill-posed or rank-deficient, and can not be solved for unknown displacement in a time window. Furthermore, a small amount of low-frequency spectral noise in measured acceleration data may significantly pollute the recon- structed displacement as in the time-marching algorithm. To overcome these two difficulties, the Tikhonov regularization scheme, which has been widely employed to alleviate the ill-posedness of inverse problems, is adopted. The 2-norm of the displacement to be reconstructed in a time window is chosen as the regularization function. An overlapping time-window concept proposed by Part et al. is adopted to enhance the accu- racy of reconstructed displacement. The reconstructed displacement only at the center of a time window is taken as the solution of the time window so that the error by inaccurate estimation of boundary conditions should be minimized. Considering the accuracy and computational effort of displacement reconstruction, the optimal time-window size is proposed through a parameter study of SDOF systems.

150 The validity of the proposed method is demonstrated through displacement reconstructions using raw acceleration data measured from laboratory vibration test of a stay cable. It is shown that the proposed displacement reconstruction scheme does not suffer from any instability caused by low-frequency spectral noise, and yields accurate and reliable results.

151 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Dynamic testing of existing bridges for high speed trains

A. Turer Civil Engineering Department., Middle East Technical University, Ankara, Turkey

ABSTRACT: Railway transportation is one of the most economical and efficient ways of trans- porting goods and people in long distances. Railway transportation system in Turkey, unfortunately has not received the interest it deserves except for the early decades of the 84 years old Turkish Republic. Recently, the transportation system inTurkey highly relies on the highway transportation; however, there are now some attempts to renovate the railway system and increase its share in the national transportation system which is commonly observed in the developed countries. Railway bridges are the most important links of a healthy railway transportation chain and should be well inspected and maintained. Railway bridges experience larger moving loads – compared to the high- way bridges – exerted by heavy and fast moving train engine and locomotives, each one exceeding 120 tonnes of mass. The condition evaluation of train bridges gained even more importance as the train systems are improving towards speed railway trains which are much faster than the conven- tional trains, easily exceeding 200 km/hr. The measurement of vibration level during conventional train passage was conducted in two steel bridges with different span lengths (8.75 and 50 m) and structural types (beam and truss). Wireless accelerometers were used for data collection. Difficul- ties with the triggering system and data storage were addressed. Simple interactions between train speed and transverse beam spacing were evaluated for possible resonance conditions. Only the first one or two modes shapes were able to be identified due to the used sensor properties. The processed results showed that the vertical accelerations are higher than practical limits of vibrations accepted for train bridges in the range of 0.5 g to 3 g. Close-to-resonance condition was detected in the truss bridge using the measured data, sensor locations, and bridge geometry. Train crossing speeds of 95, 128, 150, 211, 259, and 320 km/hr were found to be critically matching the bridge vibration modes having potential for resonance. The first (beam type) bridge had transverse beams at every 1.75 m intervals likely to cause resonance condition at 121 and 286 km/hr train speeds which would excite natural bridge modes identified at 19.19 and 45.45 Hz. The bridges are expected to have large deformations (aggravated by low damping ratios) in the resonance state when used for fast trains matching or getting close to the indicated velocities. Simple evaluation results have supported the findings obtained from measurements indicating that existing bridges would likely run into serious resonance, safety, and serviceability problems if tried to be used for fast or high-speed trains. Usage of wireless sensors in bridge monitoring under service conditions has great advan- tages such as speed of instrumentation and ease of data collection; however, the noise to signal ratio, triggering, and data downloading speed issues should be improved for ambient monitoring applications.

152 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Concrete bridge deck condition assessment with automated multisensor techniques

Dryver Huston, Jianhong Cui & Dylan Burns School of Engineering, University of Vermont, Burlington, VT, USA Frank Jalinoos Federal Highway Administration, McLean, VA, USA

The paper presents a study aimed at developing effective automated condition assessment methods for concrete bridge decks. Delaminations, reinforcing bar corrosion and bulk material properties are all important subsurface features with a direct effect on bridge deck longevity. If detected early, ameliorative actions can arrest the damage progression. If detected at a later stage of growth, accu- rate prognoses can enable effective and low-cost rehabilitation efforts. The ideal sensing system would be one that can detect, classify, locate and identify subsurface damage, operate at low- cost with modestly-skilled personnel, cause a minimum amount of traffic disruptions and produce results with a minimum of uncertainty. At present, no such instrument exists. Instead, a variety of available sensing techniques can partially fulfill these requirements. These include: electro- magnetic – Ground Penetrating Radar (GPR), acousto-elastic – chain drag and impact-echo, and electrochemical potential – half-cell. All of these methods have strengths and weaknesses. The strengths are that certain physical effects, such as delaminations or punky concrete, can give dis- tinct readings on the instruments. Weaknesses include uncertainty as to the meaning of some of the results – especially due to confounding effects. Many of the instruments are manually operated and require dangerous traffic disruptions. This study attempts to improve the state of the art by a multipronged approach: 1. Identify the strengths and weaknesses of each of the available testing methods. 2. Automate the sensing process to increase the testing speed and reduce the possibility of human error. 3. Fuse data from multiple sensors to create a superior quantification of damage and underlying conditions. The instruments under consideration include imaging GPR, chain drag, impact echo, laser ultrasound, and half-cell. The field testing results shown in this paper are from the Van Buren Road Bridge in Dumfries, VA, USA. The testing compares five different methods: 1. Visual inspection and photographic recording of position; 2. Half-cell electrochemical poten- tial; 3. Impulse type multipoint scanning ground penetrating radar, i.e. the HERMES/PERES II system; 4. Chain drag; and 5. Impact echo, i.e. Portable Seismic Pavement Analyzer (PSPA). The initial results of the tests were that each instrument nominally performed and collected data as expected. The data is registered, overlaid and compared. The potential for developing automated multi-sensor systems that fuse data for efficient and effective bridge deck measurements will also be discussed.

153 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Parameter estimation of concrete bridge using ambient acceleration measured by wireless measurement system

S.J. Lee, S.B. Kim, K.Y. Choi & G.Y. Song Institute of Construction Research, Samsung Engineering and Construction, Gyeonggi, South Korea D.O. Kang Palm Jebel Ali Project, Civil Works Division, Samsung Engineering and Construction, Dubai, UAE Y.H. Lee Korea Maintenance & Control Co., Ltd, Seoul, South Korea

In this study, the ambient vibration responses of a PSC bridge under wind load are measured by both wired and wireless acceleration measurement systems. The parameters of a marine concrete bridge such as natural frequencies, mode shapes and elastic modulus of the bridge are estimated to analyze the behaviour of a real bridge compared with that of the FEM model used in the design of the bridge. First, in order to verify the parameter estimation procedure used in this study, acceleration data of the numerical model under impact load are measured by the numerical time-history simulation. Also, the modal parameters and the elastic modulus of the model are estimated. In this paper, two modal parameter estimation methods are used and compared: one is FDD (Frequency Domain Decomposition, Otte et al., 1990) method, and the other is SSI (Stochastic Subspace Identification, Overschee et al., 1996) method. The elastic modulus of the bridge is estimated iteratively by minimizing the modal error between the measured natural frequencies and the natural frequencies of the FEM model. The modal parameters and the elastic modulus of the numerical model estimated by the used procedure are similar to the real values of the numerical model. For the real bridge, the modal parameters are obtained from ambient vibration signal under wind loads using both the wired and the wireless acceleration measurement systems. The results from the FDD and the SSI methods are almost equal, thus the reliability of these two methods are verified. Also, the identified results show that the modal parameters from the two measurement system (wire and wireless) are very similar which means that possibility of application of the wireless acceleration measurement system are validated in the construction field of coastal environments. From the identified modal parameters, the elastic modulus of the bridge is estimated and compared with the results obtained from the stress/strain gauge, compressive strength test and the FEM model used in the design of bridge. As a result, the estimated natural frequencies and the elastic modulus of a constructed bridge are relatively larger than those of the FEM model in the design. However, the results from the static load test (stress/strain gauge) and compressive strength test (28 curing-days) show that the estimated elastic modulus is a reasonable value.

154 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Impact-Echo scanning for grout void detection in post-tensioned bridge ducts – Findings from a research project and a case history

YajaiTinkey & Larry Olson Olson Engineering

ABSTRACT: This paper presents the findings from a research project funded by the NCHRP – IDEA Program. This paper discusses the experimental results from the studies of impact-echo tests which involved a defect sensitivity study of an Impact-Echo (IE) Scanner to detect and image discontinuities in post-tensioned ducts of a mockup U- shaped bridge girder and a mockup slab. Different sizes of ducts were included in this study as well as varying sizes of void defects. Detailed sensitivity study of non-destructive grout defect detection with Impact-Echo Scanning of 8-four inch diameter ducts with constructed defects was the main focus in this study. Comparisons of the IE defect interpretation and the actual design conditions of the ducts inside the bridge girder/slab are presented. The IE results are presented in a three-dimensional fashion using thickness surface plots to provide improved visualization and interpretation of the internal grout to void defect conditions inside the ducts of the girder. The Impact-Echo tests were performed with a Scanner which greatly facilitates the Impact-Echo test process by allowing for rapid, near continuous testing and true “scanning” capabilities to test concrete structures. The paper summarizes the general background of the Impact-Echo technique and the Impact-Echo Scanner. Descriptions of two mock-up specimens used in the experiment and the discussion of the results from the Impact-Echo Scanner are presented herein. Finally, a case study using an Impact Echo Scanner to locate grout voids inside the Orwell Bridge in UK is included in this paper.

155

Bridge management systems

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Design and implementation of a new bridge management system for the Ministry of Transport of Québec

Reed M. Ellis Stantec Consulting Ltd., Canada Paul D. Thompson Consultant, Canada Rene Gagnon & Guy Richard Ministère des Transports, Direction des structures, Canada

The Ministry of Transport of Québec (MTQ) owns and maintains an asset inventory of over 9000 structures, of which 4,300 are provincial bridges, 4,400 are municipal bridges, and the remaining are retaining walls and other miscellaneous structures. These structures are managed in a largely decentralized process by the Head Office and 14 Regional offices. As is the case with other agencies, the bridge inventory in Quebec is aging and the average bridge condition constantly challenges the financial and human resources of the Ministry. In its continuing effort to ensure a safe, efficient transportation infrastructure for the people of Québec, MTQ has been developing a new bridge management system. Known as Système de Gestion des Structures (SGS), the system is a new state-of-the-art bridge management system that has a number of novel features designed to meet the specific needs of MTQ:

• A decentralized software architecture, using Microsoft’s.NET framework, for maximum sharing of inventory and inspection data across the province, with more controlled access to the strategic planning analysis; • A graphic user interface provided primarily in French with certain bilingual and localizable features; • Support for a detailed bridge inspection methodology that is span-by-span, but has elements and condition states similar in concept to the Ontario Structure Inspection Manual, with separate inspection of protection systems. • Strategic planning analysis featuring several levels to fit MTQ business processes: Network- level budgeting and performance analysis; priority programming; automated project scoping and treatment selection; and a digital dashboard for interactive design of the scoping and timing of projects. • The model framework handles preservation, functional improvements, and replacement; pro- vides explicit control of the element and project alternatives to be considered; and has features to update deterioration models based on new inspection data.

Begun in Fall 2005, the SGS has an Inventory and Inspection Module, which was delivered in early 2007, and an advanced Strategic Planning Module (MPS) due to be released in early 2008. The SGS combines lessons learned from the Ontario Bridge Management System and several recent research projects into a new system that will be highly responsive to the Ministry’s needs. The SGS is the first French language BMS of its kind. The Système de Gestion des Structures will replace an older system that handled inventory and inspection data. The SGS will provide a new inventory and inspection system and add analytical capabilities to enable MTQ perform bridge level and network level analyses. The system includes an advanced Strategic Planning Module (MPS) which incorporates a set of three analysis levels 159 which are designed to stay consistent with each other, so if a change is made at the element level, the project and network levels will adjust appropriately. The budgeting and performance measure setting that is done at the network level has an immediate effect on selected projects and the results of project selection can be viewed in tabular or graphic form. The MPS incorporates an innovative digital Tactical Planning Dashboard which complements the MPS by providing a purely bridge- level perspective on the planning of future work and enabling detailed study of project analysis results and other information only available in several different screens in MPS. This paper highlights the design of the Système de Gestion des Structures (SGS) and presents some of the challenges faced by the Ministry of Transport, and gives an overview of the analytical features of the system’s analytical engine, the Strategic Planning Module.

160 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Comprehensive lightning protection technologies for mechanical and electrical systems of Sutong bridge

Bei Yao & Weisu Zhang Jiangsu Provincial Sutong Bridge Construction Commanding Department, Nantong, Jiangsu, P.R.China Guangchang Chen & Cuihong Jiang Jiangsu Provincial Lightning Protection Center, Nanjing, Jiangsu, P.R.China

ABSTRACT: Sutong Bridge is located in the south of Jiangsu Province. Its mechanical and electrical systems are all connected as a whole, therefore highly likely to subject to lightning strike, causing damage to equipment and even result in failure of systems. Therefore, complete comprehensive lightning protection provisions are vital to the safe operation of the mechanical and electrical systems of the bridge. Currently, the main problems in the comprehensive lightning protection for the mechanical and electrical systems of the bridge include the following: lacking or incomplete provisions against direct lightning; arrangement and shielding of LVpower supply lines and signal transmission lines, installation of surge protection devices, equal-potential connection of equipment and common grounding system are not to the specification; and no shield protection at all for equipment rooms. To ensure the safety of the mechanical and electrical systems of the bridge after it is put into service and avoid damage by lightning, it is necessary to implement comprehensive lightning protection design and construction according to the characteristics and actual lightning protection requirements for these systems. They include two parts: protection against direct lightning and protection against induced lightning. Normally the three-level lightning protection configuration is adopted for LV power supply systems, energy coordination should be noted in selecting the SPD’s, and the technical parameters of SPD’s of different classes shall conform to the actual lightning protection and specification requirements of the equipment. The lightning protection for signal transmission systems mainly includes protection for equip- ment and various signal transmission lines in the monitoring system, charging system and communication system. In the comprehensive lightning protection systems, shielding provisions are extremely important. This should be completed as far as possible in engineering design and construction. All line conduits (including those for power supply lines and signal lines) should be changed to metal pipes, which shall be reliably grounded; the metal doors and windows of center equipment rooms and metal housing of equipment should be reliably grounded; further shielding must be provided for important equipment rooms; all equipment or metal housing and cases of equipment of or inside the charging kiosks should be reliably grounded, and the grounding wires should be as short as possible. In general, modern lightning protection is of system engineering nature. In the implementation of the project, emphasis should be placed on all-round protection, comprehensive control and rectification and protection at all levels; in the meantime, specific provisions must be made for special problems, to eliminate factors that may result in lightning damage one by one. Only in this way can we minimize the hazard of lightning attack.

161 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Risk evaluation and management for road maintenance on urban expressway based on HELM (Hanshin Expressway Logic Model)

Yasuhito Sakai Hanshin Expressway Company Limited, Osaka, Japan Kiyoshi Kobayashi Graduate School of Management, Kyoto University, Kyoto, Japan Haruhiko Uetsuka CTI Engineering Co., Ltd., Osaka, Japan

Hanshin Expressway is an urban expressway network accommodating more than 900,000 vehicles per day. To sustain efficient and comfortable road traffic environment, Hanshin Expressway is performing various highway Rehabilitation and Maintenance (R&M) activities. The well-organized highway R&M activities (ex.; regular inspection, roadway cleaning, etc.) is fundamental for keeping traffic safety and provide road users with psychological satisfaction, but poor R&M may cause serious traffic accidents or defects. Due to the lacks of the sufficient budget for highway R&M, the highway administration bodies are requested to reduce the expenditures for R&M activities. In this paper, the logic model for R&M (named HELM; Hanshin Expressway Logic Model), which was developed by the Hanshin Expressway Corporation, is presented to administrate the whole system of R&M activities in an efficient way. The model will turn out to be the indispensable platform for the introduction of R&M contracts with the outside sectors in the near future. Some evaluation indicators for policy models comprising HELM, which are expected to be monitored and evaluated throughout the implementation of the PDCA cycles, are also illustrated in the paper. This paper shows the methodology to formulate the criteria based on risk (probability times consequence) measurement for the risk management of the potholes, sands and pebbles, upon the pavement. The criteria in this paper consist of the average line and the margin area considered variability or uncertainty factor. The criteria presented can be estimated based upon the past record on highway patrol. The risk management diagrams for the highway patrol are also presented, by which the administrative bodies can efficiently design the optimal patrol frequency given the predetermined risk levels. The paper is concluded by summarizing the ongoing research agenda to improve the HELM throughout the PDCA cycles of the real R&M activities.

162 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Asset management system development

J. Radic University of Zagreb, Faculty of Civil Engineering & Civil Engineering Institute of Croatia, Zagreb, Croatia J. Bleiziffer University of Zagreb, Faculty of Civil Engineering, Zagreb, Croatia G. Puz Croatian Motorways Ltd., Zagreb, Croatia

Over the last decades Croatia made large investments in order to complete the 1500 km long motor- way network, which shall integrate the most distant parts of the country and provide quick, easy and comfortable transportation of people, goods and services. The Croatian motorway network may not seem long, but it is worth mentioning that Croatia already has more kilometres of motorways per 100 000 citizens than UK, Ireland, Greece or Italy. According to annual figures, the construction industry’s share of Croatian GDP was 5.9% in 2006. Broken down by type of structures, the com- position shows that transport infrastructure accounted for 39.5%. Such large investments require a carefully devised and above all economical management of the existing stock. The recently completed study on asset management system for Croatian motorways was pre- pared by the Structural Department of the Zagreb Faculty of Civil Engineering for the client Croatian Motorways Ltd. aimed at development of a management system that would properly address maintenance and management of all motorways structures. The study established the framework for development of an asset management system including all physical assets on Croatian motorways – bridges, tunnels, pavements, drainage, geotechnical structures, road equipment and buildings, taking a comprehensive view to evaluate all the trade-offs that must be made. Broader approach in asset management provides framework for cost-effective decision-making and quality improvements. It also serves as an important process in short-term, long-term and strategic planning. The framework for Croatian motorways asset management system was developed on following principles: (1) asset management system should provide for daily maintenance and management activities (operation, maintenance, inspection, repair etc.) and (2) asset management system should provide support for agency’s decision in resource allocation (applying objective information to decision-making, considering alternatives, focusing on outcomes. Another principle was to identify common attributes of a variety of structures on Croatian motor- ways covered by asset management system in order to integrate the management of all motorway structures in making decisions across assets. The management system should also allow for under- standing implications of different levels and distributions of expenditures across geographic areas and sections of motorway network. Asset management system for Croatian motorways has to be compatible with existing information systems of Croatian motorways Ltd. – the entries in a database of road information, the content of which was agreed upon between different Croatian road authorities. Guidelines for development of individual components and modules of the system were defined. Further work on developing asset management system for Croatian motorways should follow shortly with drafting the standardized procedures for inspection, assessment and repair of structures (incl. comprehensive catalogue of possible damage and defects on structures), cost analysis and cost- benefit module, future condition forecasting algorithms, and criteria for deciding on an optimum maintenance strategy.

163 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Development of a smart-client based bridge management and maintenance system for existing highway bridges

Deshan Shan & Qiao Li Civil Engineering School, Southwest Jiaotong University, Chengdu, Sichuan, China

Bridge maintenance costs have increased in China to the point where it is now more expensive to maintain existing highway bridges than to build new ones since 1990s. Several bridges have deteri- orated considerably in recent years due to factors such as the increase in traffic volume, the increase in heavy vehicles and structural aging. Depending on the severity of deterioration and availability of limited funds, repair or strengthening of these bridges is essential. Fortunately, Chinese high- way network is comparatively newer than those of other countries, because several thousands of highway bridges have been constructed through the National Highway Network Project, which was initiated in 1950s. However, a report has indicated that by approximately 2015, approximately 40% of bridges will be 60 years of age or older. Therefore, the development of a comprehensive bridge management and maintenance system for existing highway bridges is vital. According to the demand of current operation and management of the existing highway bridges in China, the objec- tive and the basic framework of the smart-client based bridge management system is analyzed in this paper. And the model design of each subsystem is introduced from the following perspectives: smart-client network support, central database designing and relationship among the data, bridge managements of project-level and network-level. Besides, the latest mathematical method of data mining and artificial intelligence is employed in optimizing decision analysis. Because the bridge management on project level and network level are complementary to each other, and reflect the different requirements of traffic on the serviceability and safety of existing bridges, a integrated implement of the Bridge Management System can be met the real bridge engineering requirements on the bridge management system by the multi-tier framework proposed by Microsoft in 2004. Because the database structure and relationship among the data lies of great significance, which is related to the effective data transfer and the realization of data sharing, 84 data tables, 184 stored procedures and 12 data cubes were carried out in the database to meet the requirements of the Data. Based on the Dotnet 2.0 frameworks, the Bridge Inspection Rating Maintenance and Management System (BirMMS) is developed by employing object-oriented programming technology aided by theVisual Studio 2005 C# and SQL Server 2005. Smart-client application combined the advantages of both Fat Client and Thin Client application can not only make full use of local resources but network resources, so the Smart Client programming technology was adopted during the system development. Data mining and data analyzing function as online data processing (OLAP) offered by SQL Server 2005 were also used in BirMMS to aid the decision making of assessment for bridge operational condition and the allocation of the maintenance fund. And report service also offered by SQL Server 2005 is adopted to fulfill the report of bridge condition assessment and maintenance fund allocation optimization. In order to make the bridge maintenance systemize and program, realize the bridge management on the project level and network level and distribute rationally the maintenance fund, the system has been successfully applied to the management of three long-span bridges (one is cable-stayed bridge, the other two are continuous rigid frame bridges) of certain expressway in Chongqing, China. This guarantees the three bridges effectively and safely serve the traffic demand. Through verification of three long-span bridges of certain expressway, the sys- tem achieves the aim of systemizing and programming of the bridge maintenance, and rationally distributing the maintenance fund.

164 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

A new Bridge Management System for the National Department of Transportation of Argentina

Miguel E. Ruiz National University of Córdoba, Córdoba, Argentina Eduardo A. Castelli National Department of Transportation of Argentina (Dirección Nacional de Vialidad), Tucumán, Argentina Tomás A. Prato National University of Córdoba, Córdoba, Argentina

ABSTRACT: Highway bridges are valuable capital assets from both social and economic points of view and their correct performance has direct implications on economic, social, cultural, and military activities of a country. Bridge structures have been found to present a particular com- bination of design challenges that are not usually found in other civil works; such as structural, hydraulic (for river bridges), and traffic safety issues. Thus, in order to assess the adequacy of a bridge, its condition should be considered from a global point of view, considering the three afore- mentioned aspects. The present approach considers that a bridge in excellent condition implies that it simultaneously presents appropriate structural and hydraulic conditions and ensures safe circulation. In order to aid authorities in the rational budget allocation for maintenance and retrofit works of existing bridges, a new Bridge Management System (BMS) was developed as the result of a joint effort by the National University of Córdoba and the National Department of Transportation of Argentina (Dirección Nacional de Vialidad, in spanish). This paper provides an overview of the recently developed BMS methodology, highlighting the particular characteristics that differentiate the Argentinean BMS from others systems currently in use around the world. The newly developed methodology provides an overall assessment of the bridge condition from a global point of view, accounting for structural, social, environmental, traffic safety, and economic aspects and considering the bridge as part of a highway system. The proposed evaluation methodology is based on a group of failure risk and failure consequences indicators, which assign a global condition grade for the bridge. Grades are assigned to the three failure risk indicators to individually assess the structural, hydraulic, and traffic safety condition of the bridge. The weighted average of these grades gives the bridge Failure Risk Grade (FRG). All bridges having a FRG equal or greater than 8 are deemed to present an acceptable global condition. In the other hand, bridges presenting FRG lower than 8, are deemed to present unacceptable condition, thus the consequences of bridge failure should be studied. The assessment of the bridge failure consequences (through the failure consequences indicators) allows estimating an overall bridge condition grade. This grade is the principal parameter upon which a priority list for the execution of rehabilitation, maintenance, or retrofit of existing bridges of the Argentinean National Highway System is defined.

165 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Development of a bridge maintenance decision support module for Taiwan Bridge Management System

Hsien-Ke Liao Department of Civil Engineering, National Central University, Taiwan Chung-IYen & Nie-JiaYau Institute of Construction Engineering and Management, National Central University, Taiwan

ABSTRACT: Taiwan Bridge Management System (T-BMS) is a web-based bridge management system developed in year 2000 for the bridge management agencies in Taiwan. T-BMS has been on-line for nearly eight years. Having more then 25,000 bridges in its inventory, T-BMS has thousands of logins per month to update data in its relevant databases. The major functions of T-BMS were focused on collecting and recording inventory, regular inspection, and repairing information for the bridges in its early stage. In recent years, functional improvements of T-BMS are aimed at using inspection data to provide information that helps those management agencies in making strategies and decisions for bridge maintenance. This research first discusses the DER&U methodology that is currently utilized as a bridge visual inspection standard at Taiwan, as well as the needs for a decision support tool. A decision support module that incorporates the existing bridge inventory, regular bridge inspection data and repairing methods in T-BMS is designed and developed in this research. The decision support functions include (1) budgeting for bridge inspections, (2) suggesting repairing methods, (3) budgeting for repairing bridges, (4) classifying deteriorated bridges by degree of danger, (5) prioritizing repair works, and (6) distributing funds for repairing. A new index DD (degree of danger) is also introduced in this research to classify the safety level of a bridge. This module will serve as a decision tool not only helping personnel who are responsible for bridge inspection and repair tasks but also managers for developing budgeting and maintenance strategy in the long run.

166 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Optimization of bridge management policies on the French national roads network

N. Odent & J. Berthellemy French Ministry for Ecology and Sustainable Development and Spatial Planning, Setra, Bagneux, France C.F. Cremona & A.D. Orcesi Laboratoire Central des Ponts et Chaussées, Paris, France M. Toriel French Ministry for Ecology and Sustainable Development and Spatial Planning, General Highways Department, Paris la Défense, France

In 1994, the French HighwayAgency initiated the condition assessment of the national bridge stock, based on the scoring system named IQOA (Quality Image of Engineering Structures). Since this year, bridge condition is assessed every 3 years, on the basis of an inspection campaign performed every year on a third of the stock. Structures are thus ranged according to five IQOA scores whose meaning is explained in Table 1. The resource allocation to the local agencies however continued, until 2006, to rely on ratios such as the bridges deck areas without efficiently using the results from the IQOA condition. However, maintenance actions for the most damaged engineering structures, accompanied with periodic or enhanced bridge surveillance, allowed maintaining a good safety level and conservation of the bridge stock. This maintenance policy can be nevertheless improved. It is more pertinent to correlate financial needs with the condition of the bridge stock in a better way. Besides, a new law of finances reformed deeply the management of the State in January 2006, by introducing the concept of budget performance. The Ministry must therefore justify the efficiency of its bridge stock management strategy in the eyes of the Ministry of budget, with an annual draft and report of performance showing the evolution of quality indicators for requiring financial needs. Facing these stakes, the French Highway Agency requested Setra and LCPC to analyze policies and to assess the quality indicators evolution until 2020. The scope of this paper is to present the results of this analysis for the next fifteen years. Setra developed a simulation model of the ageing of bridges, based on the observation that there is a very good correlation between the age and the condition of bridges. The observation of the trend curves between the ages and the IQOA grades makes it possible to deduce some degradation profiles according to a Poisson process: on this assumption, for a population of bridges, the probability of transition from a score to a worse score is independent of the past life of the bridges.

Table 1. IQOA scores.

Score Apparent condition

1 Good overall condition 2 Equipment failures or minor structure damage. Non urgent maintenance needed. 2E Equipment failures or minor structure damage. Urgent maintenance needed. 3 Structure deterioration. Non urgent maintenance needed. 3U Serious structure deterioration. Urgent maintenance needed.

167 Figure 1. Evolution of quality indicators with Figure 2. Evolution of quality indicators with scenario 4 and Poisson process analysis. scenario 4 and Markov chains analysis.

LCPC built the annual transition matrix directly from the IQOA collected scores. The Markov assumption is used stating that the condition of a facility at the inspection i depends only on its previous condition at inspection i − 1. With this assumption, the present score is the only one which is taken into account to determine the future of the bridge. The access to the rating data of approximately 9,000 bridges between 1996 and 2005 makes it possible to determine the probability for 1 m2 of bridge to move from one condition rating to another one within one year. Four scenarios were defined, showing possible strategies to be adopted within the next fifteen years. These maintenance scenarios emphasize either preventive actions, or curative ones, or both, to control the annual budget and to keep the bridge stock in a good condition. The fourth scenario simulates the following situation : a constant annual budget is fitted and is distributed on all IQOA scores, verifying the quality indicators. Some results of the scenario 4 are shown in figures 1 and 2 respectively for both models. Lessons for defining the maintenance policies can be deduced from the simulation results. Even if they diverge regarding budgets, the study of the four scenarios allows to observe that continuing a policy only based on corrective actions is not the best strategy for maintaining a sufficient level of preservation and safety of the bridges, that the order of magnitude of the present budget allocated to maintain bridges on national road network in France is good, but that an important reduction of financial needs would lead to a significant degradation of the condition of the asset. Finally, simulations of evolutions associated to an economic approach enable to determine pertinent maintenance strategies for the management of a bridge stock, i.e. those that maintain the bridge stock in a good condition and that control the repair costs.

168 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Bridge management: A challenge for local authorities

B.M. Kamya Kamya Consulting Ltd, Aberdeen, UK

The management processes for bridge owners can be broadly categorised as Strategic, Tactical and Operational. These processes are often not well integrated and quite often lead to inconsistent decisions on future management objectives. The strategic policy for asset management is usually developed in consultation with stakeholders in order to establish the overall long term strategic goals and objectives. The tactical process is the formal planning to identify the beneficial and cost effective way to manage the bridge assets. The operational process involves engineering such as inspection, structural assessment, routine maintenance, scheme design and work scheduling. It focuses on choosing the right techniques, Value Engineering of schemes and carrying out the work in the most efficient way. This paper uses examples of bridges assessed as part of a routine load assessment and for a specific loading requirement for a local authority in the United Kingdom UK to discuss some of the challenges experienced by local authorities in managing their bridges. This paper reviews the current practise in the asset management in the United Kingdom UK. In so doing, an examination is carried out of the strategic, tactical and operational process used by the local authorities in identifying how decisions to assess these particular bridges were arrived at. It is hoped that this experience may strike a cord with other bridge owners in identifying that in times of restricted budgets, competition from other departmental needs, recognising the most efficient way of decision making could improve the ways that bridge owner authorities can improve the processes of asset management. Reference is made to the guidance standards and advice on the asset management in the UK. In conclusion the paper proposes that the processes of decision making are similar for most organisations and such sharing of experience can help others in identifying better and efficient approaches.

169 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

A condition index based on the concept of apparent age

D. Zonta, F. Bortot & R. Zandonini Department of Mechanical and Structural Engineering, University of Trento, Italy

In Bridge Management Systems based on AASHTO Commonly Recognize Standard Elements, the Condition State (CS) is defined at the element level only. However the system operation often requires definition of an index somehow representing quantitative measurement of the global deterioration of the bridge. This request is typically answered by introducing heuristic condition indices, which, for example, mix with appropriate weighting numerical values associated with the CS of each element. In this paper we introduce a non-heuristic global condition index based on the concept of Apparent Age. The Apparent Age of the bridge is defined as the most likely age of the bridge, given the condition state of its elements. As for individuals, Apparent Age allows an estimate of the degree of aging of the subject. Compared to the real age, Apparent Age takes account of whether the bridge deterioration is normal, or if the structure has suffered abnormal degradation. We define the Apparent Age of a Standard Element as the most likely age of the element given its CS, assuming theoretical age distributions that are consistent with the normal deterioration model adopted for the element. The model takes into account that the element, when deteriorated, is typically replaced or demolished. The age distribution of a set of geometrically identical elements follows immediately from the definition of conditional probability.The size of an element impacts the perception of degradation, and this is accounted for, by introducing exponential weights proportional to size. To estimate the age distribution of whatever combination of elements, as in the case of a bridge, we must extend the meaning of these exponent weights, also considering the different psychological impact that different types of element have on the overall perception of the bridge deterioration. This requires definition of a model that quantifies the relevance of the element in the bridge from the owner/manager perspective. This model should reflect the subjective view of the final user of the BMS and has to be chosen and calibrated consistently with the management policy of the agency. The calculation of the Apparent Age is described with a number of practical examples, taken from the bridge stock of the Italian Autonomous Province of Trento. The overall distribution of the Apparent Ages of the APT bridges respect to the actual ages is also analyzed. As expected, we observe that the older the bridge, the higher its Apparent Age, although the scatter from the average trend is very significant. The poor correlation between age and real condition suggests we must be very careful when using deterioration models for predicting the future state of a single bridge. Also, this analysis immediately lets us recognize the bridges that exhibit abnormal deterioration. In any case, it is worth remembering that the Apparent Age indices calculated are sensitive to the selected normal deterioration model: this model is merely conventional, and should be calibrated on the expected lifespan of the bridge under normal maintenance conditions. This note emphasizes that the Apparent Age is by nature a relative rather than an absolute index.

170 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Design of the standardized measuring system for the integrated safety management of bridge structure

W.S. Lee & K.T. Park Hybrid Structure Research Division, Korea Institute of Construction Technology, Korea

In this study, the research was carried out to integrate the bridge measuring systems that have been established independently during that time for the purpose of safety management of bridge structure in efficient and economic approach. At present, it is not easy to integrate them because any specific standard to select the measurement item, measurement location and amount of sensor are not defined yet in terms of the bridge measuring systems, and therefore these are prepared in different type depending on the technical level of the system designer and its budget. In this study, specification and standard were defined according to the type, scale, condition and place of the bridge in terms of the measurement item and sensor location of the bridge measuring system so that the automated design of the measuring system through the development of algorithm (Fig. 1).

Figure 1. Improved measuring system design algorithm.

171

Bridges for high-speed railways

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Steel bridges for high speed railways – Design regarding fatigue and durability

W. Hoorpah Bridge consultant MIO, Paris, France

New HSR lines with steel and composite bridges are in service in France since the 1990’s. New lines are under construction now or in the design phase. This trend is also observed all over Europe and the European network is gradually getting mature with interoperable traffic. In France, the bridge decks were exclusively in concrete in the first three HSR lines; steel first appeared in the 1990’s in the TGV North. The percentage of composite steel decks has since been rising regularly and has reached 100% in the medium and large span bridges for the TGV East line opening now. The increase in steel bridges is due to: – the design of efficient composite steel concrete decks, – better design for dynamic behaviour and fatigue consideration, – specific fabrication details, – structures that can be easily maintained and inspected, – well guaranteed durability. The bridge deck structures being built now are the results of recent experience concerning design based on Eurocodes ensuring satisfactory dynamic behaviour and criteria of safety, passenger comfort and structural durability. They have proven to be economical for fabrication and mounting. This paper gives an overview of HSR bridges evolution in France with recent examples of steel and composite bridges in all span ranges: small, medium and large span. The typical deck structures described are: the twin girder composite deck, the four girders composite deck, the half through deck, the twin box girder. For longer spans reaching 100 m and above: tied arch decks and truss box girders in some cases. The fatigue sensitive details are detailed for each type of deck, with regards to their dynamic behaviour under live loads. The structural evolution is towards more robust structures, with the required stiffness for HSR loads, more economical to fabricate, mount and maintain.

Keywords: High speed railways, steel, composite decks, dynamic behaviour, fatigue, stress concentration, maintenance, durability.

175 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Dynamic response of the Cahir Viaduct – An investigation into the derailment of a freight train

M. Majka Structural Design Section, Irish Rail, Dublin, Ireland M. Hartnett Department of Civil and Environmental Engineering, National University of Ireland, Galway, Ireland

The paper addresses a railway accident that occurred at a railway bridge across the River Suir in Cahir, Ireland. A fully laden cement train derailed immediately after locomotives had crossed the bridge. In effect, a part of the bridge deck experienced significant damage and fell into the river below together with the rear twelve wagons of the train. The paper briefly summarises the circumstances of the accident, as well as the findings of the previous investigations that followed the derailment. Despite different hypotheses being formulated in these investigations, none provided conclusive prediction of the cause of the derailment. An objective of this paper is to study effects of the bridge-train interaction and resonance, as well as their possible contribution to the derailment. Numerical analysis of the dynamic response of the CahirViaduct subjected to the derailed train is undertaken using the finite element based bridge-train interaction model DBTI (Dynamic Bridge Train Interaction) developed and extensively verified by the authors. Within the DBTI model, bridges are represented by beam-type finite elements and vehicles are modelled as multi-body dynamic systems. The solution is provided by means of the direct integration Newmark’s method combined with an iterative scheme to incorporate bridge-train interaction. The analysis has shown that some indicators of increased dynamic effects are visible in the bridge response; however, the magnitude of this increased response is rather minor and could not solely cause the derailment of the train. Nevertheless, the magnified dynamic response, in combination with other factors identified in previous investigations, could increase the risk of derailment. It is the conclusion of this paper that bridge-train interactions and resonance were unlikely to cause train derailment and, consequently, the collapse of the Cahir Viaduct. However, the dynamic effects, in combination with other factors, could have contributed to the derailment.

Figure 1. The Cahir Viaduct following the derailment. 176 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

A comfort limit for evaluating the serviceability due to bridge vibration

B.G. Jeon & N.S. Kim Pusan National University, Busan, South Korea S.I. Kim Korea Railroad Research Institute, Uiwang, South Korea

ABSTRACT: In general, deflection limit criteria of bridge design specifications have been con- sidered based on static serviceability and structural stability. Dynamic serviceability induced from bridge vibration, as a comfort limit, actually has not been included in the criteria. Thus, it is necessary for comfort limit to be considered in order to check dynamic serviceability on bridge vibration. In this study, the comfort limit of bridge structures based on the root-mean-quad (RMQ) and vibration-dose-value (VDV) considering the signal fluctuation effectively and the time duration exposed has been constructed. Figure 1 shows the time-dependent comfort limit on bridge struc- tures. The comfort limit developed in time domain was verified by using vibration signals directly measured from the existing bridges. Comparing the developed comfort limit with the conventional ones defined in frequency domain, it is shown that the comfort limit developed in time domain would be more feasible for evaluating quantitatively the serviceability due to bridge vibration. It was found that one of urban railway bridges had a value exceeding the comfort limit with increasing the time duration. Thus, a stiffness limit satisfying the bridge vibration serviceabil- ity was numerically estimated by dynamic analysis on the interaction between train vehicles and bridge. Figure 2 shows the comparison of ISO comfort limit with evaluation of dynamic service- ability by stiffness changes. From the results, a new deflection limit on bridge structures satisfying the vibration serviceability could be proposed by comparing with the conventional deflection limit criteria.

Figure 2. Comparison of ISO comfort limit with Figure 1. Comfort limit on bridge structures. dynamic serviceability by stiffness changes.

177 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Vibration control through TMDs in high-speed railway bridges

J.F. Henriques & J.M. Proença ICIST/IST, TU Lisbon, Lisbon, Portugal

ABSTRACT: This paper focuses on the use of Tuned Mass Dampers (TMDs) applied to high- speed railway bridges in order to control train-induced vibrations by providing auxiliary energy dissipation when resonance between the external train excitation and the structure occurs. This control device presents as advantages the absence of external power source, permanent service time and easy maintenance, although, on the other hand, the possible detuning effects may have significant detrimental effects on the bridge response. The effectiveness of this class of devices is analyzed through a comparison between the numerically simulated dynamic responses of a continuous precast railway bridge located in the Alcacer-Valencia branch of the Spanish high- speed network. It is composed by four spans of lengths 26.5, 33.5, 33.5 and 26.5 m, respectively, in a total of 120 m. The cross-section of the deck is composed by two parallel precast concrete U-shape beams, with a cast in-situ slab, carrying two ballasted tracks. A complete set of dynamic analyses was considered, including all ten trains of HSLM-A defined in EN1991-2 (2003) and the real trains ICE2, ETR-Y, EUROSTAR, TGV, TALGO, THALYS, and VIRGIN. Results showed that the maximum acceleration, equal to 3.438 m/s2, almost in the limit of 3.5 m/s2 for ballasted tracks as prescribed in EN1990-Annex A2 (2005), is obtained with the HSLM-A7 train, being the dynamic results obtained with the HSLM-A an envelope of the homologous results computed with the set of real trains. Three different combinations of passive TMDs were considered. In the first, it was assigned one TMD tuned to the first vertical bending mode to each girder in the interior spans, at the location where the first mode displacements are higher, in a total of 4 TMDs with equal properties. Since the contribution for the response due to the second mode for the VIRGIN train is higher than those due to the first mode, only a rough decrease was obtained in the computed results. Therefore, on the second case, a set of TMDs was tuned to the second vertical bending mode. This led to reductions from 35% to 52.7% in terms of accelerations and 31.7% to 44.7% in terms of displacements, respectively for mass ratios from 0.01 to 0.05, and for the passage of the VIRGIN train. A third configuration considers the two previous set of TMDs in order to reduce the contribution due to the two firsts vertical bending modes. Similar reductions to those obtained in the second case were obtained. It should be stressed that the reduction of the responses due to these passive systems occurs both for the steady-state and for the free-vibrations, the latter vanishing more quickly with TMDs. The effectiveness of the MTMDs system was also evaluated with the complete HSLM-A, leading to a much smoother envelope of the dynamic response in a speed band from 140 km/h to 420 km/h than that computed without this system. One of the main drawbacks of these vibration control systems is the detuning effect that may lead to a decrease in its effectiveness due to errors in the tuning of the TMDs or due to the fact that the passage of a train over a bridge is a time-variant problem. Face to the optimal tuned TMDs system with µ = 0.01, the displacement response with such a mistuned system increases about 14%, without, nevertheless, attaining the levels reached without TMDs.

178 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

DETAILS: A research project for improvement of analysis, design and durability of HS railway bridges

G. Chellini, L. Nardini & W. Salvatore Department of Structural Engineering, University of Pisa, Italy

Steel-concrete composite bridges are more and more used in the development of new European High-Speed (HS) Railway networks. Nevertheless design process highlighted many still open problems related to the estimation of dynamic and fatigue effects due to the HS train passages lacking reliable numerical models and suitable assessment methodologies. The Research Project “DETAILS: DEsign for opTimal life cycle costs (LCC) of high-speed railway bridges by enhanced monitoring systems”, funded by the Research Fund for Coal and Steel of the European Commission, aims at removing actual uncertainties on dynamic effects and inter- action phenomena, fatigue loadings, structural modelling, fatigue life and damage assessment of steel concrete composite bridges for HS railway lines. The research programme integrates suitable modelling strategies with experimental “in field” vibration test and laboratory fatigue tests on details in order to achieve optimal structural models and to improve design procedures. The research strategy will be applied to three case studies (a twin parallel girder bridge in Austria, a box girder viaduct in Italy and a composite filler beam bridge in Germany), representative of some most adopted structural solutions for HS steel-concrete composite railway bridges (Figure 1). In the present paper, the complete Research Work Programme is described, pointing out also the new aspects and methodologies adopted applying the research strategy to three case studies.

Figure 1. The three case studies of DETAILS research project: a) the twin parallel girder bridge, b) the box girder viaduct and c) the filler beam bridge.

179 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Dynamic testing and numerical modelling of a typical short span high-speed railway bridge

V.Zabel & M. Brehm Bauhaus-University Weimar, Weimar, Germany

As in many European countries also the German railway company – Deutsche Bahn (DB) – is enhancing and extending its high-speed railway network. This project does also include the improve- ment of existing links with the aim of raising the speed limits on the respective lines. In this context also all bridges have to be examined if the critical speed corresponding to respective trains is higher than the design speed limit for the complete line. This paper reports about the investigations which were carried out at a typical short span railway bridge in Germany. First the dynamic tests that were performed under very different environmental conditions are described. One of the most interesting results of the experimental investigations was that the corresponding natural frequencies were in winter significantly higher than in summer. Based on the identified modal parameters a numerical model was developed in several stages with increasing complexity. The aim was to create a model which describes the experimentally identified modal behaviour very closely. It became apparent, that a correct numerical description of the dynamic behaviour of the considered, relatively simply appearing structural system requires a relatively detailed model. The most relevant model parameters were identified by means of a stochastic sensitivity anal- ysis. A genetic optimisation algorithm was then applied to identify appropriate values for these parameters such that the model describes the dynamic behaviour of the real system with sufficient accuracy. This was the most appropriate approach which could be found to solve the identification problem described in this study. The investigations also showed that the dynamic behaviour of a short span railway bridge with tracks on ballast is significantly influenced by the properties of the bearings and the ballast, which includes the connection of the rails to the superstructure by the ballast. Neglecting these contribu- tions to the stiffness of the structure will normally not lead to a correct description of a structure of this kind.

Figure 1. Elevation of the bridge (left) and bottom view of the bridge deck (right).

180 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Development of an efficient finite element model for the dynamic analysis of the train-bridge interaction

S. Neves, A. Azevedo & R. Calçada Faculty of Engineering, University of Porto, Porto, Portugal

The design of high-speed railway bridges comprises a set of demands, from safety and serviceability aspects, to new types of equipment and construction solutions. The dynamic behavior of railway bridges can be analyzed with or without the consideration of the vehicle’s own structure. The simulation of the train-bridge system requires several independent submeshes and the consideration of contact conditions that represent their interaction. In order to perform an accurate and realistic evaluation of the dynamic behavior of the train-bridge system, adequate analysis tools that take into account the complexity of the system are required. These computational tools must be based on efficient algorithms to allow for the completion of detailed dynamic analyses in a reasonable amount of time. The classical methods of analysis may be unsatisfactory in the evaluation of the dynamic effects of the train-bridge system and fully assessment of the structural safety, track safety and passenger comfort. A direct and versatile technique for the simulation of the train-bridge interaction was imple- mented in the FEMIX code, which is a general purpose finite element computer program. The formulation of the contact between nodal points of the vehicle and internal points of a finite ele- ment is briefly described in this paper. The dynamic equilibrium equations are solved by a direct integration strategy, based on the Hilber-Hughes-Taylor method. Dynamic equilibrium equations in non prescribed degrees of freedom, in contact degrees of freedom and in prescribed degrees of freedom are separately developed. Contact compatibility equations between points of the vehicle and internal points of a finite element are also separately developed. All these equations consti- tute a single system of linear equations involving displacements, contact forces and reactions as unknowns. After the solution of this system of linear equations the displacements, velocities and accelerations at the current time step can be calculated and a new time step is started. This het- erogeneous system of linear equations can be efficiently solved by means of the consideration of several submatrices with specific characteristics. The new formulation was applied to the analysis of the dynamic behavior of the São Lourenço bridge, which is a bowstring arch bridge. The bridge is located in the North Line of the Portuguese railway system, in a section that was recently upgraded to allow the passage at greater speeds of the Alfa pendular train. The obtained results provide precise indications on the structural safety, track safety and passenger comfort. The track safety is guaranteed for speeds up to 310 km/h and the comfort level at all the train coaches is “very good” for speeds up to 345 km/h. The vertical acceleration at the last car body is much higher than for the case of the first car body, for a speed of 395 km/h (resonance of the first vertical mode of the bridge). The new formulation proved to be very accurate and efficient.

181 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Fatigue assessment of composite bridges for high speed railway traffic

H. Figueiredo, R. Calçada & R. Delgado Faculty of Engineering of the University of Porto, Porto, Portugal

Over the years, one aspect that has been regarded as one of the main disadvantages of using steel in railway bridges is its fatigue behavior, a phenomenon highly dependant on the structural detailing and on the dynamic effects usually involved. Fatigue is particularly important in the design of High-Speed Railway (HSR) infrastructures, which must withstand the passage of very long trains capable of travelling up to speeds of 350 km/h. In these conditions the passage of the train’s regularly spaced axle loads can lead to resonance phenomena and, consequently, to a decrease of the fatigue life. Recent advances on the dynamic behavior of HSR bridges have been introduced in the new European standards, EN1991-2 (2003) and EN1990-Annex A2 (2005). These include general guidelines for performing dynamic analysis, as well as design rules for checking structural and traffic safety, and passenger comfort. Despite these improvements, the design codes are still rather incomplete concerning the fatigue assessment of this type of bridges. The design philosophy used in the Eurocodes and in all major the fatigue design codes or standards is based on the concepts of damage accumulation and fatigue strength curves, also known as S-N or Wöhler curves. Two design methods, the damage accumulation method and the equivalent damage method, are proposed. The methods are based on 12 train types, suited for bridges that carry mainly standard traffic, heavy freight traffic or light traffic. No specific guidelines are given for bridges conceived for high speed passenger traffic. In this paper a complete assessment of the fatigue behavior of a composite HSR bridge is performed. The results are shown for the two methods proposed in the European standards and for a methodology which enables a correct and consistent assessment of the fatigue behavior of the bridge under HSR traffic. The methodology is based on the damage accumulation principle, using the stress time histories taken from a dynamic analysis of the bridge. The damage included both single train passages and train crossings, the later determined by doubling the stress history of a single passage. Analyses were conducted with the traffic scenario proposed for the TGV Eastern HS line which consists in 70 daily train passages, in each direction, for a period of 100 years. It is also assumed that in 5% of these cases the bridge is loaded simultaneously by two trains. Dynamic calculations were conducted for the European HS Trains for speeds varying between 145 km/h and 420 km/h. Fatigue calculations were conducted for each train at the design speed of 350 km/h and in resonance conditions, i.e., at the speed where the dynamic amplifications were greater. Finally, some conclusions concerning the fatigue behavior of this type of deck are pointed out.

182 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Experimental modal analysis of a twin composite filler beam railway bridge for high-speed trains with continuous ballast

T. Rauert & B. Hoffmeister Institute of Steel Construction, RWTH Aachen, Germany R. Cantieni rci dynamics, structural dynamics consultants, Duebendorf, Switzerland M. Brehm & V.Zabel Bauhaus University, Weimar, Germany

In the framework of the research project DETAILS(Design for optimal life cycle costs of high-speed railway bridges by enhanced monitoring systems, contract-No.RFSR-CT-2006-00032), supported by the European Research Fund for Coal and Steel, dynamic tests are performed on a railway bridge which is part of the high-speed line betweenAachen and Cologne in Germany.The monitored bridge consists of two parallel composite filler beam girders separated by a gap, each with one ballasted railway line. Both parts of the bridge are single span systems with a span of 24.6 m. The bridge is used by normal- and high-speed passenger trains but also by freight trains. As a first step ambient vibration tests have been carried out in April 2007. In a set of four configurations accelerations at 44 points at the bottom of the bridge have been measured due to ambient loads but also due to train passages. By using output only methods (EFDD- and SSI- algorithms) 13 mode shapes in the frequency range f = 3.68 ...41.1 Hz have been identified. The critical damping ratio for the first eigen frequency ranges between 2.5 and 3.0%. The identified mode shapes have been successfully used to calibrate a FEM-model of the bridge. The next step will be the development and installation of a permanent monitoring system on the bridge, which will provide system information for a time period of two years. By means of the monitoring campaigns the dynamic behaviour of the bridge and the changes in the dynamic behaviour due to long term effects are to be investigated. Aspects like concrete behaviour, slip between concrete and steel, contribution of tracks and ballast, bearing characteristics and climate influence are focused on. The main goal is to improve the current design concept, so that a better prediction of the dynamic behaviour of filler beam bridges will be possible.

Figure 1. Erfttal Bridge, cross section. Dimension in meters. HE-M steel profiles embedded in concrete.

183 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

A study of the lateral dynamic behaviour of high speed railway viaducts and its effect on vehicle ride comfort and stability

R. Dias, J.M. Goicolea & F. Gabaldón Computational Mechanics Group, Escuela de Ing. de Caminos, Univ. Politécnica de Madrid, Spain M. Cuadrado, J. Nasarre & P. Gonzalez Fundación de Caminos de Hierro, Madrid, Spain

The study of the lateral behaviour of railway bridges and vehicles is an important issue on bridges with low lateral stiffness, which has been defined by ERRI (1996) as those with lateral natural frequencies below 1.2 Hz. This limit applies to the deformation of the deck in one span, and was demonstrated to be a real issue on measurements and models of bridges with open deck sections and supporting trusses, of low lateral bending stiffness for the deck. Although not included in the above category, modern long viaducts for HSR with continuous decks on tall piers may also exhibit very low lateral stiffness and frequencies, which could produce undesired effects for the comfort or even the stability of the railway vehicles. In this work a simple model has been developed and applied to consider worst-case scenarios in a representative bridge, the “Arroyo de las Piedras” viaduct in Spain. The trains considered are representative of those circulating in the Spanish HSR network, as well as a freight wagon. Three-dimensional dynamic models were developed with finite elements. The actions considered include the lateral deformation of the bridge in response to vertical eccentric loads, track alignment irregularities and finally lateral motion of vehicles due to conicity of wheel-rail contact. The results show that there is, at least in this case, no cause for concern. However, for some scenarios the results in terms of lateral motion and forces are not negligible and should be considered in the design. Representative results are shown in Figure 1.

Figure 1. Viaduct “Arroyo las Piedras” for the HSR line Córdoba-Málaga, Spain (2007). Continuous deck of length 1209 m, spans of 63,5 m and piers taller than 94 m. In bottom left virtual path of deformations of viaduct due to eccentric vertical loads, in blue. At bottom right response of vehicle in terms of lateral accelerations, which are below 0.5 m/s2.

184 Damage assessment

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Evaluating composite steel girder-concrete slab bridge beams using simplified plastic analysis

P.S. McCarten Opus International Consultants, Napier, New Zealand

ABSTRACT: Ideal asset management requires Bridge Managers to have a high level of under- standing of the structural adequacy of the bridges under their control and be able to report the information. The Bridge Manager is also required to track the structural adequacy reflecting the changes in condition and/or structural risk through the life of the structure. Over the years the 460 steel bridges on the state highways in New Zealand, many of which rely on partial composite action, have had the load carrying capacity evaluated by a variety of methods. Steel beam design in New Zealand is currently to NZS 3404 and this standard is also used for evaluating existing bridge beams. Using NZS3404 to assess the flexural capacity of the many bridges with a degree of shear connectivity less than 50% fail to accommodate legal vehicles loads because the evaluator is required to assume non-composite action. To overcome this discontinuity and ensure consistency in the method for evaluating partially composite bridge beams, with a focus on determining the reliable shear connector strength, a simplified plastic analysis method was developed. The paper presents the rationale used for developing the assessment method, the details of the proposed method and a summary of the calibration that was undertaken to confirm the method. The proposed method allows application for the full range of beam configuration, covers the many neutral axis locations and the full range of shear connectivity. The method follows the philosophy of the design method set out in NZS 3404 but includes specific guidance for use with bridge beams which have specific loadings, flexure/shear stress distributions, deterioration due to fatigue effects and material deterioration due to durability effects. The method focuses on overload rating using the HO vehicle, the material properties and the ultimate limit state load factors defined in the Transit New Zealand Bridge Manual. A feature of the method is the inclusion of capacity reduction factors to allow for shear connector details, spacing or configuration not complying with the standard details in NZS 3404. Calibration of the method was undertaken in broad terms, accounting for shear connector frac- ture, covering the full range of partial composite action, by comparison to the working stress design method and by expert bridge engineer review. The calibration confirmed the appropriateness of the proposed method and it is concluded that it provides a reliable assessment of the beam flexural capacity. The method provides an approximate 20 to 30% increase in flexural capacity for the critical bridges allowing legal vehicle loads to be carried and of course deferring bridge replace- ment. Due to the simplifying assumptions for the applied rating load the method includes a small degree of conservatism that can be used for a specific overload vehicle assessment where axle configurations and loadings are known with some certainty. While the parametric study by Burnet and Oehlers (2001) offers a holistic view of the critical composite beam actions, appropriate for calibrating the method for overloads, it is suggested more research be undertaken to confirm the zones of applicability for shear connector fracture modes of failure and the reliability of the shear connectivity under the heavy and repetitive loadings experienced by bridge beams.

187 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Improving bridge component deterioration forecasting precision

Harold S. Kleywegt Keystone Bridge Management Corp.

ABSTRACT: One of the biggest knowledge gaps in bridge asset management is reliable informa- tion on the deterioration rate of bridge components. The overwhelming need for bridge rehabilitation in North America demands good field information and reliable forecasting data to determine appropriate bridge management strategies. Current North American bridge inspection practices may actually hinder our appreciation of bridge deterioration. The National Bridge Inventory Sys- tem (NBIS) is examined together with PONTIS® and the Ontario Bridge Management System (OBMS). Subjective rating systems are demonstrated to result in significant variability in outcomes. This variability confounds efforts to meaningfully model bridge deterioration. A novel approach to improving this information shortfall consists of characterizing a bridge and its components in terms of Depreciation, Defects, and Damage. Depreciation is a simple deterministic measure based on the age and deemed life expectancy of the component. Examples of the depreciation of a bridge with multiple components, and a network of 751 bridge decks are provided. The most compelling reason for modeling a bridge in terms of Defects and Damage is the simple elegance of only two easily understood definitions that can be applied to every component of a bridge. Defects are undesirable but non-threatening component attributes not caused by normal depreciation and are field identified and quantified during inspection. Similarly, Damage is any transformation of a component to a structurally weaker condition. Simple definitions and eliminating the task of rating a bridge’s“goodness” will facilitate concentrating bridge inspection efforts on exceptions that truly merit attention. The regular and timely recording of Defects and Damage on a component basis will yield a less subjective repository of data that will more precisely reflect the growth of bridge deterioration. As a result, it is proposed that a more consistent bridge condition metric will evolve.

188 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Stochastic subspace-based structural identification and damage detection – Application to a long span cable-stayed bridge

W.Zhou&H.Li Harbin Institute of Technology, Harbin, China

ABSTRACT: Automatic global vibration-based monitoring techniques have proven useful for SHM of structural and mechanical system. SHM techniques based on processing vibration mea- surements basically handle two types of characteristics: the structural parameters (mass, stiffness and damping) and the modal parameters (modal frequencies, associated damping values and mode- shapes). A central question for monitoring is to compute changes in those characteristics, to assess their significance and to handle correlations among individual changes, while controlling the complexity of the processing of the collected data. In this spirit, a damage detection method based on a residual associated with output-only subspace-based modal identification and global or focused chiˆ2-tests built on that residual has been proposed by Basseville et al. and successfully experimented on a variety of test cases. In the proposed damage detection algorithm, it is assumed that a signature of the structure in its safe state is available. This signature is usually identified using reference data, possibly recorded under an unknown non-stationary excitation. The damage detection algorithm processes new data by first generating a residual, and computing its sensitivity with respect to damages. The residual is shown to be asymptotically Gaussian under both on damage and small damage assumptions. This results in a global test, which performs a sensitivity analysis of the residual to the damage (modal changes), relative to uncertainties in the estimated modal parameters and noises on the available data. The purpose of this work is to describe the damage detection method and apply this method to assess a long span cable-stayed bridge. In the study, the undamaged data will be the reference output to obtain the test value (damage index) for the other damaged case. The results of identification and damage detection will be presented. These final results indicate that the algorithm is able to assess the damage in the example bridge.

189 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Structural safety of historical stone arch bridges in Korea

N.K. Hong & H.-M. Koh Korea Bridge Design & Engineering Research Center, Seoul, Korea S.G. Hong Department of Architecture, Seoul National University, Seoul, Korea B.S. Bae National Research Institute of Cultural Heritage, Daejeon, Korea

The majority of surviving history bridges in Korea belongs to stone arch bridges. However, historic stone bridges have suffered not only structural degradation but also historic or cultural degradation due to several reasons including urban development without a comprehensive consideration for cultural heritage; restoration of cultural heritage without systematic maintenance management plan; and rehabilitation considering only a few aspects such as structural safety impending at hand. This paper studies the structural behavior of stone arch bridge systems in Korea whose geometric shapes are varied from semi-circular to segmental with some variations focusing on structural safety aspect. Traditional practice used in the construction of stone arch bridges in Korea is first overviewed to understand what types of stone arch bridges are distributed. Previous studies on safety assessment for masonry arch bridges are then reviewed including the concepts of limit analysis and load’s thrust lines. Using the limit analysis that are performed in conjunction with collapse mechanisms, stone arch bridges in Korea are investigated under several simulated cases how they behave structurally regarding the structural safety issue. The results of the limit analyses are represented using a load factor, which refers to a multiplier of an applied load at which an arch structure under consideration has reached to a collapse state, and correlated geometrical and material parameters for practical purposes. Most old heritage stone arches in Korea seemed in semi-circular shape at first appearance. However, many of them are deviated somewhat from the original semi-circular shape which might be resulted from variations due to local site conditions. One of the variations was made by adding piers below the semi-circular arch ring. Accordingly their structural behavior is different from that of semi-circular arch ones. For example, the lowest load factor of Goheung bridge is much lower than that of typical semi-circular arch under the same loading condition (Fig. 1). There are many influencing factors affecting structural safety of old stone arch bridges including the followings: geometry, loading conditions, material properties, degree of material deterioration, and their interrelated conditions. This study provides nothing but a starting point.

Figure 1. Goheung bridge.

190 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Realistic estimation method of moment redistribution in reinforced concrete beams based on the analytical methods

Ju Hyoun Cheon & Jae Geoun Park Department of Civil and Environmental Engineering, Sungkyunkwan University, Kyonggi-do, Korea Sang Cheol Lee Korea Infrastructure Safety & Technology Co., Koyang, Kyonggi-do, Korea Myoung Seock Oh Department of Structure, Seoyeong Engineering Co., Ltd, Seoul, Korea Hyun-Mock Shin Department of Civil and Environmental Engineering, Sungkyunkwan University, Kyonggi-do, Korea

Moment redistribution in beams has traditionally been considered as an Ultimate Limit State (ULS) phenomenon closely associated with considerations of reinforced ductility and it reveals the realistic strength of reinforced concrete beams. The evaluation of the ductility of reinforced concrete beams is very important, since it is essential to avoid a fragile collapse of the structure by ensuring adequate deformation at the ultimate limit state. One of the procedures used to quantify ductility is based on deformations, namely, the plastic rotation capacity. From the many researches, the coefficient of redistribution is depends on the relation between the stiffness and the plastic rotation capacity at the critical regions. Significant differences exist among various design codes on the previous for moment redistribution bases on the approximate ductility of structures and provide empirical rules for moment redistribution. A nonlinear finite element analysis program named RCAHEST (Reinforced Concrete Analysis in Higher Evaluation System Technology) was used to evaluate the ultimate strength and degree of moment redistribution.

Keywords: Moment Redistribution; Performance Based Design; Finite element method; Plastic analysis.

191 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Probabilistic analysis of the structural behaviour of a bridge prestressed concrete beam

C. Cremona, S. Mohammadkhani-Shali & B. Richard Laboratoire Central des Ponts et Chaussées, Paris, France C. Marcotte & B. Tonnoir Centre d’Etudes Techniques de l’Equipement Nord-Picardie, Lille, France

TheVIPP (Viaducs à travées Indépendantes à Poutres préfabriquées en béton Précontraint par post- tension) are simple span viaducts made of precast concrete girders prestressed by post-tension. This type of construction was largely developed between the years 1955 to 1970, thanks to its girders launching system which allowed the crossing of non classical obstacles with reasonable height (10 to 25 m above ground-level, range between 30 to 60 m). This technique is not longer used, largely competed by other construction techniques making it possible to carry out more economic and safer redundant structures with respect to rupture. The enthusiasm which reigned at the construction time of the first VIPP generation moreover resulted in many design and execution mistakes, and by the absence of corrective maintenance actions to restore the defective waterproofing of these bridges. The detailed description of the problems encountered in this family of bridges shows that the diagnosis of these structures is essential to conclude a relevant assessment of their performance. It is in this context that the network of the Public Works Laboratories of the Ministry of Transport decided to initiate in 2005 a set of experimental and numerical investigations. The Merlebach bridge suffered from all the classical VIPP’s deficiencies and its replacement offered the opportunity to conserve one the most degraded beams in order to test its strength. The test specimen (figure 1a). is 33.50 m long and 2.10 m high (32.50 m between bearings). It is composed of ten 12Ø8 cables. The loading system is made by a framework place at mid-span on which a 150 t jack has been installed (figure 1b). The loading is performed by 5 t steps with zero returns after each step. The first loading cases are designed to determine the transition between a pure elastic behaviour and the beginning of the cracking process. A reliability analysis for complex structures requires in general a finite element reliability approach. Each realization of the random vector of basic variables is considered as a set of data which has to be sent to the finite elements solver. A response is obtained and then the construction

Figure 1a. Test specimen. Figure 1b. Framework and loading jack.

192 of a local approximation of the implicit limit state can be performed. In this paper, it is proposed to show the capabilities of the Quadratic Response Surface Methods (R.S.M.) and Support Vector Machines (S.V.M.)approaches for analyzing the structural reliability of the Merlebach’sbeam. This probabilistic numerical analysis has been performed to obtain a relevant structural model. Median and mean models are therefore determined as well as 90% confidence intervals. These numerical investigations are compared with the 90% measurement confidence intervals in order to assess the degree of uncertainties introduced in the probabilistic numerical model.

193 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Gi-Lu Cable Stayed Bridge – From earthquake damage to full recovery

Z.K. Lee National Center for Research on Earthquake Engineering, Taipei, Taiwan K.C. Chang Department of Civil Engineering, National Taiwan University, Taipei, Taiwan C.C. Chen Yunlin University of Science and Technology, Yunlin, Taiwan C.C. Chou National Taiwan University, Taipei, Taiwan

The Gi-Lu Bridge is a modern pre-stressed concrete cable-stayed bridge located in central Taiwan. The bridge crosses the longest river in Taiwan. With two 120-meter spans and two row-fanned 68 cables, the bridge is symmetrically supported by a single pylon and two bents. On Sept. 21, 1999, when construction of Gi-Lu Bridge was nearing completion, a significant earthquake of 7.3 ML magnitude (the Gi-Gi Earthquake) struck central Taiwan. Only three kilometers away from the earthquake epicenter, the bridge was subjected to strong ground shaking and some structural members failed. After the earthquake, repair work was conducted in two phases. Phase (1): retrofit mending of the concrete structure. Phase (2): replacement and adjustment of the cable system. This study introduces the damage and repair procedures; as well as analyzes the cable loading test and safety loading test of Gi-Lu Bridge.

194 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Stress monitoring of steel girder bridges with different boundary conditions

N. Namatame Nichizo Tech Inc., Osaka, Japan K. Tani Takadakiko Co., LTD., Osaka, Japan T. Tsuji Hitachi Zosen Steel Structures Corp., Osaka, Japan M. Kawatani Department of Civil Eng., Kobe University, Kobe, Japan M. Kano JIP Techno Science Corp., Osaka, Japan N. Tanaka KATAYAMA Stratec Corp., Osaka, Japan

H. Hattori Department of Informatics, Kansai University, Osaka, Japan

ABSTRACT: This study presents stress measurements and analytical studies of two bridges (Figure1), both of which have almost the same type at the same place, but the newly constructed bridge is in good condition, the older one is in poor condition. It was confirmed that A2 abutment moved toward the center of the river by ground movement. The poor condition bridge, therefore,

Figure 1. Two target bridges with location of strain gauges and thermo couples, General plan.

195 Figure 2. Stress Measurement; Longitudinal Stress variation due to seasonal temperature change. is constrained in the longitudinal direction and as a result the bearing lost its controllability under seasonal temperature changes. Figure 2 shows the stress measurement due to the temperature variations. In a large sense, the stress range of the poor condition bridge is larger than that of the good condition bridge. The analytical studies have been conducted to complement the measurement. The analyses are performed in three cases with different boundary conditions. Comparing by the results of both bridges, their present condition is grasped.

196 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Evaluation of compressive stiffness of elastomeric bearings

H.J. Yoon,Y.J. Kim, C.B. Cho & I.J. Kwahk Korea Institute of Construction Technology, Goyang, Korea

For the elastomeric bearings to fulfill their functions as device transmitting superstructure‘s load, appropriate stiffness should be developed in the vertical direction. Accordingly, USA, Europe and Japan are proposing specifications for the elastomeric bearings based on analytical and experimental results. In Korea, the Korean industrial standard KS F 4420 is dealing with overall features of elastomeric bearings. However, KS F 4420 is not establishing clearly the compressive elastic modulus and is proposing erroneous load range for the performance evaluation, which impedes the verification and comparison of the performance of the bearings. Therefore, the bearings designed in compliance with KS F 4420 are likely to produce measured values different to design or may not develop the performance required in design. On the other hand Japan proposed empirical formula based on actual stiffness of elastomeric bearings to reduce the gap between actual stiffness and the one proposed specification. This study performed the compressive stiffness test on domestic elastomeric bearings and com- pared its measured stiffness with the stiffness in the KS F 4420 and empirical formula in the Japanese bearing manual. Table 1 showed that the KS standards are expressing adequately the

Table 1. Comparison of measured and design compressive stiffness.

Actual Actual compr. Deviation compr. Deviation Shape stiffness Shape stiffness ∗ ∗ factor Kv KSF JBM factor Kv KSF JBM Specimen – (kN/mm) 4420 (2004) Specimen – (kN/mm) 4420 (2004)

S5#1 5.1 129.43 59% 9% RBC2-1 33.5 2084 −1% 125% S5#2 5.1 124.31 53% 5% RBC2-2 33.5 2057 −3% 122% S5#3 5.1 135.53 66% 14% RBC2-3 33.5 2043 −3% 120% S5#4 5.1 135.95 67% 15% RBC2-4 33.5 1955 −7% 111% S5#5 5.1 134.69 65% 14% RBC3-1 31.42 1616 −8% 109% S5#6 5.1 136.53 68% 15% RBC3-2 31.42 1712 −3% 122% S7#1 6.8 140.33 −21% −30% RBC3-3 31.42 1643 −7% 113% S7#2 6.8 137.92 −22% −32% RBC3-4 31.42 1914 9% 148% S7#3 6.8 141.55 −20% −30% RBC3-5 31.42 1651 −6% 114% S7#4 6.8 155.39 −12% −23% RBC3-6 31.42 1702 −3% 121% S7#5 6.8 164.28 −7% −18% RBC3-7 31.42 1711 −3% 122% S10#1 10.2 172.73 −65% −59% RBC3-8 31.42 1759 0% 128% S10#2 10.2 157.34 −68% −62% RBC3-9 31.42 1609 −9% 108% S10#3 10.2 167.43 −66% −60% RBC3-10 31.42 1629 −7% 111% S10#4 10.2 184.4 −63% −56% RBC3-11 31.42 1698 −4% 120% S10#5 10.2 179.02 −64% −57% RBC3-12 31.42 1664 −5% 116% S10#6 10.2 206.99 −58% −50% RBC3-13 31.42 1632 −7% 111% RBL#1 9.07 844.09 −19% −16% RBC3-14 31.42 1627 −8% 111% RBL#2 9.07 873.52 −17% −13% RBC3-15 31.42 1590 −10% 106% RBL#3 9.07 857.79 −18% −14% RBC3-16 31.42 1598 −9% 107% RBL#4 9.07 874.02 −16% −13%

∗ Japanese bearing manual (2004)

197 behavior of elastomeric bearings with large shape factors but fails to represent their behavior for shape factor ranging below 10 that is bearings commonly used for bridges. The empirical formula is expressing adequately the behavior of elastomeric bearing for bridges, while failed to represent their behavior for large shape factor. Accordingly, proposal of a regression equation for the com- pressive elastic modulus should be done by means of stiffness test fitted to the shape factor range adopted for bearings purposed for bridges.

198 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Effect of soil-bridge interaction and continuity on live load distribution in integral bridges

M. Dicleli & S. Erhan Middle East Technical University, Ankara, Turkey

In bridge design, most engineers prefer using simplified two dimensional (2-D) structural mod- els and LLDFs available in current bridge design codes to determine live load effects in bridge components. Although these LLDFs are developed for the girders of regular jointed bridges, they are also used for designing the girders of IABs since LLDFs specifically for IAB components are not available in current bridge design codes such as AASHTO (American Association of State Highway Transportation Officials) Bridge Design Specifications. However, the LLDFs developed for regular jointed bridge girders may not be suitable for IAB girders due to the effect of mono- lithic construction of IABs at the abutments forcing the bridge superstructure to interact with the abutments, piles as well as backfill and foundation soil. Furthermore in IAB design, the live load effects in the abutments and piles are generally calculated by using some rules of thumb or the LLDFs developed for the girders. This may obviously lead to either conservative or unconservative estimates of the live load effects in the substructure components of IABs. Moreover, in spite of the monolithic construction of IABs forcing the abutments to interact with the backfill under gravi- tational load effects, the current state of design practice in North America and Europe normally neglects backfill-abutment interaction effects in live load analyses of IABs. This may also result in conservative or unconservative estimates of the live load effects in the components of IABs. There- fore, a research study on the effect of soil-bridge interaction as well as the substructure properties on the distribution of live load effects in IAB components is needed to enlighten the above-mentioned uncertainties. For this purpose, numerous three dimensional (3-D) and corresponding two dimen- sional (2-D) structural models of typical IABs are built and analyzed under AASHTO live load. In the analyses, the effect of various geotechnical and substructure properties such as foundation soil stiffness, absence and presence of backfill, backfill compaction level, absence and presence of wingwalls, abutment height and thickness as well as size, orientation and number of piles are considered. The results from 2-D and 3-D analyses of IABs are then used to calculate the LLDFs for the components of IABs as a function of the above mentioned properties. It is found that the effect of the foundation soil stiffness, backfill, wingwalls, and the size and orientation of the piles, on the LLDFs for the girder and pile moments and shear is negligible. However, the foundation soil stiffness, the backfill and number of piles are found to have a remark- able effect on the LLDFs for the abutment moment and to a lesser extent on those for the abutment shear. Therefore, these parameters must be considered in estimating the distribution of the live load effects within the abutments of IABs. Furthermore, the analyses results revealed that the AASHTO girder LLDFs can be used to obtain reasonable estimates of live load effects in short to medium span IAB girders where the calculated live load shear for IAB girders using AASHTO LLDFs may be increased by 5% to assure safety. However, LLDFs for IAB abutments and piles are still needed to estimate live load effects in these components for design purposes. The findings from this research study can be used as a starting point to formulate the LLDFs for the abutment and piles of IABs.

199 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Application of the analytic hierarchy process in performance evaluation of existing concrete cable-stayed bridge

Qiao Li, Deshan Shan & WenYan Civil Engineering School, Southwest Jiaotong University, Chengdu, Sichuan, China

There are three main stages in the lifespan of major bridges, including macro-planning, designing and construction, and operational management. In the past, the previous two stages were received more attention, while the importance of operational managing is ignored. With the construction and development of traffic network in the whole China, a great amount of capital is invested in major bridges. Owing to the gradual great influence on bridge condition by scientific management, operation management (including bridge health monitor and assessment) receives more and more attention to fully bring its function into play and guarantee its normal operation. During the life- cycle of the concrete cable-stayed bridges, it is inevitable to have some degradation in different degrees because of the inherent and extrinsic factors, such as the increase in traffic volume, the increase in heavy vehicles and structural aging. There are many concrete cable-stayed bridges in China, and the deterioration of them is increasing in service, some of them are even worse. Therefore, the requirements for health monitor and condition assessment of existing cable-stayed bridge increases. The longer time the bridge serves, the worse its deterioration is in general. So it is indispensable to inspect the existing cable-stayed bridges for determining their condition. Only after the visual inspection and monitor of the existing cable-stayed bridge, the proper assessment can be made and the real condition about the bridge can be obtained. Based on this assessment and real condition about the existing concrete cable-stayed bridge, a set of realistic and objective maintenance decision system can be established for treating the bridge damages and defects. And this makes the decision-making of bridge technical rehabilitation more scientific and technical plan for bridge renovation more economical and rational with great social significance, economical value and widespread application prospect. The current conditional assessment of concrete cable-stayed bridges, however, is on the level of simple evaluation and qualitative analysis for each component of the bridges, and there is no appropriate integrative assessment system for the inspection of them. In this paper, an integral assessment model and the synthetic grading level for existing concrete cable-stayed bridge is figured out. These assessment indexes are classified as qualitative indexes, quantitative indexes and measurable group indexes, and the Analytic Hierarchy Process method is adopted to evaluate the qualitative indexes. Based on the systematic analysis of existing concrete cable-stayed bridge defects, a hierarchy assessment model is established. Based on the experts’ investigation, the judgment matrix of multiple comparisons is carried out, and the weight of single element is figured out and judgment consistency is checked, then group decision method is used to calculate the judgment results of all experts. The corresponding assessment software, which as the assessment tool of the Bridge Inspection Rating Maintenance Management System (BirMMS), is complied based on the .net framework 2.0 aided by the powerful development tool of visual studio 2005 C#. One certain concrete cable-stayed bridge with single pylon is inspected and assessed by the proposal method in this paper, and the presented assessment method is proved feasible and reasonable through application to one certain concrete cable-stayed bridge. The result suggests that the proposal method can be applied in the real bridge performance assessment.

200 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Analysis of box girder bridges using finite elements and AASHTO-LRFD

R.R. Doerrer Strand Associates, Inc., Joliet, Illinois, USA R.A. Hindi Bradley University, Peoria, Illinois, USA

ABSTRACT: In the last twenty years, the bridge design philosophy has had major changes. The basic principle of resistance must be greater than the applied loads still remains unchanged. The bridge must be able to support all loads that will be applied to it. In order to design a safe and economical bridge correct modeling of the structure must be used. When a vehicle or load travels across the bridge the load of the vehicle is transferred to the bridge deck and then to the girders. The load is not taken entirely by a single girder, rather it is shared over the girders; this effect is know as live load distribution. Distribution of the load is not uniform over the girders; rather the girders that are closest to the load will carry most of it. Most bridges are commonly designed by taking the distribution factor and multiplying it to the beam line analysis results to determine the design forces for the bridge. In a beam line analysis the superstructure is idealized as a single beam and is subject to a single lane of traffic loading. Thus it is easy to see why a correct distribution factor is needed. If the factor is too small the girders will be over design and not be economical. If the factor is too large the girders will be under design and possible be unsafe. This paper presents a comparison between the live load distribution factors of cast in place concrete box girder bridges based on the AASHTO-LRFD 2007 specifications and finite element analysis. In this comparison, the range of applicability limits specified by the current AASHTO- LRFD specifications is fully covered and investigated in terms of span length, girder depth, girder spacing, and number of cells. In order to investigate the accuracy of the live load distribution factor for both positive and negative moment. Both simple and continuous bridges were investigated in this research. The use of three span bridges was used to compare the distribution factors for exterior and interior spans. To cover all the limits of the AASHTO-LRFD specifications, a variety of bridges were each modeled with different geometric properties that pushed the limits of what the specifications allow. There are two main categories of bridges in this study, which are based on the limits of theAASHTO- LRFD specifications that are being investigated. The first category, Category One, will include all bridge groups that have a varying span length and depth, but all other bridge geometric properties are held constant. The depth will vary as a function of span length to accurately represent bridges that are currently being constructed. This category has bridges that have one, two, and three spans. The second category, Category Two, will include all bridge groups that have varying number of cells in bridge, but keep a constant width and length. This category has bridges that have one and two spans. All the bridges were modeled using SAP 2000 Version 10.0.2. In order to gain confidence in the modeling process bridges previously modeled by others were modeled again using SAP 2000, and their results were compared to the results from the finite element analysis. The finite element modeling process produced results with in five percent of the existing modeled bridges. The live load used in the analysis is as specified by AASHTO-LRFD. Lanes were defined along the span of the bridge and moved transversely across the width of the bridge. This load was then moved longitudinally along the lanes defined on the bridge. All possibilities of one lane or multiple lanes loaded were investigated. 201 In order to make a comparison between the AASHTO-LRFD distribution factors and the Finite Element Analysis distribution factors, a ratio was made. The ration “R” was the AASHTO-LRFD distribution factor over the finite element distribution factor. If the ratio is greater than one, the AASHTO-LRFD values are being conservative and overestimating the distribution factor. If the ratio is less than one, AASHTO-LRFD is not being conservative and underestimating the live load distribution factor. By introducing this ratio “R”, it becomes much easier to see trends in the distribution factors and to see the accuracy of the AASHTO-LRFD values when compared to the finite element values. Based on the results of this study, the following findings were concluded. 1. The use of the lever rule is very inaccurate and can produce distribution factors that are more than three times the finite element results. 2. AASHTO- LRFD distribution factors for moment both interior and exterior girders subject to one or multiple lanes loaded become increasingly conservative as span length increases. Typically span lengths greater than 32 meters have AASHTO-LRFD at least ten percent greater than FEA. The increase of the number of cells can also cause AASHTO-LRFD to become more conservative. When the number of cells increases from four to seven the ratio “R” increase by up to 15 percent. 3. AASHTO-LRFD distribution factors for shear interior girders subjected to one lane loaded are ten to 30 percent higher than the FEA distribution factor and for exterior girders subjected to one lane loaded are 25 to 65 percent higher. For interior and exterior girders subjected to multiple lanes loaded AASHTO-LRFD is with in ten percent of FEA. The effect of a different number of cells is minimal. When the number of cells increases from four to seven the ratio “R” does not vary by more than ten percent. 4. For distribution factors on multiple span bridges the effects of multiple spans on distribution fac- tors are small. For moment distribution factors, FEA averages within three percent for multiple spans compared to single spans; for shear, FEA averages within ten percent. The AASHTO- LRFD distribution factors are the independent to the number of spans. Two and three span bridges’ distribution factors are incredibly similar. The FEA distribution factors, when compar- ing the distribution factors for exterior spans of three span and two span bridges, are within one percent of each other. The FEA distribution factors are approximately 20 percent higher for exterior spans than for interior spans.

202 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Nonlinear finite element analysis of precast segmental prestressed concrete bridge piers

Heon-Min Lee, Dae-Jeong Seong, Jae-Geun Park, Keo-Sung Kim & Hyun-Mock Shin Department of Civil and Environmental Engineering, Sungkyunkwan University, Kyonggi-do, Korea Tae-hoon Kim, Young-Jin Kim, Sung-Woon Kim Civil Engineering Research Team, Daewoo Institute of Construction Technology, Kyonggi-do, Korea

The main objective of this study is to investigate the inelastic behavior of precast segmental con- crete bridge piers that has many advantage as mentioned above. A computer program, named RCAHEST (Reinforced Concrete Analysis in Higher Evaluation System Technology), for the non- linear finite elements analysis of reinforced concrete structures was used. Material nonlinearity is deliberated by comprising tensile, compressive and shear models of cracked concrete and a model of reinforcing steel. The proposed system uses unbounded or bonded tendon to join the precast segments. An unbonded tendon element based on the finite element method, that can represent the interaction between tendon and concrete of prestressed concrete member, is used. An interface element is newly developed to predict the inelastic behaviors of interface between the segments in the precast segmental prestressed concrete bridge piers. The proposed numerical method for the inelastic behavior of precast segmental prestressed concrete bridge piers is verified by comparison with reliable experimental results.

Keywords: Precast Segmental PSC Bridge Pier; Bonded and Un-bonded tendon; Finite element method.

203 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Advanced numerical study of asphaltic surfacings on orthotropic steel deck bridges

X. Liu Section of Structural Mechanics, Delft University of Technology, Delft, The Netherlands T.O. Medani Road and Railroad Research Laboratory, Delft University of Technology, Delft, The Netherlands A. Scarpas Section of Structural Mechanics, Delft University of Technology, Delft, The Netherlands

Light weight orthotropic steel bridge decks have been widely utilized for bridges in seismic zones, movable bridges and long span bridges. Nowadays more than 1000 orthotropic steel bridges have been built in Europe, out of which 86 are in The Netherlands. In the last three decades, several problems were reported in relation to asphaltic surfacing materials on orthotropic steel deck bridges such as rutting, cracking, loss of bond between the surfacing material and the steel plate. The sever- ity of the problems is enhanced by the considerable increase in traffic in terms of number of trucks, heavier wheel loads, wide-base tires etc. Current design methods have a very limited success in esti- mating correctly the life span of the surfacing material. The main cause of the limited success of the current design methods originates from the fact that the mechanism of the interaction between the structural components is poorly understood. The current design methodologies are merely based on the assumption that behavior of all materials is linear elastic. The geometry of the orthotropic steel deck bridge is often simplified to that of a layered beam supported on two or three supports. According to many experimental and numerical investigations, when traffic load is imposed on the surface of orthotropic steel deck bridges, non-uniform displacement fields develop resulting to a variety of states of stress in the surfacing material. This necessitates search for a more powerful computational tool for increased understanding of complicated response of asphaltic surfacing on orthotropic steel deck bridges. The tool should incorporate a more realistic material behavior and more realistic representation of the structural geometry. In order to gain insight into the mechanical response of asphaltic surfacing materials and their interaction with orthotropic steel decks a large project of experimental and computational inves- tigation has been undertaken at Delft University of Technology. One of the primary goals of the investigation was the development of a constitutive model capable of describing the non-linear response of asphaltic surfacings. Once the model was available and verified, its implementation into the CAPA-3D finite element system has enabled the simulation of the dynamic non-linear response of orthotropic steel deck bridges accounting for the effects of material non-linearity, complex geometry of orthotropic steel deck and moving load patterns. In the first part of this contribution, details of the constitutive model development and formulation are presented. Model predictions are compared with actual test data from Mastic Asphalt (MA) surfacing material under various temperatures and displacement rates. In the second part, various aspects of the finite elements response of a 3D orthotropic steel deck bridge subjected to dual wheel loads are presented. The importance of the strain distribution through the depth of the surfacing asphalt and steel is demonstrated as well as the role of other influencing factors, such as non-linearity of surfacing material and membrane, loading location, loading speed and environmental temperature.

204 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Estimation of damping characteristics for cable using system identification scheme

S.-K. Park, K.W. Lyu & H.S. Lee Seoul National University, Seoul, Korea

Cable-supported structures are appropriate structural type for large-scale structures such as long- span suspension bridges and cable-stayed bridges. As cable-supported structures become larger and longer, dynamic problems caused by the rain-wind induced vibration and vortex vibration can be an important issue to design the bridges. This is because the inherent cable damping ratio is very low. Since the damping characteristics of the cable are different from other members, the accurate estimation of the cable damping ratio is important in designing damper and dynamic analyzing of cable-structures. This paper proposes a method to estimate the damping ratio of the stay-cable, based on a system identification scheme in time domain using measured acceleration data. The error function is defined as the time integral of the least square errors between the reconstructed displacement and calculated displacement by a mathematical model. Since the accelerations have instabilities with loading condition and are very sensitive compared with the displacements, the direct use of measured acceleration data to SI problem is difficult. However, the displacements of the cable are difficult to measure caused by support problem to fix the measuring device. The displacements are reconstructed using measured acceleration data through displacement reconstruction technique proposed by Hong et al. Displacement of the cable is calculated by linearization of equation of motion and discretization by finite element method. Linearized incremental equation of motion is integrated by Newmark-β method. Displacement is defined as the relative position between the current position and the position when the cable is loaded by its own self-weight. The structural damping is modeled by the Rayleigh damping. Damping coefficients of Rayleigh damping model are considered as system parameters. A regularization technique is used to alleviate the ill-posed characteristics of inverse problems. Determination of the regularization factor is very important in obtaining stable and physically meaningful solution. An optimal regularization factor is determined by geometric mean scheme. The validity of the proposed method are demonstrated through a laboratory test of the stay-cable. The damping ratio is estimated using displacement for forced vibration and free vibration, separately. The estimated damping ratio is compared with the damping ratio by logarithmic decrement method.

Keywords: Damping ratio, Cable, System identification in time domain, Rayleigh damping model, Displacement reconstruction.

205 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Self-adapting models of bridge degradation

J. Bien & A. Banakiewicz Wroclaw University of Technology, Wroclaw, Poland

Prediction of bridge degradation processes is crucial for bridge lifetime prognosis and for optimal planning of maintenance as well as for distribution of financial resources. In contrast to the majority of presently used pre-defined degradation models the proposed solution enables dynamic modifi- cation and individualization of the self-adapting models for each structure. Application of flexible “intelligent” models of degradation with ability to automatic self-adaptation based on systemati- cally collected data. This paper is focused on the deterministic self-adapting solutions created by means of two techniques: symbolic – based on the self-adapting degradation functions and non- symbolic – using the self-adapting neural networks. General conceptions of both techniques are described and discussed. In the first technique the initial degradation function – specific to the considered type of bridges – is defined on the basis of statistical data describing degradation of the analysed population of bridges. By means of the results of consecutive inspections of the considered structure the initial degradation function can be transformed to the individual degradation function specific to the analysed bridge. The transformation formula is proposed in the paper. Artificial neural networks are composed of interconnected neurons and the networks are able to learn by examples. For learning of the neural networks a set of various degradation models, defining the bridge condition against the structure age, was used. On the basis of history of the condition rating the neural network is able to define an individual degradation model specific for the considered structure. As a number of information about the structure condition is increasing the neural network can be automatically re-trained to obtain more precise prognosis of the bridge lifetime. In the presented study few degradation models based on neural networks were tested and comparison of the accuracy of lifetime prognosis by means of two considered neural degradation models is presented in Figure 1.

4.5

4

3.5

3 model A(25) model B(25)

2.5

2 A(25) model - testing B(25) model - testing 1.5 mean error [years] A(25) model - training 1 B(25) model - training

0.5

0 3 5 10 20 30 neuron numbers in hidden layer

Figure 1. Mean error of training and testing dependent on neuron number in hidden layer.

206 Proposed technologies of self-adapting degradation models can be used in expert tools supporting decisions concerning bridge safety and maintenance. The possibility of automatic creation of the dedicated dynamic model of degradation for individual structure seems to be the most promising innovation.

207 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Computational model generation based on 3D CAD digital data of RC bridges

Jeeho Lee & Min-Seok Kim Dongguk University, Seoul, Korea

In this paper an algorithmic method for generating computational RC bridge models from their 3D CAD digital data is presented. The present method has two procedure steps: structural object identification procedure and rebar generation procedure (Figs 1–2). The first procedure consists of sub-steps, rebar data acquisition step and object identification step. The second procedure, the

Figure 1. Procedure steps for computational RC bridge model generation.

Figure 2. Computational model generation procedures.

208 rebar generation procedure in the context of the finite element method and the multi-level analysis, includes node, element and connector generation steps. Numerical examples are given to show how the present method works in the multi-level analysis of RC bridges. The proposed method provides an efficient and practical model generation process especially for reinforced concrete structures due to its conveniency in constructing a multi-level analysis model along with 3D CAD digital data.

209 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Artificial Intelligence: Historical development and applications in civil engineering field

L. Sgambi Department of Structural and Geotechnical Engineering, University of Rome “La Sapienza”, Italy

The Artificial Intelligence field (AI) collects all studies regarding the understanding and construc- tion of intelligent entities. Mixing up within another disciplines, the study of intelligence and artificial intelligence are at the same time, an ancient and modern science [Russel 1995]. The birth of artificial intelligence has been fixed in 1956 when Mc Carthy, along with other researchers as Minsky, Shanno and Rochester, coined the definition in a historical convention at Dartmouth University. The soft-computing was introduced by Zadeh in 1994: exploit the tolerance for imprecision, uncertainty and partial truth to achieve tractability, robustness and low solution costs ...The principal components are fuzzy-logic, neurocomputing and probability reasoning. The reference model for soft-computing is the human mind. Many applications was developed in the civil engineering field, using neural networks ([Safi 2004] ,[Arangio 2005]), Evolutive algorithms ([Biondini 2000], [Sgambi et al. 2004]) Fuzzy-logic ([Savoia 2002, [Sgambi 2004], [Sgambi & Bontempi 2004], [Moller]). These applications shown that soft-computing methods are very effective approaches to improve the reliability of the analyses and design of complex structures.

REFERENCES

Arangio S. 2005. Heuristic methods for solving direct and inverse problems of complex structural systems, Proceedings of the Eighth International Conference on the Application of Artificial Intelligence to Civil, Structural and Environmental Engineering, 29 August – 2 September 2005, Roma. Biondini F., Bontempi F. & Malerba, P.G. 2000. Fuzzy Theory and Genetically-Driven Simulation in the Reliability Assessment of Concrete Structures. Proceedings of 8-th Conference on Probabilistic Mechanics and Structural Reliability, July 24–26, Notre Dame. Moller, http://www.uncertainty-in-engineering.net/ Russel S. & Norvig P. 1995. Artificial Intelligence. A modern approach. Prentice-Hall, Englewood Cliffs, NJ. Safi M. & Tehranizadeh M. 2004. Artificial intelligence for study of ground motion parameter effects on design spectra, European Earthquake Engineering, Vol. 2. Savoia, M. 2002. Structural reliability analysis through fuzzy number approach, with application to stability. Computer and Structures, Vol. 80, No. 12, pp. 1087–1102. Sgambi L., Bontempi F., Biondini F. & Frangopol D.M. 2004. Handling Uncertainties in Optimal Design of Suspension Bridges with Special Emphasis on Loads. Proceedings of PSAM 7 – ESREL’04, Berlin. Sgambi L. 2004. Fuzzy approach in the three-dimensional non linear analysis of reinforced concrete two-blade bridge piers. Computers & Structures, Vol. 82, pp. 1067–1076. Sgambi L. & Bontempi F. 2004. A fuzzy approach in the seismic analysis of long span suspension bridge. Proceedings of the 13th World Conference on Earthquake Engineering, Vancouver, B.C., Canada.

210 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Safety factor prediction for steel cable-stayed bridges by iterative eigenvalue analysis

Dong-Ho Choi, Hoon Yoo, Dong-Soo Kim & Ho-Sung Na Hanyang University, Seoul, Korea

The paper illustrates a critical load analysis of steel cable-stayed bridges considering axial-flexural interactions of girders and towers in the bridge. The analysis described in this study is based on the concept of the bifurcation stability using a tangent modulus approach and the column strength curve. Criteria for an iterative eigenvalue analysis are proposed in order to consider the primary bending moment as well as the axial force by using the interaction equation for beam-column members in cable-stayed bridges. Inelastic critical loads are evaluated for numerical models of example cable- stayed bridge that has a center span of 600 m. The results show that the illustrated method correctly evaluates the critical loads of steel cable-stayed bridges and can be a good alternative for predicting a safety factor of overall bridges.

Figure 1. Numerical model of the example bridges.

Table 1. Critical loads of the example cable-stayed bridges.

Elastic buckling analysis Inelastic buckling analysis Refined plastic hinge analysis

Model Depth LC1 LC2 LC3 LC1 LC2 LC3 LC1 LC2 LC3

600-m 1 3.62 3.63 3.82 2.60 2.61 2.77 2.37 2.38 2.67 2 8.43 8.28 8.38 2.87 2.89 3.24 2.95 2.96 3.15 3 13.37 12.93 13.07 3.09 3.01 3.02 3.12 3.13 3.19 4 17.83 17.40 17.38 2.87 2.97 2.97 3.11 3.13 3.04 5 22.20 21.89 21.93 2.84 2.92 2.93 3.08 3.14 3.15 6 23.92 23.80 23.88 2.84 2.86 2.86 3.05 3.07 3.10

211 Figure 2. Variation of critical loads with respect to the girder depth.

212 Damage assessment of existing bridges

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Evaluation of load-carrying capacity of the damaged bridge model using the updated FE model

D.S. Jung & C.Y. Kim Myongji University, Yongin-si, Gyeonggi-do, Korea

Load-carrying capacity is usually used to describe the structural performance considering phys- ical conditions of structures. This procedure consists of field test, Finite Element (FE) analysis, and load rating. Generally, load-carrying capacities of most bridges under service for a long time are decreased by site-specific environments such as aging of materials, increasing heavy vehi- cles, and structural damages. To decide maintenance policy of old bridges, rigorous evaluation of load-carrying capacity in current state is very important. In actual applications, assessment results often show extreme differences due to modeling error of FE model and measurement error. If mea- surement errors are reduced by applying high-accuracy sensors, reliable data acquisition systems, and well-developed signal processing tools, the difference can be generally minimized through FE model updating procedure. Especially, the accurate analytical model is more important in case of old and damaged bridges. Therefore, the modification of analytical model considering the current state is necessary in order to accurately evaluate load-carrying capacity of the damaged bridge. In this paper, FE model updating using measured data and assessment of load-carrying capacity by the updated FE model are investigated. First, the laboratory test by several damage scenarios for the simply supported small-scale bridge model was conducted. Through the laboratory test, static deflections, natural frequencies up to the first six modes and the first six corresponding mode shapes were considered for FE model updating. Finally, finite element model updating using the measured data and load-carrying capacity assessment according to the damage severities were carried out. The Hybrid Genetic Algorithm (HGA) method based on the genetic algorithm and the modified Nelder-Mead simplex method are used to update the initial FE model. The objective function suggested for the HGA is formulated as a linear combination of the fitness functions related to the natural frequencies, mode shapes and static deflections. The MathWorks’ MatLAB toolbox is used to optimize the objective function for the HGA, and the ABAQUS is used for FE analysis. The FE model was modeled as a grillage using only beam elements. Also, The Allowable Stress Rating (ASR) was utilized to evaluate load-carrying capacity for each damage case. Through the model updating on simply supported small-scale laboratory bridge model, it is verified that the FE model of the damaged structure can be correctly improved if the upper limit value of updating variables is controlled by design calculation. Also, it is shown that the load- carrying capacity using the updated FE model were well estimated. Consequently, it is found that the use of the updated FE model to evaluate the load-carrying capacity is more effective than that of initial FE model in the case of damaged structure. Also, it is expected that this technique can be widely applied to reinforced concrete girder bridges as well as steel plate girder bridges.

215 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Detection of sudden damages of structure by regularized autoregressive model using measured acceleration

J. Kang & H.S. Lee Department of Civil and Environmental Engineering, Seoul National University, Seoul, Korea

This paper presents a novel damage detection algorithm by regularized autoregressive model using measured acceleration. The autoregressive (AR) model is utilized to make non structural model based system. The autoregressive model is extremely useful in the representation of certain practically occurring time series. The autoregressive model with time windowing technique is employed to overcome the perturbations of measured signals. The perturbations due to environmental changes are commonly changed gradually during long time period and time window size is relatively very smaller than environmental perturbation period so it can be assumed that perturbation of environmental changes can be neglected within the time window. The covariance between residuals and first order coef- ficients of autoregressive model is proposed as a new damage feature in order to use information of residuals and autoregressive coefficients instantaneously. The absolute value of residuals and autoregressive coefficients are used because the directional information of damage feature is not necessary. Making decision whether the considered structure is sound or not using damage features from each sensor in every time step is also very important. The extreme value distribution is utilized to detect outliers because damage information lie in the tail of distribution and the extreme value distribution is well established for the tail distribution. The Generalized Extreme Value distribu- tion (GEV) is adopted for simplicity. The average method with normalization is adopted to draw integrated decision using results of damage detection from each sensor. The validity and accuracy of the proposed damage detection algorithm will be demonstrated through a numerical simulation studies on a two-span continuous truss bridge. A normal operational condition is simulated randomly from three types of vehicle, car, bus and truck. The numerically generated acceleration data with proportional noise under normal operational condition are utilized as measured signals for the numerical simulation example.

REFERENCES

Castillo, E. 1988. Extreme ValueTheory in Engineering, ACADEMIC PRESS, 1250 Sixth Avenue, San Diego, CA 92101 Box, G.E.P.,Jenkins, G.M. and Reinsel, G.C. 1994. Time SeriesAnalysis forecasting and control, Prentice-Hall, Englewood Cliffs, New Jersey 07632 Gumbel, E.J. 1960. Statistics of Extremes, Columbia Univ. Press., New York (1960) Park, H.W.and Sohn, H. 2006. “Parameter estimation of the generalized extreme value distribution for structural health monitoring”, Probabilistic Engineering Mechanics, 21(4), pp. 366–376 Ljung, L., System Identification Theory for the User, Prentice Hall, Inc. (1999) Sohn, H., Czarnecki, J.J. and Farrar, C.R. 2000. “Structural Health Monitoring using Statistical Process Control”, Journal of Structural Engineering, ASCE, Vol.126, No.11, pp.1356–1363

216 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Internal damage localization in a thick plate using moving sensing windows

Yong Han Kim Infrastructure Design & Management Co., Seoul, Korea Hyun Woo Park Donga University, Pusan, Korea Jae Woong Whang Yooshin Engineering Corporation, Seoul, Korea Hae Sung Lee Seoul National University, Seoul, Korea

As MEMS technique and IT technology rapidly develop, monitoring technique on infrastructures using high technologies has been adopted in civil engineering. Researches on applying method of PZT sensor endowed with active excitation and sensing for structural damage detection has been reported. In previous researches, elastic wave excited from PZT sensor applied damage detection on the plate. The plates used in civil infrastructures usually are relatively thick compared to those used in other engineering fields such as aerospace engineering and mechanical engineering. Therefore, the elastic waves propagating on the plate surface are of Rayleigh waves. Originally, Rayleigh waves are uni-directional plain-strain waves which attenuate exponentially with respect to depth in a semi-infinite body. Due to the finite thickness of the plate, there also exist bulk waves propagating through plate thickness. Note that the analytic solutions for these bulk waves propagating through plate thickness have not been available so far. In this paper, damage such as internal crack is localized by using bulk waves through plate thick- ness. Note that the input driving frequency is set to properly produce the Rayleigh wave and bulk wave modes on and within the thick plate. Finite Element Method (FEM) is employed to approx- imately solve a two dimensional elasto-dynamic problem in which both guided waves and bulk waves propagate for a prescribed traction and displacement boundary conditions of the thick plate. Incorporating the conventional wave propagation theories into approximate solutions from FEM, elastic wave modes in the thick plate are investigated semi-analytically. In particular, the most sensitive wave modes for internal cracks have been found to be diagonal bulk transverse wave modes. The sensing ranges of surface bonded active sensors are to be limited because bulk waves attenuate as they propagate through the medium. To overcome this limited sensing range of the active sensor, a new moving sensing window concept is developed for internal damage localization. Each sensing window encompasses of a pair of the surface bonded active sensors. The maximum size of sensing window is determined based on the signal-to-noise characteristics of bulk transverse waves. If there exists an internal crack on the effective diagonal wave path in a moving sensing window, the internal crack can be confirmed by measuring the attenuation of diagonal bulk transverse waves due to the internal crack. Considering all damage-confirmed wave paths of moving sensing windows, the location of the internal crack can be identified on a maximum likelihood basis. The validity of the proposed method is demonstrated by localizing an internal crack in 50mm thick plate. It is shown that the proposed moving sensing window scheme improve the confidence of damage detecting for longitudinally long and thick plate.

217 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Hybrid health monitoring technique for PSC girder using wireless sensing and embedded monitoring algorithm

J.H. Park, J.T. Kim & Y.S. Ryu Pukyong National University, Busan, Korea D. Mascarenas & M.D. Todd University of California San Diego, San Diego, CA, USA

Recently, the interest on structural health monitoring of prestressed concrete (PSC) girder bridges has been increased. For a PSC girder bridge, the flexural stiffness in concrete girder and the prestress force in tendon are two important parameters that should be secured for its serviceability and safety against external loadings and environmental conditions. Since as early as 1970s, many researchers have focused on the possibility of using vibration characteristics of a structure as an indication of its structural damage. Based on the previous works, however, vibration-based approaches can not easily distinguish the two damage-types, stiffness-loss and prestress-loss, unless the information on real damages is known. Therefore, other nondestructive evaluation techniques that are comple- mentary to vibration-based approaches should be sought. Although a proper monitoring system is installed in a PSC bridge, however, the following problems still remain to be solved: (1) the costs associated with installation and upkeep of monitoring systems, which employ wires for the transfer of measurements to a centralized data server, can be high and (2) data repositories with high capacity is needed for future engineering analysis. In this study, we present a hybrid health monitoring technique to alarm damage occurrence, to classify damage-types, and to identify damage locations and severities in PSC girder bridges. In order to achieve the objective, the following approaches are implemented. Firstly, we newly design a hybrid damage monitoring scheme for PSC girder bridges. The hybrid damage monitoring scheme consists of 3 Steps. In Step 1, the occurrence of damage is alarmed in global manner by using fre- quency response changes (that is, correlation coefficient) in a target structure. In Step 2, the alarmed damage is classified by using the change in impedance signatures. Damage in local area near a piezoelectric sensor can be detected by impedance-based method. That is, by attaching the sensor onto tendon or anchor, damage can be classified as: prestress-loss occurred or not. In Step 3, the clas- sified damage is examined in detail for damage localization and severity estimation. Secondly, we develop a smart sensor platform that has wireless sensing capacity with embedded monitoring algo- rithms as shown in Figure 1. Finally, the performance of the hybrid monitoring technique is evaluated using a laboratory-scale PSC girder bridge model with an internal unbonded tendon. The proposed hybrid monitoring technique for PSC girder successfully alarmed and classified the damage. Also, the developed smart sensor platform alarmed the damages of PSC girder in stand-alone.

Figure 1. Schematic of Smart Sensor Platform. 218 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Regularization of inverse problem for damage detection

I. Yoshida Musashi Institute of Technology, Tokyo, Japan C.W. Kim & M. Kawatani Kobe University, Kobe, Japan

An effective maintenance of infra-structures is indispensable in infra-management in Japan. It is required in making decisions as to repair timing and methods based on proper prediction of the structural condition in the future. Development of a damage detection method for existing struc- tures is important in order to perform effective maintenance. However, the first step to effective maintenance is to estimate the current condition of existing structures. Damage detection problem can be interpreted as an inverse problem, of which ill-posedness is a known difficulty. Regular- ization method, which is a technique to overcome the problem of ill-posedness, can be classified into adding information and reducing the size of solution space. In this paper, two regularization methods are introduced for the damage detection problem of a bridge with vibration induced by a moving vehicle. The first approach is to add a priori in-formation. A formulation is proposed to estimate uncertainty level of the a priori and posterioi information, in addition to unknown parameters, using maximum likelihood method. Gaussian or Laplace distribution is used for the uncertainty of a priori information. A bridge model for the numerical simulation of inverse problem is shown in Figure 1. When a vehicle moves on a lane of the bridge, the dynamic response of the bridge is observed at several points. The bending stiffness of an element of the bridge is assumed to decrease at the damaged point. Damage index is defined to be the ratio of the damaged stiffness to undamaged stiffness. For example, when stiffness of an element decreases to 70%, the damage index of the element is 0.7. It is important to estimate stiffness distribution of initial model without damage because the damage index is defined as the ratio of the damaged stiffness to the undamaged stiffness or initial stiffness. Assumed initial stiffness distribution is shown by the line denoted “true” in Figure 2. The vertical axis expresses the ratio to nominal stiffness. Response induced by a moving vehicle is calculated using the stiffness distribution and then response data at the observation points are obtained, which are subsequently used as observation data in damage detection simulations. Three cases are performed, without a priori information, with Gaussian type and Laplace type a priori information. Mean value of damage index in a priori information is assumed to be 1.0. The estimated distributions are also shown in Figure 2. The distribution estimated without a priori information is fluctuating and different from that of the true distribution. The damage indices estimated with Laplace type a priori information is close to 1.0, and also different from that of the true distribution. The distribution estimated with Gaussian a priori information is almost the same as that of true model. It is suggested that when the distribution is smooth, Gaussian type a priori information has an advantage in the identification problem. The position and level of the damage are estimated in damage detection. The damaged point is assumed to be Element 6, and its damage index is 0.7. Three cases of simulation, without a priori information, with Gaussian type and Laplace type a priori information, are also performed. The estimated distributions are shown in Figure 3. Though all cases indicate small damage index around Element 6, the distribution estimated without a priori information has some fluctuations. The case with Gauss type a priori information shows better agreement with the true distribution, however still some estimation errors are observed. On the other hand, the case with Laplace type

219 Figure 1. A bridge model for numerical simulation.

Figure 3. Damage detection of one damaged points Figure 2. Estimation of initial model. model.

Figure 4. Damage detection of three damage points model. shows very good agreement. It is suggested that Laplace type a priori information has advantage in identification problem when only a specific part is damaged. The other approach is to reduce the unknown parameter space, assuming only limited parts are damaged. In the above damage detection, damage indices of all elements are estimated. Unknown parameters are the location and level of damage in the second formulation. The number of unknown parameter is 2nd , when the number of pre-assumed damaged points is nd and the model is one dimensional as shown in Figure 1. Damage detection by the second formulation is conducted using a model with three damaged points, element 3, 6, and 8. Four cases that pre-assume one to four damage points are performed. The 5% noise level is considered in observation data. Figure 4 shows the estimated distribution. Each case shows different result, therefore we can not tell the damage points only from these figures. AIC (Akaike’sInfomation Criteria) of these four cases are estimated. AIC of the case pre-assuming three damaged points is the smallest in these cases. The estimated distribution pre-assuming three damaged points agree well with the true distribution.

220 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Feasibility investigation of health monitoring from traffic-induced vibration data of bridge

Mitsuo Kawatani, Chul-Woo Kim & Tatsuki Fujimoto Department of Civil Engineering, Kobe University, Kobe, Japan

ABSTRACT: For a safety consideration, bridges are checked by periodic monitoring with the aim to minimize the safety risk on the one hand and to keep the costs for the maintenance as low as pos- sible on the other hand by carrying out rehabilitation at the right moment. The technique of Bridge Health Monitoring (BHM) based on vibration data recently has been attracted considerable attention by many civil engineers. This paper proposes a time-domain bridge condition screening procedure based on a pseudo-static approach considering the coupling vibration between bridges and mov- ing vehicles including the effect of roadway surface roughness. The basic idea of the pseudo-static approach for bridge health monitoring in this study is to detect change of bridge’scondition directly from the change in element stiffness. An advantage of using the traffic-induced vibration data of bridges in the damage identification may be ease of exciting the bridge. Another important point is that some highway bridges are tested before/during in-service such as a moving vehicle test. It means that there exist traffic-induced vibration data available for their health monitoring. In this study model reduction methods such as Guyan reduction and SEREP (System Equivalent Reduction Expansion Process) which are essentially substructuring methods are also applied. An object function for damage identification is defined using a pseudo-static formulation in the form of a linear system of equations subtracted from equations of motion for the bridge-vehicle interaction. The governing equations of motion for the bridge-vehicle interaction system generally contain time-variant coefficients, which is the reason why this study adopts a time domain approach. The feasibility of bridge health monitoring using the traffic-induced vibration data is investigated through a numerical analysis. A simple-supported girder bridge with span length of 40.4 m is adopted as a numerical example. The damage is idealized by decreasing the element stiffness of the bridge. Assuming that the damage changes the bending rigidity to simplify the derivation, the change of the element stiffness is obtainable using the Element Stiffness Index (ESI). Dynamic responses taken from a traffic-induced vibration analysis of the bridge idealized as a bridge model with fine mesh (see Figure1) are assumed as measured data. The simulation for the traffic-induced vibration of the bridge is based on the modal analysis. The model for damage identification and damage scenarios are shown in Figure 2. Estimated ESI values under a single moving vehicle with 20 km/h are summarized in Figures 3 and 4. Figure 3 shows damage identification results by using Guyan Expansion, while Figure 4 gives those results by using SEREP.The RESI in the figure denotes the reference ESI value, whereas the IESI indicates the identified one. Comparing Figure 3 with Figure 4, SEREP gives a better identification than Guyan reduction for undamaged elements. On the other hand, SEREP procedure overestimates the rigidity of the damage element in comparison with that of Guyan reduction.

Figure 1. FE model of bridge for modal analysis.

221 Figure 2. Damage scenarios: (a) Scenario 1 (SCN1); (b) Scenario 2 (SCN2); and (c) Scenario 3(SCN3).

Figure 3. Damage identified results using Guyan (10% noise; v = 20 km/h): (a) SCN1; (b) SCN2; and (c) SCN3.

Figure 4. Damage identified results using SEREP (10% noise; v = 20 km/h): (a) SCN1; (b) SCN2; and (c) SCN3.

The summarized result shows that the proposed method is able to identify damages’ location. The error of identified rigidity of each elements, however, varies from 2% to 33% by using Guyan reduction and from 0.3% to 19% by using SEREP. Investigations through the numerical study of this paper suggest feasibility of the proposed method to damage identification of bridges, even though many tasks remain to be solved.

222 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Localization of damage in a bridge using measured response signals

S. Shin & H. Park Inha University, Incheon, Korea

ABSTRACT: A damage detection and localization algorithm is proposed using a single-input- single-output (SISO) ARX model that provides a correlation between input and output signals. The input signal is the measured acceleration at the impact location and the output is also the measured acceleration at a place of the other accelerometers uniformly distributed in a structure. To detect and locate damage, it is assumed that a baseline statistical property for an intact condition is provided as a priori known information. By comparing the statistical properties of the current condition with those of the baseline, damage can be detected and localized. Three damage indices are proposed by using the statistical properties as the absolute difference in the area of the function, the ratio of the maximum values, and the ratio of the variances of the intact and damaged conditions. A laboratory experiment on a steel plate girder bridge model has been carried to examine the usefulness of the proposed algorithm. The results are compared in the figures below.

Figure 1. Computed damage index DI1.

Figure 2. Computed damage index DI2.

Figure 3. Computed damage index DI3.

223

Design and analysis

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Stonecutters bridge – Design for operation

Matt Carter & Naeem Hussain Ove Arup & Partners, Hong Kong

Highways Department (HyD) of Hong Kong SAR is building a new bridge across the Rambler Channel at the entrance to one of the world’s busiest container ports. The bridge will become the second longest spanning cable-stayed bridge in the world. Due to the spectacular location adjacent to Hong Kong’s harbour, HyD decided to hold an international design competition to select the bridge concept. This led to the scheme of a cable-stayed structure with freestanding towers on land and a clear span of 1018 m to leave the navigation channel unobstructed. The twin box girder deck, which is steel in the main span and prestressed concrete in the back spans, passes either side of the mono-column towers. The paper describes the measures taken during the design to provide a durable structure to achieve the 120-year design life. Where suitable, corrosion resistance materials are provided, such as the stainless steel skins to the composite upper towers, and the stainless steel reinforcement in the lower sections of the towers. The steel deck is coated with a high-build epoxy paint system on the outside, and is dehumidified on the inside. Through the life of the bridge, maintenance, repair and some component replacement will be necessary. A comprehensive set of facilities for inspection and maintenance of the structure was included in the design. This includes fixed access facilities and a suite of motorised access machines (Fig. 1). The bridge is fully instrumented with a real time Wind And Structural Health Monitoring System (WASHMS). This paper is submitted with the permission of the Highways Department, HKSAR.

Figure 1. Overview of access facilities.

227 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Strut-and-Tie Method for FRP strengthened deep RC members

S. Park Urban Management Division, Seoul Metropolitan Government, Seoul, Korea R.S. Aboutaha Department of Civil & Environmental Engineering, Syracuse University, Syracuse, NY, USA

Fiber Reinforced Polymer (FRP) composites have been successfully used for strengthening existing reinforced concrete bridges with inadequate traffic load carrying capacity, and inadequate seismic resistance. During the last two decades, a great deal of research on FRP retrofitted concrete bridge flexural members has been conducted. However, most of these investigations have primarily focused on the flexural and the shear strengthening of slender flexural members. Very limited work has been done on FRP strengthened deep members. Therefore, FRP strengthened bridge piers are still being analyzed using approximate procedures that have been developed for more slender members. For analysis of FRP strengthened deep reinforced concrete members, the Strut-and-Tie Method (STM) offers a powerful analysis tool, as the externally bonded FRP sheets would act as an additional tension ties. In this paper, a practical analysis and design process of FRP strengthened deep reinforced concrete members using the Strut-and-Tie Method is presented.

228 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Design of the Machang Mainbridge

Eul-Ho Bae, Soung-Yen Kim & Ryang-Gyun Kim Hyundai Engineering Co., LTD., South Korea Siegfried Hopf, Armin Patsch & Peter Walser Leonhardt, Andrä und Partner, Consulting Engineers VBI, GmbH, Germany

ABSTRACT: The Machang bridge is part of the highway connecting the cities Masan and Chang- won crossing over the Masan Bay. This high level crossing has a total length of 1,700 m and is composed of a 546 m long EastViaduct, the 748 m long Mainbridge and a 406 m long WestViaduct. The mainbridge is a two tower cable-stayed bridge with a mainspan of 400 m, 170 m sidespans and 4 m cantilevers at each holddown pier. The 21 m wide bridge deck carries the dual two lane carriage- ways including 2.0 m wide reserve areas at each side. The superstructure consists of two 2.50 m deep steel composite plate girders and of the 29 cm thick concrete slab. Steel crossbeams are provided at regular spacing of 4.15 m to support the concrete slab. The deck slab is made of prefabricated concrete panels which are connected over the crossbeams by cast in-situ joints. The stay cables, which are of parallel wire type, are provided at regular spacing of 12.45 m at the deck. The stay cable arrangements in vertical planes has been achieved by providing additional cable-crossbeams. The aesthetically pleasing H-shaped pylons have an height of about 158 m above foundation. The 3 upper crossbeams, which consist of diameter 2.0 m steel tubes, provides the structure with a very distinctive appearance. At the upper pylon legs, a steel composite cable anchorage type has been designed. The pylons are supported on deep foundations each with 28 bored piles dia. 3.0 m which are designed also to resist impact loads from aberrant ships. In order to reduce the seismic loads on the foundation, a combination of shock transmission units and shear keys are provided at the deck connection to the pylons. Non-structural wind fairings with rounded shapes are attached to the superstructure edges to improve the aerodynamic stability of the deck and to reduce wind loadings. The piers are build using jump forms, while the pylons are constructed by slip forming. The steel composite deck is build using the free cantilevering method. The starter segment at the pylons is lifted in place completely with the concrete slab by a 3000 to capacity floating crane. It is secured by the crane until the first cable is installed and stressed. The steel work of the starter segment at the hold down piers is also lifted in place by the floating crane onto temporary piers and the concrete slab is cast after lifting. For the typical deck segment erection a standard derrick with approximately 110 to capacity was used to lift the steel grid and the concrete panels from the delivery barge. The project of the Machang Bridge has been ordered as private investment method in Korea. The design of the Machang Bridge has been performed in 2003/2004 in close collabo- ration between the Korean designers of Hyundai Engineering (HEC) and the German consultant Leonhardt, Andrä & Partner (LAP). Construction has been carried out in 2005/2008 by Hyundai Engineering & Construction (HDEC), Korea and the French contractor Bougyues Travaux Publics.

229 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Durability of suspension bridge with multi main spans

Manabu Inoue IHI Corporation, Tokyo, Japan Mitsuhiro Kudo & Kazuma Doi IHI Corporation, Aichi, Japan Yukiaki Takizawa IHI Corporation, Tokyo, Japan

Recently, the continuous suspension bridges with multi main spans have been planned in some countries mainly because of the aesthetical and the economical backgrounds. Some bridges have main spans of more than 500 m. To take a step forward for the realization, this type of bridge should be competitive to alternative bridge types such as cable-stayed bridge in the points of durability, stability both in service and during the construction, both initial and life cycle costs and so on. Understanding the behavior of towers, especially the inside towers, is of importance to keep stabilities both in the points of structural safety and serviceability. For example, to minimize the error sensitivity is one of key points to achieve the bridge with adequate durability. In reality, several kinds of imperfection must be inevitable during the fabrication and the construction. Espe- cially, the imperfection of center tower always deteriorates the ultimate capacity of tower for the suspension bridge with even main spans. Also, the suitable rigidity of tower is essential to construct the bridge with higher quality. Furthermore, the bridge articulations may have much influence on serviceability and LCC. In this study, the bridge configuration for a suspension bridge with multi main spans has been investigated mainly in the points of the stability and the durability of towers. The fundamental characteristics both during the construction and in service such as the effect of sag ratio to cable slippage resistance were shown using a 5-span continuous suspension bridge model with each main span of 800 m. Then the relationship between the ultimate capacity of bridge and the tower imperfections has been investigated by the parametric analyses. Based on the results, some remarks for the characterize features and the possible criteria during the construction were noted for the towers. For a continuous suspension bridge with multi main spans, understanding the behavior of towers is of essential and of interesting to realize with suitable stability and durability.The major concerning issues are the cable slippage resistance, the running stability and the durability. These are inevitable subjects to be solved for the realization of a continuous suspension bridge with multi centre spans. Especially, how to handle the imperfect sensitivity during the fabrication and the construction has a big effect on stability, durability, cost and so on.

Figure 1. General geometrical layout.

230 Finally, there are still many problems or concerning issues to realize a suspension bridge with multi main spans. However, some special features of towers and a part of required accuracy for tower fabrication/construction can be clear through this study. The well-prepared plan/design and the well-controlled construction method with a consideration of the imperfect sensitivity and with a possible care can make a durable and a stable bridge possible.

231 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Effect of main steel corrosion on the stiffness of corroded reinforced concrete beams

F.J. O’Flaherty, P.S. Mangat & P. Lambert Centre for Infrastructure Management, Sheffield Hallam University, Sheffield, UK E.H. Browne Halcrow Group Limited, Birmingham, UK

Deterioration of reinforced concrete is principally caused through corrosion of the reinforcing steel when it becomes subjected to a severe corrosive environment containing chlorides and carbon dioxide. However, before repair is carried out, it is important that the structural capacity of deteri- orated members can be estimated to ensure repairs are executed at the appropriate time. This paper presents information on the influence of corrosion on the structural stiffness of corroded beams reinforced with different percentages of main steel reinforcement. Influence of design parameters such as concrete cover to the main steel is also related to the structural performance. Influence of main steel corrosion on the serviceability limit state is also considered. A total of thirty-eight reinforced concrete beams were tested in flexure. The results show that stiffness reduces with increasing main steel corrosion. In addition, beams with lowest covers of 26 and 36 mm retain a higher residual stiffness at an arbitrary 10% corrosion whereas those with higher cover (56 mm) exhibit a lower residual stiffness of 43% and 49%. Similarly, the rate of stiffness reduction due to main steel corrosion generally tended to increase in negativity for beams with higher covers. For example, the slope of stiffness versus degree of corrosion for beams reinforced with 2T8 with 26 mm cover is −0.43 whereas the slope for beams with 56 mm cover, reinforced with 2T8 and 2T12 is −0.53 and −0.79 respectively. Steeper slopes mean that beams suffer a more rapid decrease in stiffness with higher corrosion. The paper also relates the serviceability limit state to the degree of corrosion. A service load is estimated as a percentage of the ultimate (control) load as is assumed as 40% of the ultimate limit state. This accounts for factors of safety included in the design for dead (1.4) and live (1.6) loads and materials (1.5 for concrete and 1.05 for steel). The allowable deflection under serviceability conditions is assumed as span/250. The results show that beams exhibiting main steel corrosion up to 10% exhibit deflections less than the allowable deflection (3 mm). Beams exhibiting main steel corrosion greater than 10% generally failed in flexure before reaching the service load. Therefore, concrete beams in practice with main steel corrosion approaching 10% should be considered as reaching their serviceability limit state and repair and maintenance is required to extend their service life. If a factor of safety is applied to this figure, say 1.5, then beams exhibiting main steel corrosion of 6–7% should be considered as requiring attention in the near future. The main conclusions from the results reported in the paper are: – reinforced concrete beams show a loss in stiffness with increasing corrosion of the main steel reinforcement; – reinforced concrete beams with lower concrete cover to the main steel generally tend to retain a higher residual stiffness as opposed to those with higher cover at the same degree of reinforcement corrosion; – reinforced concrete beams with higher covers tend to suffer a more rapid decrease in stiffness with higher corrosion; – beams exhibiting main steel corrosion greater than 10% generally failed in flexure before reaching the service load. 232 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Optimum life-cycle-cost design for bridge structures considering damage probability

Y.S. Shin, J.H. Park & T.H. Kim Dept. of Construction and Transportation Engineering, Ajou University, Suwon, Korea

The importance of the Life Cycle Cost (LCC) analysis for construction projects of bridge has been recognized over the last decades. Accordingly, theoretical models, guidelines, and supporting software have been developed for the LCC analysis of bridges. However, it is difficult to predict LCC precisely since the costs occurring throughout the service life of the bridge depend on various parameters such as design, construction, maintenance, and environmental conditions. This paper presents a methodology for the optimal design of bridge structure. Total LCC for the service life is calculated as the sum of initial cost, damage cost, maintenance cost, repair and rehabilitation cost, and user cost. The optimization method is applied to design a bridge structure with minimal cost, in which the objective function is set to LCC and constraints are formulated on the basis of Korean Bridge Design Code. Initial cost is calculated based on standard costs of the Korea Construction Price Index and damage cost on damage probabilities to consider the uncertainty of load and resistance. An advanced first-order second moment method is used as a practical tool for reliability analysis using damage probability. Maintenance cost and cycle are determined by a stochastic method and user cost includes traffic operation costs and time delay costs. Optimal design is performed for various bridge types such as steel-box girder bridge, plate girder bridge, PSC-I girder bridge including the substructure and the effects of various parameters are investigated. This study performed optimal LCC design of a 4@40 m bridge with width of 15.6 m according to the type of superstructure and supporting piers. The types of superstructure were concrete slab, steel box girder and PSC-I girder, and those of the piers were single-column or double-column piers.

Figure 1. Annual cost for piers. Figure 2. Annual cost of bridge by road type.

233 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

The study on the methods for slimming bridge

Kyeong Ho Kim, Chul Soo Lee & Kyu Sun Hong Chungsuk Engineering, Structural Div., Seoul, Korea C.-G. Lee HTTI, Hwaseong, Gyeonggi-Do, Korea

Modern highways have increasing number of bridges and tunnels as straight alignments are being preferred, and therefore, continuous research and technical advancement as per design, construc- tion as well as maintenance are becoming more important. Besides the technical aspect of a bridge, demands for various aesthetic elements, optimal design for minimal space and environment- friendly structures that harmonize with the surrounding are also rising, and in order to achieve these elements, modeling of a bridge into a slim type can be thought as the most crucial factor. In this study, status and trends of Korean and foreign bridges are compared, while analyzing elements that affect slim structures, which are divided into 4 categories in large, the design speci- fication, characteristics of materials, design customs and landscape of accessory bridge structures. As the first element in the category, the seismic design specification in Korea is over-evaluated. The seismic load is the main factor that determines the cross section of a pier, and therefore, slim structure is designed through a seismic analysis appropriate for Korea’s condition. In terms of material characteristics, slim structure is designed as reducing weight of the superstructure and making the substructure slimmer by applying concrete and steel with higher strength. In terms of design custom, unlike USA or Europe where they use the LRFD, Korea uses design based on allowable stress for steel bridge. In design custom of Korea, structural calculations are performed based on the initial assumption of cross section, instead of optimal cross section, and slim structure design is implemented by repeating calculations to come up with an optimal cross section. In terms of the last element of the category, the landscape of bridge’s accessory structures, slim design can be very different due to various landscape elements and aesthetics of accessory structure even if the girder heights are similar or cross sections are same Through such establishment of slim design methods, it can be seen that economic bridge design as well as technological advance of construction and maintenance and minimization of environmental damage can be achieved.

Keywords: slimming bridge; aesthetic elements; LRFD; optimal cross section.

234 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Design of deck continuity details for steel and prestressed concrete bridges

A.P. Ranasinghe & W.L. Haugeto Michael Baker Jr. Inc.

Leaking of storm water and deicing salts through bridge deck joints is a major cause for the deteri- oration of bridge joints, bridge bearings and bridge girders. It also results in unsightly discoloring of abutments and piers of steel bridges. Based on these reasons, many bridge agencies all over the world are getting rid of the deck joints in existing simple span bridges and replacing them with continuous decks over piers. Unfortunately, many continuity details also experience maintenance issues and have been found to crack and leak at the point of continuity. By modifying the continuity details for both steel and concrete bridges, these maintenance issues can be significantly reduced, resulting in continuity details that will experience minimal maintenance issues. Numerous continuity details for both steel and concrete bridges are currently used throughout the United States, and several states have developed standard continuity details for use on their bridges. The existing details shown in this paper have been used in New Jersey, USA by the New Jersey Department of Transportation or by the New Jersey Turnpike Authority. However, similar details have also been used in other states. Continuity can be achieved in simple span steel girder bridges by connecting the simply supported girders with a plate to the top flanges of girders in adjacent spans then pouring the deck continuously over the location of the former joint. In simple span prestressed concrete girder bridges, the continuity is achieved by pouring the deck continuously over the piers between the girders of adjacent spans. Many existing and new steel and concrete bridges with simply supported girders where the deck is made continuous over the supports have shown cracking of the deck and often complete separation at the point of continuity.After examining the behavior of these details in the field, several maintenance issues were observed. For the steel continuity detail, cracking in the continuity plate was observed after the plate was in service for over 10 years. Cracking of the steel plate results in subsequent cracking in the top of the concrete deck. For the concrete continuity detail, cracking in the top of the deck is observed almost immediately after construction. When cracking is observed in the top of the deck in both details, maintenance issues will ensue when deicing salts penetrate the deck through these cracks. Improvements to the continuity detail can be made for both steel and concrete superstructures to decrease long-term maintenance costs of this detail. By slightly adjusting the details shown in this paper by modifying several parameters, including the deck thickness, concrete strength, effective deck width, top and bottom reinforcement steel areas, girder rotations, temperature, crack control provisions, exposure conditions, the cracking of the deck over the supports can be minimized. The paper details several continuity options used in both steel and concrete bridges, their maintenance issues, and provides suggestions to improve them.

235 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Reliability-based optimum design of high-speed railway bridges considering structure-rail interaction

Jong-Soon Lee Bridges & Rails Engineering Consultants Co., Ltd., Seoul, Korea Hyo-Nam Cho Department of Civil & Environmental Engineering, Hanyang University, Ansan, Korea Yeong-Rok Ihm Department of Civil Engineering, Kyonggi University, Suwon, Korea

Nowadays, the constructions of high-speed railway in Korea and many other countries are gradually increasing. But 3 years after partial opening of Kyeongbu High-Speed Railway, it was confirmed that the dynamic response of the constructed high-speed railway bridge exceeds the limit of the design standard, which is vertical acceleration of deck and rotation of end deck affecting the operation safety of train, in some bridges. This result was caused by conservative bridge design without respect to bridge structure-rail interaction in existing bridge design. Therefore, in this study, it is suggest that it has to reliability-based design methodology with respect to bridge structure-rail interaction. To consider the structure-rail interaction, it was performed the sensitivity analysis for main effect factor and numerical analysis of initial cost optimization, LCC (Life Cycle Cost) optimization and equivalent LCC optimization to secure the structural safety, track safety and vehicle safety, and it compares and analysis the result of that. For the structural analysis, commercial package, ABAQUS, are used for a three-dimensional finite element analysis. It was modeled the frame element of 6 degree of freedom for deck of bridge and rail and modeled the bi-linear element to consider the nonlinear for ballast. And it is used the nonlinear spring to simulate the longitudinal stiffness for bearing. The optimization process utilizes a well-known optimizer, ADS (Automated Design Synthesis). Optimization technique is utilized the ALM-BFGS method for global area search and Golden Section Method for 1-D search. The RSM (Response Surface Method) is used for reliability analysis of strength and serviceability of the bridges, the strength of CWR (Continuous Welded Rail) and the comfort of passengers. In general, ALM-BFGS method is not need the 1-D search, and that algorithm converge a 0.1∼0.2 of Push-Off factor. But in this study, value of Push-Off factor is used 90, therefore 1-D search should be need for effective convergence. That algorithm contains the "heuristic decision method". To reduce the structural analysis running time, it was performed the analysis for dynamic behavior analysis of structure and time history analysis of CWR, and it was reduced the 50% of analysis time. To improve the effectiveness and economics the bridge design methodology considering the structure-rail interaction suggested in this study, the LCC effective optimum design is applied to a 2-main steel girder bridge, 5 × (1@50 m) for comparison with conventional design, initial cost optimization and equivalent LCC optimization. As a result of the optimum design based on reliability, it may be stated that the design of High-Speed railway bridges considering the structure- rail interaction is more efficient than typical existing bridges and LCC optimization without respect to structure-rail interaction. The result of optimization design considering the interaction is higher than result of initial cost optimization design in initial cost, but that has the advantage than result of initial cost optimization design in expected LCC. Therefore, design methodology suggested in this study was demonstrated more economic and efficient than existing design and LCC optimization not considering structure-rail interaction. Keywords: Reliability-Based Optimum Design, CWR, Interaction of Structure-Rail, ALM-BFGS

236 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Cyclic response of the precast SRC bridge piers

Y.-S. Chung, C.S. Shim, J.-Y.Yoon & J.-H. Park Department of Civil Engineering, Chung-Ang University, Ansung, Korea

In the design of bridge piers in congested area, it is necessary to reduce the dimensions of columns considering design criteria and to minimize the construction period. In the civil engineering com- munity, prefabricated structures and systems have received increasing attention as a way to increase construction speed, to enhance work zone safety and to minimize the environmental impact. A pre- fabricated substructure system has been suggested and is shown in Fig. 1 was suggested. The embedded steel member can increase ultimate strength and seismic performance, and also pro- vides a connection to other parts without prestress. This concept also can be applied to building structures. Especially in seismic area, the ductility requirement is the most important factor. In order to enhance the seismic performance of bridge piers under combined actions, it is necessary to make the ductility of columns larger by covering RC columns with steel tubes or confining the columns by arranging transverse reinforcement such as hoop ties closely. Using core steel composite columns is useful as one of the reinforcing RC columns for increasing static and seismic performance. In this paper, several tests on concrete encased composite columns, which are prefabricated, were performed to investigate the seismic performance of the composite columns. Quasit-static tests were carried out and their performance was evaluated and compared each other. The cross-sections of these specimens are composed of concrete-encased circular tubes with partial in-filled concrete. Through the tests, it was evaluated the ductility of SRC composite specimens. It has become clear from the test results that encased steel elements makes the deformation capacity of the columns to be large. For relatively small size columns, precast SRC columns can be excellent alternatives for bridge substructures without any prestressing. Based on the test results, design considerations on the details of the reinforcements were proposed.

Figure 1. Prefabricated composite column.

237 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

General design method for the shear connection according to failure modes

C.S. Shim & S.-M. Jeon Department of Civil Engineering, Chung-Ang University, Ansung, Korea P.-G. Lee Civil Engineering Research Department, RIST, Hwasung, Korea

Stud shear connectors are influenced by several parameters according to previous researchers, with major factors categorized into shank diameter, height and tensile strength of studs, compressive strength and elastic modulus of concrete, and direction of concrete casting. In addition for shear connection in precast deck bridges, material properties of filling material and bedding height must also be considered for the evaluation of structural performance of stud shear connection. The failure modes of shear connection can be categorized as in Figure 2. Mode-1 is defined as stud failure without considerable concrete damage. Mode-3 means the concrete failure without stud failure. When the connectors are failed after considerable concrete damage, it can be defined as Mode-2. In high shear region, we need stronger shear connection. Group stud shear connection is dealt with in this paper in terms of static strength and fatigue endurance. Push-out tests were conducted for the shear connection in composite trusses and in precast decks. For the group stud connection of composite truss bridges, it is adequate to use current design codes for shear connection which satisfy the requirement of minimum spacing of current design codes. Since the shear load is concentrated at the joint structure of composite truss bridges, additional confining reinforcements are needed to enhance the horizontal shear strength of the shear connection. For the specimens of precast deck bridges, the effect of the stud spacing and confining reinforcements was clearly observed. Decreasing the stud spacing resulted in lower ultimate strength of the shear connection. The confining reinforcements inside and outside of the shear pocket can enhance the shear strength of the shear connection. The requirement of the minimum spacing for the stud connectors needs to be revised for precast decks. However, the shear connection with smaller spacing should have adequate reinforcement details to increase the failure load of concrete slab. In this paper, the empirical equation was proposed and fatigue endurance of the shear connection with group arrangement was verified.

Figure 1. Shear connection for high shear. 238 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Structural system durability through jointless bridge decks

U.B. Attanayake Western Michigan University, Kalamazoo, MI, USA A.E. Ulku Wayne State University, Detroit, MI, USA H.M. Aktan Western Michigan University, Kalamazoo, MI, USA

Bridge decks are expected to satisfy two functional requirements: (1) distribute live loads between the girders and (2) provide shelter to the superstructure and substructure components. Bridge deck joint failure compromises the intended purpose of providing shelter to protect the components from detrimental actions such as exposure to deicing salts and other chemicals transported with surface water. Corrosion induced deterioration is a major problem and it is more pronounced in older bridges especially near the girder-ends (Whiting et al. 1993 and Aktan et al. 2002). Upon deck joint failure, in colder climatic zones, water saturated with deicing salts downpours over the girder ends exposing the surface to chlorides and chloride ingress. With repeated exposure girder end deteriorate prematurely. Girder end condition is important on two aspects: first, load path from superstructure to the substructure is through the girder end and, second, in the case of prestressed concrete (PC) girders, girder ends are within the prestress transfer zone to the concrete. Hence, loss of any prestress will compromise bridge safety. Field investigation of 20 PC I-girder bridges was carried out by Aktan et al. (2002) to identify the causes and progression of girder end deterioration and to develop recommendations for abating or minimizing the distress. One of the major causes of girder end deterioration was attributed to the girder ends being located directly below poorly maintained deck joints and thus exposed to deicing salts draining from the deck surface. Caner and Zia (1998) recommended using a slab connecting the neighboring spans, which is called the link slab, to eliminate the deck joints over the piers. The intent of using link slab is to improve bridge durability and to achieve smooth riding surfaces. The link slab is designed according to the rationale developed by Caner and Zia (1998). Design procedure allows fine cracks to develop in the link slab. Water penetration through the link slab may still be possible through the fine cracks but is a more tolerable situation than those of jointed decks. However, recent field inspections of eight bridge decks with link-slabs in Michigan revealed full-depth cracking with significant width directly over the piers. Finite element models were developed to evaluate the effects of link slab design parameters in conjunction with live and thermal loads on the link slab performance. The design parameters of the link slab used in this analysis are: debond length, girder height, span length ratio, and support conditions. An assemblage model that consists of two simple span girders was developed. The effects of varying debond lengths on link slab stresses were investigated by using debond lengths of 0%, 2.5%, 5.0%, and 7.5% of the span length. The effect of girder height is examined using two different standard PC I-girder sections (Type III and VI). Analyses were performed evaluating the effects of adjacent span ratios. Vertical and horizontal springs were incorporated into the model simulating the effects of bearing stiffness. Influence of three different support configurations on link slab behavior was investigated. They are: (a) HRRR (Hinge-Roller-Roller-Roller), (b) RHHR (Roller-Hinge-Hinge-Roller), and hinge supports at either side of the interior supports (RRHR or RHRR). 239 Live load and thermal gradient load cases were considered independently for all models. Dead load was not taken into account since its effects can be eliminated by placing link-slab segments upon completion of the deck placement sequence. Live load was applied symmetrically to both spans. HL-93 (AASHTO LRFD 2004) loading with an impact factor of 1.33 was placed so as to create maximum girder end rotation. A lane load of 9.34 kN/m was used in conjunction with the axle loads. Negative and positive thermal gradient loads were applied to the composite girder-deck cross-section. Thermal gradient loads were calculated based onAASHTO LRFD (2004) for Zone-3. Negative temperature values were obtained by scaling the positive temperature values by −0.30. A uniform thermal expansion coefficient of 10.8 × 10−6/◦C was used for both deck and girder concrete. Following conclusions are derived from the analysis results: 1. The link slab is under combined bending and tension for the HRRR, RHHR and RRHR/RHRR support conditions. 2. Refined FE analysis showed that the link slab bottom fiber is in compression for HRRR and RRHR/RHRR cases even though the link slab resultant forces are in tension under live load. 3. Link slab moment decreases with increasing debond length. Increasing debond length is not recommended since this would lead to shrinkage and thermal hydration cracking. Debond length should be established based on the beam support configuration. 4. Live load moments developed at the link slab decrease with deeper girder sections due to the reduction in link slab relative stiffness. 5. Negative thermal gradient load case will increase live load induced moments when superposed. However, AASHTO LRFD (2004) does not provide load factors under Strength I limit state for combining the load effects for the design. Load factors are provided only for the Service I limit state that is used for controlling crack widths of reinforced concrete components. 6. In the case of full deck replacement or deck joint replacement projects, link slab concrete is placed last to eliminate the effects of dead load. Under this implementation, link slab is restrained at both ends. Drying shrinkage and heat of hydration may generate substantial strains at early ages when tensile strength of concrete is low. Formation of the full-depth cracking is attributed to drying and thermal hydration stresses induced in the link slab. From this study the following recommendations are developed: 1. Link slab should be designed considering flexural interaction with axial loads. 2. Link slab debond length should be selected as short as possible for minimizing the thermal and shrinkage effects. 3. Link slab debond length and the axial force and moment demand on the link slab should be determined considering girder geometry, adjacent span ratio, and support conditions for the particular structural system.

REFERENCES

AASHTO. 2004. AASHTO LRFD Bridge Design Specifications. Third Edition. Washington, D.C. Aktan, H., Koyuncu, Y., Rutyna, J., Ahlborn, T.M. & Kasper, J.M. 2002. Causes and cures for prestressed con- crete I-beam deterioration. Research report RC-1412. Michigan Department of Transportation. Lansing. Michigan. Caner, A. & Zia, P. 1998. Behavior and design of link slab for jointless bridge decks. PCI Journal. May–June: 68–80. Whiting, D.A., Stejskal, B.G. & Nagi, M.A. 1993. Conditions of prestressed concrete bridge com- ponents: Technology review and field surveys. FHWA-RD-93-037. Federal Highway Administration. Washington, D.C.

240 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Modal flexibility-based FEM model updating for bridges

Jintao Cui & Dookie Kim Department of Civil Engineering, Kunsan National University, Kunsan, Korea Ki Young Koo Department of Civil and Environmental Engineering, KAIST, Daejeon, Korea Hie Young Jung Department of Civil Engineering, University of Seoul, Seoul, Korea

This paper represents a new FEM model updating method for bridges using deflections obtained by modal flexibility matrix from dynamic structural response measurements. Generally, natural frequencies and mode shapes have been widely used in FEM model updating procedures instead of deflections since direct deflection measurements are hard to be obtained in practical situations. In this study, a new FEM model updating method is proposed based on static deflections obtained by dynamic measurements via modal flexibility. The proposed model updating method is easy in implementation and fast in computation since only static analysis is needed during optimization avoiding eigen-analysis. Experimental verification was carried out on a steel-box girder bridge model and it was found that the proposed method reasonably fits the estimated deflection. The stiffness and the flexibility matrices related to the modal data can be simply transferred from Equations (1) and (2) as following:

Figure 1. The cost function values during optimiza- tion process. Figure 2. The optimized values at the final step.

241 where K is the stiffness matrix; M is the mass matrix; = [φ1, φ2, ..., φn] is the mode shape = 2 2 matrix, φi is the ith mode shape, diag(ωi ) is the modal stiffness matrix, ωi is the ith modal frequency, n is the number of degrees of freedom, and G is the flexibility matrix. The mode shapes and the natural frequencies can be obtained from either analytical or experimental data.

242 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Optimal design of a steel box girder bridge considering aesthetics

Y.S. Shin, J.H. Park & G.O. Kim Department of Construction and Transportation Engineering, Ajou University, Suwon, Korea

Former bridge design in Korea focused essentially on money and time savings. Such design was the natural answer to the rapid industrialization, which needed to construct transportation infras- tructures in records of time within the cheapest costs. Recently, the growing public demand for aesthetically pleasing bridges resulted in the introduction of aesthetics concern in design. How- ever, engineers are encountering huge difficulties in applying aesthetics to the design of bridge because of the absence of absolute value to evaluate aesthetics. This paper presents a method to determine the optimal design of a steel-box girder bridge superstructure considering aesthetics. Optimal design is performed for a variety of span ratios standing as a major factor determining the aesthetics of bridge in the design and evaluation is done by comparing the economical efficiency of each alternative. Therefore, construction cost is set as objective function. Three-span contin- uous and four-span continuous bridges are used as numerical examples. Results show the effect of the consideration of aesthetics in design on the construction cost. The analysis results can be summarized as follows. – Total construction cost of bridge is correlated to span ratio. – Construction cost of optimal design applying aesthetic span ratios is comparatively economical. The concept of aesthetic is so subjective that the bounds of research are extensive. Research on the aesthetic bridge requires much time and labor. But engineers have the responsibility to design safe, economical and aesthetic bridges. This paper is a simple attempt to apply aesthetic concepts to the design of bridge. Proper consideration of aesthetics together with life cycle cost will undoubtedly result in more reasonable design of bridges.

Figure 1. Construction cost of 3-span continuous Figure 2. Construction cost in 4-span continuous bridge according to span ratio. bridge according to span ratio.

243 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Complex shapes and innovative technologies for bridges

I. Paoletti Researcher of Technology in Architecture, Faculty of Architecture, Polithecnic of Milan, Italy

Complex shapes are becoming part of our architectural scenario. In this field bridges has an important role pursuing the need to merge morphological shape with structural constraints. More- over innovative materials and technologies are in those projects a challenge and an opportunity to increase performances. This paper will show examples of bridges around the world where challeng- ing structural and construction constraints were converted to opportunities showing the possibility of merging artistic beauty and structural efficiency. Today the influence of surface and solid geometrical modeling powerful algorithms allows an unprecedented morphology freedom in architecture. This freedom has brought innovation in the traditional design methodology of structural engineering, generating process uncertainties in reli- ability assessment in general and structural safety warnings in particular. This statement is even more truth for bridges, where the architectural shape and creativity really need a technical support in order to be realized. Starting from the scenario above described some challenges seem to raise on which bridges will probably have to face in the next years. The first one is the research of lightness., the second theme that can also stimulate innovation is mass-customization, a third one is is calculation potentialities, where the advantage offered by informatics and automation has been very important in the field of structural design in general and particularly significant in the case of special structural systems. Last but not leastThe Sheikh Zayed Bridge project by Zaha Hadid falls within the broader project designed to connectAbu Dhabi with the express highways and expressways of the rest of the United Arab Emirates. The structure of Hadid looks like a wave through the channel. The plans are road bump on the two sides of the undulating structure that constitutes the backbone of the bridge. The structure which cling to the lane reaches a height of 60 feet above the water level, while the road rises to an altitude of 20 metres. This project has been designed in a virtual environment in a manner that structural constraints are already verified, event roughly, on the design phase. The last topic is energy. As per other building typology also bridges are involved in energy saving polices. Undoutably, some works achieve the level of architectural and sculptural art and the role played by structures is merely to support architectural design, however a concern on economic value and best practice methodology must be included. Conversely, a structural forgery and/or morphological sculptured shapes making any prismatic configuration building look outdated, may induce students and professionals without sufficient knowledge and expertise to elaborate imitations that introduce dangerous design uncertainties.

244 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Reliability-based design optimization using a response surface method

S.C. Kang & H.-M. Koh Seoul National University, Seoul, Korea

In conventional structural design, deterministic optimization which satisfies codified constraints is performed to ensure safety and maximize economical efficiency. However, uncertainties are inevitable due to the stochastic nature of structural materials and applied loads. Thus, determi- nistic optimization without considering these uncertainties could lead to unreliable design. Re-cently, there has been much research in Reliability-Based Design Optimization (RBDO) taking into consideration both the reliability and optimization. RBDO involves the evaluation of prob- abilistic constraint that can be estimated using two different approaches: the Reliability Index Approach (RIA) and the Performance Measure Approach (PMA). RIA defines a probabilistic constraint as reliability and PMA defines a probabilistic constraint as the performance measure. It is generally known that PMA is more stable and efficient than RIA. Despite the significant advancement in PMA, RBDO still requires large computation time for large-scale applications. Therefore, in this study, an effective RBDO algorithm is presented by integrating Response Surface Method (RSM) with PMA. The RSM has been successfully applied to many application of the reliability analysis (Myers et al. 1995). It is to approximate an implicit performance func- tion or the structural responses to a polynomial in explicit form and to perform the first order re-liability method (FORM) or the Second Order Reliability Method (SORM) with the obtained explicit function. RBDO often includes large numbers of time-consuming function evaluation. This is particularly the case when finite element analysis is involved. In this sense, RSM can help the RBDO alleviate the computational burden. The key issue of the RSM is how to select the design of experiments reasonably and how to approximate the performance function to the exact model as exactly as possible. For selecting the design of experiments, Latin hypercube sampling is employed to generate uniform sample points over the region that covers the constraint in PMA. For the approximation of the performance function, the least square method is used. After constructing the response surface, PMA is used to assess the performance measure with the obtained approximate performance function. For more accuracy, several itera-tions are carried out with new sample points and pre-determined sample points. In order to demonstrate the effectiveness of the proposed method, mathematical problem and ten-bar truss are considered as numerical examples and the results are presented.

Keywords: reliabilty-based design optimization, performance measure approach, response surface method.

245 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Durability design procedure of concrete structures in Korea using partial safety factor format

J.S. Kim Department of Civil Engineering, Seokyeong Univerisity, Seoul, Korea. K.J. Shin Department of Civil Engineering, Seoul National University, Seoul, Korea J.H. Kim & K.M. Lee Department of Civil and Environmental Engineering, Sungkyunkwan University, Suwon, Korea S.H. Bae Department of Civil Engineering, Andong University, Andong, Korea

Traditionally codes of practice contain requirements for design which are formulated directly in terms of the load carrying capacity of the considered element. Durability often has been considered as being of secondary importance. Hence, requirements to the durability of the design have typically been given implicitly. Recently, a variety of researches has been carried out to conduct the durability analysis of concrete structures. To obtain a more controlled durability and long-term performance of concrete structures under chloride attacking environments, new procedures for probability-based durability analysis/design, which treats major design variables such as the diffusion coefficient, chloride corrosion threshold value, and chloride ion content at surface as random variables, have been proposed. Although there is still a lack of relevant data, this approach has been successfully applied to the durability design of some new concrete structures. In this paper, the diffusion equation of chloride ingress into reinforced concrete structures based on Fick’ssecond law was solved with a time dependent diffusion coefficient. The time-dependency was included with a new parameter, mean diffusion coefficient for interested time interval, and verified with some experimental results. The probabilistic analysis of the durability performance was carried out using a Monte Carlo Simulation with a limit state function which is defined as the violation of pre-determined chloride ion concentration at the depth of reinforcement. The durability design procedure of Korean Standard was examined in the point of structural reliability concept. As the code, adopting level-1 partial safety factor method, was not based on an explicit probabilistic approach, the design according to KS code may give non-consistent result. A durability design of real concrete structures in Korea was re-designed following various design methods such as DuraCrete method, full-probabilistic method and Korean Standard methods. The results of former two show little difference and variances but that of KS show many variances according to the values of safety factors and intended service lives. Therefore, further researches and efforts should be given to improvement of durability design methods in Korea and the probability- based design concept also should be taken by the KS code.

246 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Analysis of static behavior of CFTA girder

Hak Lee Civil, Environment and Architecture Engineering, Korea University, Seoul, Korea Kyung-Hoon Park Hybrid Structure Research Division, Korea Institute of Construction Technology, Korea Jung-Sik Kong Civil, Environment and Architecture Engineering, Korea University, Seoul, Korea

Nowadays various studies related to superstructures of bridges are conducted to pursue more effec- tive section of superstructure, materials and more economical application of composite members. Particularly the concrete, the steel and the strand are the most effective materials and widely used. Therefore, many researches are ongoing to maximize merits of traditional structure members such as CFT (Concrete Filled Steel Tube) or SCP (Steel Confinement Pre-stressed) girder. This paper introduces CFTA girder (Concrete-Filled and Tied Steel Tubular Arch Girder) which is combined with the traditional CFT structure, the arch effect and the pre-stressing effect. CFT structure is a composite style that steel box is filled with concrete, therefore deformation and stiffness of the members will be improved by confinement effect and due to the arch effect and pre-stressing effect, the induced moment and tensile stress will be reduced. In this study, linear and nonlinear static analysis were carried out on 12.2 m CFTA girder which force and deformation-controlled load are applied to and is pre-stressed to verify the performance of the CFTA girder by ABAQUS 6.5-1 and MIDAS/civil 2006. This study is carried out to verify the performance The characteristic of the innovation type of the composite girder. In order to comprehend the characteristic, ABAQUS 6.5-1 and MIDAS/civil 2006 are used. The conclusion by the FE and Frame analysis of the CFTA girder is summarized as follows: (1) When the load is applied to the CFTA girder, the stress of the strands is increased by the loads. It is judged that the increase of the stress of the strands makes the axial force of the girder reduce to apply the compression load to the girder. Therefore, the CFTA girder acquires more margin of the tensile fracture than the girder which doesn’t have the tied-arch effect. (2) By the displacement aspect of the CFTA girder, the position of the inflection point is placed within 1814.29mm from the end point of the girder. The reason why the inflection point is positioned inside of the girder is the tension force of the strands induces the negative moment to the girder, and that the increase of the positive moment offsets the negative moment at the position of the inflection point. (3) The results of the FE and the Frame analysis are similar. Therefore the frame model is verified by the FE analysis and it is judged that the frame model is adequate for the optimum design of the 50 m CFTA girder. (4) Although the nonlinear behavior is started, there is no rapid decrease of the stiffness by the non- linear analysis of the CFTA girder. Therefore if the CFTA girder is applied to the superstructure of the bridge, it has the sufficient margin of safety.

247 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Durability analysis of RC bridges using Monte-Carlo simulation

W. Raphael, R. Faddoul & J. Hokayem Saint-Joseph University, ESIB – CST Mkalles Mar Roukos– Beirut, Lebanon

Models for describing the evolution of the degradation processes of concrete and thus used to predict the lifetime of existing or future concrete structures are an essential requirement for a rational optimization of the design, execution, maintenance and the eventual decommissioning of concrete structures. Until this day, the design of concrete used in bridges, considered under the durability angle, is essentially based upon deem-to-satisfy empirical rules, such as, minimum reinforcement cover or maximum water/cement ratio. Under particular conditions specified in the design codes (related to the environment, manufacturing, material properties and loading), such an approach allows us to obtain sometimes an acceptable durability. However, the durability of concrete used in bridges, more particularly the carbonation, is a difficult problem and is often unreliable. The aim of the present paper is to develop a performance based durability design and assessment model of concrete structures. Such a model will be necessarily probabilistic, due to the stochastic nature of almost all the input variables and also to the lack of knowledge of the exact physical relations governing the phenomena participating in the deterioration of concrete. In the first part of the study, some detailed mathematical models describing the carbonation process and the chloride ingress process are adopted. The proposed models take into account the influence of the environmental conditions (temperature, humidity, pollution ...), the mate- rial properties (cement type, admixtures, type of aggregates ...), and the concrete manufacturing (water/cement ratio, casting of concrete, curing ...). Then, a probabilistic model of the input variables is defined, taking into account, for the carbonation process, the time dependant variation of RH. In the third part of the study, a response sample is generated using Monte-Carlo simulations techniques, degradation stochastic processes are inferred and used to describe the evolution of a carbonation front and of a critical chloride concentration front. Reliability indexes at different times are calculated and the number of years after which the concrete element is considered as “not sufficiently safe”, is determined for both processes and for different values of the input parameters. This study shows the interest of reliability considerations in the qualification of design codes and allows the designer to take account for carbonation models uncertainties. Very interesting results are obtained and the adoption of such a design approach would improve long-term serviceability of bridges.

248 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

System stability design of cable-stayed bridges based on elastic/inelastic system buckling analyses

Y.-S. Kyung, J.-S. Lee & M.-Y. Kim Sungkyunkwan University, Suwon, Korea

This study presents a system stability design method of cable-stayed bridges based on the elas- tic/inelastic system buckling and the system P-Delta analysis. Firstly an initial configuration of cable-stayed bridges is determined under full dead-loads. And then the elastic and geometric stiff- ness matrices are evaluated for the whole structural system and a system buckling analysis is utilized to calculate the effective buckling length of beam-column members such as main girders and pylon members. Next the second-order analysis considering P-Delta effect is performed and the amplified bending moments are evaluated under live load combinations. Finally an improved stability design method is proposed and the results are compared with those by the current stability design method through the numerical example.

Figure 1. Effective buckling lengths in main members of cable-stayed bridges.

249 Figure 2. Inelastic Procedure for the calculation of effective length.

250 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Experimental research on passive cable dampers’ performance

Sang-sup Ahn Incheon Bridge Construction Office, Korea Highway Corp., Incheon, Korea Jong-Heon Park & Sanghoon Lee GS E&C Research Institute, Yongin, Korea

ABSTRACT: The development of cable-stayed bridges in the world has been taken since the first modern cable-stayed bridge, the Stömsund Bridge, designed by German engineer Dishinger in 1955 because of the bridges’ efficient constructability and aesthetic appeal. The desire of constructing longer spans is contribute to have longer stay cables. The slender cylindrical geometry of the cables causes for experiencing undesirable vibrations during a bridge’s lifetime. In general, large-amplitude cable vibrations are known as harmful problems in cable-stayed bridges and have been reported worldwide. In order to suppress an excessive cable vibration various mechanical dampers are often selected as one of the most adequate mitigation measures. There are several different forms of passive dampers available, and each damper has a different mechanical behavior in response to cable dynamic characteristics. For the different kinds of mechanical dampers, an efficiency factor between the designed and the practical damping was studied by previous full-scale experimental research in Shanghai Tongji University. However, practical engineers require more precise behavior characteristics of passive dampers. In this study, experiments of cable-damper system were conducted by using a 100m long full-scale stay cable. 5 types of passive dampers, including high damping rubber damper, viscous damper and friction damper, were adopted. Damper performances are evaluated according to the classified damper efficiency categories such as damper efficiency, operation efficiency and total efficiency.

Cable-damper system test bench. Sensor arrangements for cable-damper system.

251 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Conformity control of concrete based on the “concrete family” concept

R. Caspeele & L. Taerwe Magnel Laboratory for Concrete Research, Department of Structural Engineering, Faculty of Engineering, Ghent University, Belgium

EXTENDED ABSTRACT: Concrete factories are asked to supply a very wide range of concrete mixes, with different characteristics. As a result of this largely differentiated request, a plant does often not produce enough of certain concrete mixes in order to apply the conformity criteria applicable to an individual concrete according to the European Standard EN 206-1. Concretes that can be reliably related to each other can be grouped into families and the combined data from the family can be used for conformity control. In order to check whether the concrete production complies with the specified properties, conformity criteria are used. For the application of the family concept, the strength results of the family members are transposed into an equivalent value of a reference concrete. This larger group of transposed data is then used to check the conformity criteria for the compressive strength. The family concept thus allows to obtain a sufficiently high number of strength results and allows a more continuous control of the production process and consequently a more rapid detection of significant changes in quality level. There exists however also a disadvantage in using the family concept. Because test results of different concrete qualities are combined, the statistical data of the individual family members interact and influence the decision making on the conformity assessment. Although EN 206-1 provides some guidelines on the use of the concrete family concept, no probabilistic appraisal of the concept is available until now. For the probabilistic evaluation of the available conformity criteria for concrete families, the operating characteristic curves are calculated and the AOQL concept is used in order to evaluate the obtained OC-curves. Some analytical formulae of the operating characteristics are mentioned for some special cases. The conformity criteria for concrete families in EN 206-1 are explained and analyzed with some Monte Carlo simulations for the special case of a concrete family with 2 members. These simulations show that the specified quality of the non-reference family member is in some cases not obtained when the family members are investigated individually. Simulations with other family compositions and other σ-values confirm the same conclusion. About the global quality of the family, it was found that the average outgoing pooled fraction defectives remains lower than 5%.

252 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Effective slab width in steel-concrete composite girder bridges

D. Bae School of Civil & Environmental Engineering, Kookmin University, Korea S.G. Youn Department of Civil Engineering, Seoul National University of Technology, Korea Y.S. Park Department of Civil Engineering, Myongji University, Korea

ABSTRACT: This paper present the results of investigating the effective width of deck-slabs in multi-girder steel-concrete composite bridges. A general solution technique is developed for analyzing these structures. A set of biharmonic equations are employed to solve a set of differential equations for plane stress and flexural displacement of the portion of deck between girders. By using these analytical solutions, a parametric study is made on the effects of several factors influencing the effective slab width of multi-girder composite structures. The factors include type and portion of load, ratio of girder spacing and span length, geometric configurations of composite girders and thickness of slab.

1 INTRODUCTION

The main objective of this study is to provide, through an extensive parametric study, a reason- able understanding of the effective slab width in multi-girder composite bridges. To study the behavior and diverse responses of multi-girder composite bridges with many different geometric factors, a method of analyzing multi-girder composite bridges with simply supported spans is developed. Based on the above solutions, a parametric study for the effective slab width of multi-girder composite structure is made. The parameters are type and position of load, deck-slab aspect ratio, configuration of the cross section and deck thickness.

2 ANALYTICAL DEVELOPMENT

A solution technique for analyzing multi-girder composite bridges which are simply supported, single span structures are developed. The method of solution in this paper will consider that the deck-slabs of multi-girder composite bridges as plates on elastic supports. The general response of a bridge deck which is elastically supported by two girders at the edges, can be expressed by two fourth-order differential equations. Eight constants of the series solutions for these two quantities may be determined by con- sidering the equilibrium and compatibility conditions between the girders and the slab at the elastically supported edges (with four boundary conditions at each edge). These differential equa- tions and solutions can be expanded to examine the behavior of deck-slabs of multi-girder composite bridges. 253 3 INFLUENCE OF PARAMETERS ON EFFECTIVE SLAB WIDTH

The effective width, Be, of six different geometrical arrangements of multi-girder systems (two, three and four girder system with and without overhanging part) are investigated by varying the several parameters, which are the type and position of load, configuration of the cross section, deck-slab aspect ratio and deck thickness. The effective widths are increased as the load moves from above the girder towards the center line between the girders. The minimum effective width occurs at mid-span for all cases. An important conclusion can be drawn that the effective slab width for two girder composite structures without overhang is smallest when the load is applied directly above one of the two girders. The smallest effective slab width for a load position always occurs under the load.

4 CONCLUSIONS

– The deck-slabs of multi-girder composite structures under uniform loads have larger effective slab width than when under concentrated loads. – The lateral or longitudinal position of loads affects the effective slabs width strongly. Generally, a load directly above a girder, or equal loads on every girder controls the minimum effective slab width for all cross sectional configurations of multi-girder composite structures. Under point loads, the minimum effective width occurs in the cross section containing the load Loads away from the mid-span produce smaller effective slab widths. – The ratio of girder spacing to span length (the aspect ratio) is the most significant parameter influencing the effective slab width. – Increasing the deck thickness reduces the effective slab width slightly, contrary to the provisions in AASHTO specification. Overall, the effective width is not sensitive to the change of deck thickness.

254 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Preference-based optimal maintenance planning for deteriorating bridges

S.Y. Lee Seoul National University, Seoul, Korea W. Park Korea Bridge Design & Engineering Research Center, Seoul, Korea H.-M. Koh & H.J. Kim Seoul National University, Seoul, Korea

Because the structural performances are of critical importance in bridge maintenance planning, many researches have shown that both the cost and the bridge performance could be treated simultaneously by the multi-objective optimization problem. However, due to the nature of the multi-objective optimization problem, a decision making process is always needed to determine an optimal maintenance planning and it is sometimes very hard to choose the optimal one from Pareto optimal solutions set of which, for example, hundred of optimal solutions are the elements. In this study, the maintenance planning of deteriorating bridges is formulated as multi-objective optimization problem that treats the maintenance cost and the condition grade of the bridge deck and girder. To effectively address the multi-objective optimization problem and decision making process for the solution set obtained, we adopt a preference-based optimization model and apply a genetic algorithm as a numerical searching technique. In preference-based optimization, the relative impor- tance between the different objective functions can be normalized and measured by introducing a preference function. Therefore, the optimal maintenance solution is derived based on the preference of the bridge manager such as the importance or the priority of the maintenance methods applied. A numerical example for a typical 5-span prestressed concrete girder bridge shows that the opti- mized results using the preference based optimization method meet most design objectives without excessively sacrificing one specific objective compared to another (Fig. 1). This implies that the proposed approach can satisfy multiple conflicting objectives and therefore, obtain an optimal main- tenance solution that has desirable minimum condition grade of deck, girder and pier with balanced maintenance cost based on the manager’s preference. The optimal sequence and schedule of main- tenance actions such as repair, reinforcement and rebuild are determined by the proposed method. To model the deterioration process of the bridges, actual condition grade data in Korea is used.

Figure 1. Optimized values and preference range of design objectives.

255 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Analysis of steel-soil bridge structure made of corrugated plate

D. Beben Civil Engineering Department, Opole University of Technology, Opole, Poland Z. Manko Institute of Civil Engineering, Wroclaw University of Technology, Wroclaw, Poland

ABSTRACT: The way in which a new road bridge made from corrugated steel plates was tested is described and the test results, for three static load schemes in which one ballasting vehicle was used as the load, are presented. The bridge has an effective span of 12.315 m and a clear height of 3.555 m. The steel shell of the span is founded on two reinforced concrete continuous foundations. The average measured displacements and unit strains (normal stresses) in selected points and elements of the steel shell structure were found to be much smaller than the ones calculated for the same load. The conclusions drawn from this research can be useful for assessing the behaviour of such steel shells and their interaction with the surrounding backfill. Since such steel-soil structures are used more and more often for small and medium-sized bridges on road and railway lines in Poland and in the world, the conclusions from the static load tests can be generalized to a whole class of similar bridge structures.

1 INTRODUCTION

The way in which a single-span, flexible-structure bridge made from corrugated steel plates was tested is described, and the results of the tests and FDM calculations that serve as the basis for determining the quality and durability of the bridge and accepted it for normal service in view of its quite large effective span and prototypical character are presented. The primary aim of the static load tests was to determine the effort of the structural components of the bridge and to assess the workmanship and the performance of the shell structure under a known load in order to verify the assumptions made in the static calculations and analyses of the span and in the test load program and to determine the actual load-carrying capacity of the bridge. In particular the actual rigidity of the corrugated plates in the arch structure was to be evaluated and the width of the corrugated plate interact with the soil in carrying service loads and the transverse distribution of the loads among the individual ribs were to be determined. The paper presents field tests and calculations of road bridge which the reinforcement of shell was made from corrugated steel plates with an application of new localization and relatively small height cover of soil over the shell in comparison to different bridge structures of this type. The tests were to be carried out in the full range of static loads and they were to include measure- ments of deflections and strains at selected points of the steel shell structure in three cross-sections along the length of the span: in the crown – cross-section I–I, at the end of the reinforcement (cross-section II–II), and in the haunch (cross-section III–III). As originally planned, three load schemes, i.e., 2 asymmetric load schemes (the truck positioned at the barrier on the headwater side – scheme I or on the tailwater side – scheme III) and 1 symmetric load scheme (the truck positioned on the longitudinal axis of the roadway in such a way that its rear axle was located in the middle of the effective span length – scheme II were used). 256 Based on the practical experience gained from the tests and FDM calculations, the observations concerning the behaviour of the shell structure in this type of bridge made during the tests and a comprehensive analysis of the measurement and computation results, the following general conclusions about the actual behaviour of the bridge can be drawn.

257 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Introduction of design for concrete filled steel tubular arch girder with external tendons

Eunho Lee, Ho Park & Myoung Gyun Park Sambo Engineering Co., Ltd., Songpa-Gu, Seoul, Korea Kyung Hoon Park, Sang Yoon Lee & Jung Ho Kim Korea Institute of Construction Technology, Goyang-Si, Korea

There have been many trials to enhance material performances by maximizing the superior charac- teristics of each material or by complementing each other for the demerits. This paper introduces the concept and design method of the newly-developed girder, Concrete Filled Steel Tubular Arch girder (hereinafter CFTA girder), which structural efficiency is promoted by its arch configuration and additional prestressing force (Fig. 1). Figure 1 shows the composition system of CFTA girder bridge conceptually which is constituted by steel surface, concrete arch, PS tendons, and concrete deck components. In order to transfer the loads of inner arch rib concrete and self-weight to arch rib after manufacturing steel surface of girder with arch shape, column concrete is filled so that the composition of steel and concrete is completed with empty space in the inside of steel surface. In addition, the prestressing tendons which are located in the lower part of composite girder ends is installed to control self-weight loads and live loads effectively. Therefore, comparing to the existing girder bridge, the benefits of CFTA girder bridge in terms of technical, economical and aesthetic aspects can be derived by providing the structural efficiency of arch and the effective combination of merits of each structural material such as concrete, steel, tendons, and so forth. In this study, the design method of composite struc- ture is established for the purpose of designing CFTA girder and the investigation on the efficient construction procedure also has been performed. In addition, the small scale specimen test has been performed to verify the structural efficiency of CFTA girder so that the difference between structural analysis and actual test is investigated.

Figure 1. Concept diagram of CFTA girder.

258 As the result of this study, it is expected that the practical use of CFTA girder bridge through the further research and development can lower the cost of construction and maintenance as a role of mid-long span bridge. In addition, CFTA girder bridge can be applied to highway bridge and foot over bridge as well so that the presented CFTA girder bridge can be a help to the growth of design and construction technique in Korea.

259

Design expectations, monitoring response and maintenance decisions

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Design expectations and monitoring response of the Certosa fly-over in Milan (Italy)

P.G. Malerba Department of Structural Engineering, Politecnico di Milano, Milan, Italy A. Giussani IC&T, Politecnico di Milano, Milan, Italy G. Pezzetti Field S.r.l., Lallio (BG), Italy C. Malerba Consultant Engineer, Milan, Italy

The fly-over on the Milan Certosa railway-junction, built between 1987 and 1989, represents a particular application of a cable-stayed bridge into a urban viability context. Design specifications were exacting and required: (a) a section with a minimum depth; (b) maximum reduction of any interferences with the rail traffic during all erection stages; (c) a preliminary evaluation of the costs caused by the slowing down of the trains, required by the State Railways. After several design strategies were examined, the choice fell an a solution based on a cable-stayed bridge with spans of 45–90–45 m and with a deck depth of 1.60 m. The erection procedure was unusual: firstly, the two halves of the structure were manufactured outside the railway area, and then they were juxtaposed. Once the structure was in place, the two halves were made continuous by means of a connection casting and of pre-stressing cables. Such an erection technique allowed to minimize the interferences with the rail traffic and allowed an autonomous organization of the construction for almost all the structure. The particular geometry of this bridge, having relative proportions quite different with respect to the usual scheme of other cable stayed bridges, strongly characterized also the structural design. A particular attention was paid to compute the pretensioning of the stay cables. Due to the time dependent behaviour of the concrete deck and of the concrete towers, a pretensioning level suitable to remain stable with time was sought. After about fifteen years of service, the Milan Municipality decided for a general and detailed inspection of the bridge. Together with the inspections, surveying and monitoring activities were started. The main data achieved until now regard the relative movements of the bearing supports, the horizontal displacements at the top of the pylons, the vertical displacements along the two sides of the deck and the tension force in the cables. Temperature was registered nearby each instrument. The movements of the bearing supports follow with regularity the cyclic seasonal variation of the temperature. Their relative displacements are close to design estimations and, in this, reveal a regular behaviour of the bridge as a whole. The horizontal movements of the top of the pylons and the vertical displacements along the deck are comprised into the band associated with thermal range. Such displacements can be considered as physiological movements of the structure, without any particular correlation between causes and effects. The tension force in stays, after fifteen years, appear substantially stable in time, with a little trend to increase the forces in the external cables and to de-tension the internal ones. Globally, one can say that design expectations result fulfilled, with regards to both the static behaviour and durability performance.

263 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Statistical monitoring of concrete structures and cable-stayed bridges

Alessandro Fassò Dept. IGI, Università di Bergamo, Italy Giorgio Pezzetti FIELD Srl, Lallio BG, Italy

Modern monitoring of complex structures such as a cable-stayed bridge includes a number of real time operating instruments including jointmeters, load cells, inclinometers and thermometers. Assessment of field uncertainty and adjustment for environmental factors has been considered by Fassò et al. (2004 and 2005) for this kind of monitoring systems. In order to detect slow changes, such as foundation settlements or other structural damages due to diffusive attacks from external aggressive agents, see e.g. Biondini et al. (2006a), we consider a statistical approach allowing for the assessment of the monitoring system uncer- tainty and the physiological or normal dynamics of the structure at hand. The modelling approach allows for block segmenting the whole structure in an appropriate number of simpler sub-structures. Correspondingly, we define a hierarchical change detection which allows as to have a top level single statistics which is time-varying and gives a signal when the overall structure health shifts from the expected behaviour. After a signal is given, a battery of lower level statistics, related to the various model blocks defined above, allows us to localize the most probable source of drift. In this paper, we consider in details a case study related to Certosa Bridge in Milan, which illustrates the performance of the surveillance system on the basis of some artificial anomalies applied to Certosa bridge data. These data are interesting because, after a structural rehabil- itation, whose lifetime impact has been discussed in details by Biondini et al. (2006b), a number of new instruments have been installed including jointmeters (JM), deflection meters, clinometers (CL), and thermometers, which complement the data of the previously installed Load Cells (LC). As a result, the proposed methodology is shown to be capable of assessing field uncertainty and compensating for environmental biases. Moreover, it is shown that modeling the correlations among measurements improves the efficiency in supervising the monitored system and in localizing the sources of anomalies with a prefixed rate of false detections, especially for instruments strongly related to temperature.

REFERENCES

Biondini, F. Bontempi, F. Frangopol, D.M. Malerba, P.G. 2006a. Lifetime nonlinear analysis of concrete structures under uncertainty. In Cruz, Frangopol and Neves (eds) (2006) Bridge, Maintenance, Safety and Management, Taylor and Francis. Proceedings of IABMAS’06 – Third International Conference, July 16–19, 2006, Porto-Portugal. Biondini, F. Frangopol, D.M. Malerba, P.G. 2006b. Time-variant Performance of the Certosa Cable-stayed Bridge. Structural Engineering International. 16(3): 235–244. FassòA., Nicolis O., Bruzzi D., Pezzetti G. 2004. Statistical modelling and uncertainty reduction of monitoring data in geomechanics, Working paper n.3/MS, Dept. IGI, University of Bergamo (www.unibg.it) Fassò A., Nicolis O., Bruzzi D., Pezzetti G. 2005. Modelling and reducing uncertainty of field monitoring data in geomechanics by computerized statistical methods. In Proceeding of 11th IACMAG Conference, Torino

264 19-24 June 2005. Vol.1, 595–602. Edited by Barla G. and Barla M. (2005) Prediction, analysis and design in geomechanical applications. Pàtron, Bologna. Fassò, A., Pezzetti, G. 2007. Statistical Methods for Monitoring Data Analysis. In DiMaggio, J. & Osborn, P. (Eds) FMGM 2007: Proceedings of the 7th International Symposium on Field Measurements in Geomechanics, ASCE.

265 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Short and long term monitoring for maintenance and retrofitting of existing bridges

C. Modena, P. Franchetti, M. Grendene & M. Frizzarin University of Padova, Padova, Italy

In the developed countries situations become more common in which the public road network authorities and owner’s expenditure must be more targeted not only for routine maintenance but also for rehabilitation, refurbishment and retrofitting interventions. In particular, focusing on bridges, at present only in specific cases, increasing attention is paid to all the activities aimed at maintaining an acceptable level of their structural safety over time: it is noted that suitable interventions are essential for the reduction of costs, lengthening their structural life and delaying their complete restoration. On the other hand, due to the intrinsic weakness of some structural components, to deterio- ration phenomena and to the updating of structural codes, many existing bridges show currently inadequate structural performances. Their structural rehabilitation is usually coupled with refur- bishment interventions aimed to improve the safety and comfort of road users and static and dynamic retrofitting. Their design, choosing the proper intervention in terms of both material and application techniques, require a very complex approach, starting from the assessment of the current and the prognosis of future structural performances by means of an appropriate campaign of experimental and theoretical investigations. In this paper, reporting practical applications, the problems related to the “structural” – static and dynamic – characterization of existing bridges and to the observation and evaluation of the most relevant physical parameters sensitive to faults or damages are discussed. A refurbishment and seismic retrofit procedure was carried out in an arch r.c. bridge over Piave river, located in Sedico near Belluno, Italy. After a static rehabilitation and refurbishment, a model updating procedure was developed, in order to calibrate the model on the basis of experimental dynamic data. The parameters that describe the modal model of the structure are obtained from measurements of mechanical vibrations due to the excitation of an harmonic force (shaker), and from the analysis in the frequency domain of the more significant signals acquired. The results of the analysis carried out on both the preliminary and the identified FE model, highlight the deficiency of the structural system against the horizontal-transversal actions. Therefore, the main purpose of the seismic improvement interventions is the protection of the arches against the horizontal actions. Therefore the intervention consists in increasing the safety factor against the overturning and stiffening and strengthening the transversal braces, so as to avoid a strengthening intervention on the arches themselves. In this context, two kinds of interventions are foreseen: – a strengthening of the lateral braces above each of the piers, on the sides of the spans; – the consolidation and the improvement of the link between piers and foundation caissons. At the end of the works, dynamic monitoring and identification will give information on the overall structural behavior, allowing, via model updating procedures, for controlling the efficiency of the applied interventions.

266 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

The structural analysis of the Messina Strait Bridge

F. Bontempi University of Rome “La Sapienza”, Rome, Italy

The purpose of this paper is to present in a synthetic form aspects of the structural analysis of the Messina Strait Bridge that were developed during the years from 2002 to 2005 to support the basis of the design and the definition of the performances levels for the Messina Strait Bridge. The essential role of the structural analysis supporting the decisional process is enlightened referring to the fundamental static behavior assessment of the bridge connected to serviceability and ultimate scenarios.

267 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Influence of large displacements on the structural stability of cable supported bridges

F. Biondini, P. Limonta, P.G. Malerba & R. Stucchi Department of Structural Engineering, Politecnico di Milano, Milan, Italy

The paper intends to investigate the role of large displacements on the stability of the pylon of cable supported bridges. As known, the behavior of long span supported bridge is strongly influenced by non linear geometrical effects. Live loads, asymmetrically distributed on the bridge span, make the pylon bend and maybe twist. The deformed configuration of the pylon in the vertical plane makes vertical loads eccentrics and produces bending and twisting moments. The principal purpose is to evaluate the state of stress of the pylon due to these actions in relation to the geometric characteristics of the section and taking into account the role of large strains and displacements. In this work a numerical model suitable to deal with large displacements and large strains and provided with numerical techniques for solving non linear problems in presence of singular points and of unstable behaviours has been developed. Comparative studies examine the influence of the geometrical characteristics of the pylon and of the imposed imperfections, both in the hypothesis of small displacements and of large displacements and strains. The influence of geometrical nonlin- earities is shown through a comparative application. In the first case the pylon is studied by means of an Eulerian analysis, while in the second case a Lagrangian approach is adopted. When studying the problem by means of an Eulerian analysis, it is interesting to note that if a pylon is characterized by a great stiffness, the parameter α h is small and the effective bending moment is relatively large. To prevent this situation, it is necessary to reduce the pylon stiffness. Designing the pylon with α h in this range, means also that the maximum bending moment is not at the base, but is progressively rising towards the middle of the pylon. Beyond this interval the pylon is too thin and the normal force approaches the second Euler load. It can be noted that the same behavior is obtained for any intensity of the horizontal imperfection. The same design criterion for the dimensioning of the pylon by limiting the design parameter is valid also in Lagrangian approach, however in this case more information is available. In fact it can be noted that small imperfections lead to the same behavior obtained with the Eulerian approach, while, as the horizontal force increases, different solutions can be found. Therefore, a distinction can be introduced according to the angle between the pylon and the equivalent stay. When this angle assumes very small values the Eulerian analysis is sufficient to define the range within it is possible to choose the pylon stiffness. For large values of the angle the design must be revised because it means that great displacements will occur, while for intermediate values the position of the maximum bending moment must be evaluated carefully.

REFERENCES

Biondini, F., Di Domizio, M, Malerba, P. G., On the Structural Analysis of Cable-Stayed and Suspension Bridges, 2nd Int. Conf. on Bridge Maintenance, Safety and Management (IABMAS’04), Kyoto, October, 19–22.2004. Bontempi, F., Malerba, P. G., 1997. The Role of Softening in the Numerical Nonliner Analysis of Reinforced Concrete Frames, Structural Engineering and Mechanics, 5(6), 785–801. Gimsing, N., J., 1988. Cable supported bridges, John Wiley & Sons.

268 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Wind-induced fatigue assessment in main cables and hangers of suspension bridges

F. Ubertini University of Pavia, Pavia, Italy F. Bontempi University of Rome “La Sapienza”, Rome, Italy

Wind action on long-span suspension bridges may produce large amplitude vibrations, eventually leading to fatigue ruptures. In order to obtain reliable results, one should appropriately model the stochastic wind field, and the wind forces acting on the bridge deck. Regarding this last point, frequency domain methods become inadequate due to the inherent geometric nonlinearity of the structure and time domain methods must be preferred. Given these considerations, the paper explores the fatigue damage produced in the main cables and in the hangers of a long span suspension bridge when subjected to wind excitation. A three-dimensional nonlinear simplified FEM model of the bridge is developed, at first, and validated on the basis of a more costly FEM model and of the experimental results available in the literature. The wind field is modeled as a multivariate Gaussian stochastic process, exploiting the properties of the Proper Orthogonal Decomposition (P.O.D.) of the spectral matrix. The theory of aeroelastic derivatives in the time domain and the quasi-steady formulation are adopted to model the wind forces on the bridge deck. Time domain analysis are performed in a nonlinear framework and the fatigue damage is calculated, according to Palmgren-Miner’s rule, including bending effects near the terminations. A relation between fatigue damage and mean wind velocity at the bridge mid-span represents the final result of the paper. All the other excitation parameters (spectra, turbulence intensities, etc.) are fixed on the basis of literature results.

269

Development of the advanced robot systems for bridge inspection and monitoring

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Introduction of the Bridge Inspection Robot Development Interface (BIRDI)

J.S. Lee, I. Hwang, H.S. Lee & S.H. Hong Hanyang University, Ansan, Korea

Entering into the technology-driven 21st century, we are challenged to make creative changes and progressive innovations in every aspect of our engineering endeavor in order to maintain our competitiveness. Emerging technologies such as Information Technology (IT) and Robot Tech- nology (RT) provide us with opportunities through innovative technology fusion for upgrading out construction technologies and practices. The Bridge Inspection Robot Development Interface (BIRDI) was launched in December 2005 as a new R&D initiative in the Program for Cutting-Edge Technology Fusion for Construction under the sponsorship of MOCT (Ministry of Construction & Transportation) and KICTEP (Korea Institute of Construction & Transportation Technology Eval- uation and Planning). BIRDI is a research consortium whose primary goal is to develop advanced robot systems for automated bridge inspection and monitoring for immediate field application. The developed robot system will allow remote inspection of bridges while reducing human risks and improving efficiency and data reliability, thereby enabling more rational bridge maintenance pro- cedure. This paper describes briefly the robot system under development and other related research activities ongoing at BIRDI.

273 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Robotic diagnosis system for detection of bridge structures

Dae-Joong Moon Engineering Research Center, EJ Tech Engineering, Korea Kyung-Taek Yang Department of Mechatronic Engineering, Daelim College of Tech, Korea Soon-Sung Nam & Kwan-Hyuk Im Engineering Research Center, EJ Tech Engineering, Korea

Bridge Maintenance operation and the appearance investigation are necessary to be carried out because that structural performance of bridge was reduced by the vehicles load and wind load under using condition. However, in a wide inspection area, it is given rise to a certain inspection error and omission possibility due to difficult approach and weak safety. Traditional method causes to a lot of problems on economic and utility parts and it becomes a technical obstacle factor on bridge appearance inspection. Automatic diagnosis technology is efficiently established to the calculation of data defect, satis- fied safety for the inspector and reliable data base concerning the appearance inspection drawing. Automatic diagnosis technology is composed of the machine vision system and linear motion control system. The Machine Vision System consists of pan/tilt drives, zoom lens, CCD camera, controller, and calculation program of the crack width/length. The electric signals from the sensors on the focus and zoom lens are processed through micro-processing technology to calculate with the severity cracks length and width on structures. In this study, two more cameras are added to improve performance and to enhance efficiency when measuring width and length of cracks. Linear Motion Control System holds the Machine Vision System and moves perpendicular to the longitudinal axis of the bridge for inspection. The controlled camera system is installed on the mechanism to transmit video images through wired or wireless communication for inspectors to read them on their computers. The visual image data is connected to the existing Bridge Management System (BMS) and combined into a single unit of data for higher efficiency. Then, users can scroll down to enter information on cracks or to convert it to a CAD file to open on the existing BMS. As the existing BMS has variable configurations and input/output formats, the new program improves compatibility with BMS by allowing it to save image files in Auto-CAD format.

274 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Development of bridge inspection robot system: Wall climbing robot and flying robot

Ig Mo Koo & Chang Min Lee School of Mechanical Engineering, Sungkyunkwan University, Suwon, Korea Se-Hee Whang, Duk-Hoo Kim, Min-Sung Kang, Kuk Cho, Woong-Hee Son & Sangdeok Park Korea Institute of Industrial Technology (KITECH), Korea Sun Kyu Park School of Civil Engineering, Sungkyunkwan University, Suwon, Korea Hyouk Ryeol Choi School of Mechanical Engineering, Sungkyunkwan University, Suwon, Korea

Periodical inspection and regular health monitoring are very important to secure safety and extend the life of bridges. Up to now, various destructive and nondestructive bridge inspection approaches with new equipments have been developed. However, most of current bridge inspection methods, especially bridge appearance inspection methods, require massive assistance facilities and can not guarantee inspectors’ safety. Hence, it is highly required to develop easy, fast, safe, accurate and automatic bridge inspection methods. In this paper a flying robot system is introduced, which is combined with an UAV(UnmannedAerialVehicles)and a mobile robot system for visual inspection of bridges. The UAV has two coaxial rotors to fly and the mobile robot system has omni-directional wheels to navigate the lower surface of the bridges. The robot is attached to the lower surface of a bridge using the technology of UAV,navigates the space autonomously using the mobile platform under local position recognizable condition with ultrasonic localization system, and sends camera images of the surface through wireless communication. Also, we introduce a wall climbing robot on the way of development in this paper. The robot is capable of quick moving on various wall surfaces with an impeller adhesion mechanism. The adhesion mechanism which is composed of an impeller and flexible suction seal provides sufficient adhesion forces for supporting its body on the surface by keeping the air pressure less than the ambient one. Locomotion mechanism to obtain a smart moving robot is also described and evaluated the process of the design concept. Robot’smanual operating abilities are tested and its performances are validated.

Keywords: bridge inspection robot, wall climbing robot, flying robot

275 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Bridge inspection robot system with novel image processing

Je-Keun Oh, An-Yong Lee, Se Min Oh, Youngjin Choi, Byung-Ju Yi & Hai WonYang Department of Electronic, Electrical, Control and Instrumentation Engineering, Hanyang University, Ansan-Si, Gyeonggi-Do, Korea Jeong Ho Lee & Young Shik Moon Department of Computer Science and Engineering, Hanyang University, Ansan-Si, Gyeonggi-Do, Korea

This paper proposes the design and control methods for a bridge inspection robot system with the novel image processing system. The bridge inspection robot system has been developed with the aim of checking the safety status of a real bridge and gathering accurate data such as crack width and length. The developed robot system is composed of the moving mechanism mounted on the specially designed car for bridge inspection and the novel image processing system. Especially, this paper emphasizes on the system integration method including the design and control of the entire robot system. Finally, we suggest the experimental results to show the effectiveness and robustness of the developed bridge inspection robot system.

276 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Intelligent bridge management system based on the image data from robotic devices

Sungkon Kim Department of Civil Engineering, Seoul National University of Technology, Seoul, Korea Jung Seok Lee Korea Infrastructure Safety and Technology Corp., Ilsan, Korea Youngjin Choi Department of Electronic, Electrical, Control and Instrumentation Engineering, Hanyang University, Ansan, Korea Young Shik Moon Department of Computer Science and Engineering, Hanyang University, Ansan-Si, Gyeonggi-Do, Korea

This paper addresses features of a Bridge Management System which is a computerized total infor- mation system relating to inventory data management, bridge inspection, condition assessment, repair, scheduling, and budgeting. A BMS consists of software and hardware system. Software system covers database software for managing all bride maintenance related items, and applica- tion software. The database system manages overall physical description, structural and inspection reports, repair and strengthening history, assessment results. The application program consists of software engineered overall bridge maintenance work program, input and output program for database, assessment and analysis program to convert stored data to usable information, and report- ing program. The heart of the program is the assessment and analysis program that covert individual items stored in the database to useful information for the bridge managers. The core part of BMS operation will be on the data structure in which all the information from the field inspection is well archived, so that the system could perform the bridge assessment tasks. Conventional systems have been relied on the hand writing data for the damage aspects by field engineers. In this paper, an advanced technique is introduced to acquire damage aspects of the bridge members from the image data which are taken by the robotic systems.

277 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Optimum NDT using infrared thermography for defected concrete

Goangseup Zi & Jun Gi Sim Department of Civil and Environmental Engineering, Korea University, Korea Hongseob Oh Department of Civil, Jinju National University, Korea Jongseh Lee Department of Civil and Environmental Engineering, Hanyang University, Korea

In this paper, optimal conditions for the use of infrared thermography technique to detect defects in concretes were studied experimentally.Two different kinds of defects were considered: voids located few centimeters below the surface and debonds between concrete surface and FRP sheet which may be used to reinforce the concrete. Those specimens were heated by three different heating sources including natural light, infrared lamp and halogen lamp, before applying the thermography to detect the defects. The optimum heating time and the sensitivity are given based on the experimental.

278 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

A development of repair mechanism and control technologies for bottom part of concrete bridge

Kye-Young Lim Korea Polytechnic University, Korea

Inspection technologies of the bottom part of bridge are under development in colleague teams BIRDI (Bridge Inspection Robot Development Interface). The main purpose of this inspection is to find some existence of damaged part, or to find possible on going rust which is depending on crack, leakage, or some other reasons. The goal of this project is developing a mechanism and corresponding control technologies to repair the crack of bottom part of concrete bridge. Due to the special environment, several difficulties exist for this development. One difficulty is the gravity because the workspace on the bottom part of bridge on the base which is not firm is just like that of slowly vibrate ceiling on soft ground. The vacuum technology is applied to get supporting force on the surface at robot frame. Every four part of suctions are consist of multiple pads, and the vacuum level of pads are controlled according to the leaks of air on the rough surface to get hanging force. The injection of liquid epoxy will be applied into the crack with low pressure and osmotic action. The first procedure is that seal the most of cracking trajectory except injection hole area to prevent the leak out of liquid epoxy during injection. The second procedure is that the robot mounts the injector on the hole, and to do the injection start for a while. Finally the injectors are removed, and the final work is done for the clean surface. Many control technologies are developed, and tested to do the whole procedure for injection. One of that technology is force control. The well known PID controller is applied with force feedback, and the output of this controller is used as tracking reference of position command of robot. A robot system at bottom part of concrete bridge is introduced, and main technologies arising this mechanism those are vacuum handling and force control are considered in this paper.

279

Fatigue analysis

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Evaluation of fatigue strength of one riveted historical railway bridge

A. Pipinato, C. Pellegrino & C. Modena Department of Construction and Transportations, University of Padova, Italy

ABSTRACT: As a consequence of the lack of new riveted steel bridges, built in the last decades, less attention has been paid for fatigue strength behaviour than to structures containing contempo- rary fastening elements such as bolts or welds. But, the behaviour of riveted members is a matter of considerable economic importance to owners and regulatory authorities. A series of diagnostic tests were performed on one short-span, two-lane, historical railway steel-girder bridge, in service from 1918, in the line Conegliano-Casarsa, in Italy . In the first phase, were performed physical and physicochemical tests to characterize the component material. Then strains and stress were calculated via FEM analysis, to preview the ultimate strength behavior of the bridge: these results were compared with a literatures review done, and with technical Italian normative of the National Railway society (FS-RFI, partner of the investigation). In the second phase, were done real scale tests on one half part of the bridge, dismantled, and recovered in the University laboratory: static and fatigue load tests were performed to the beam, subjected to a concentrated load in various position. The first results here presented, highline the good condition of the members also after a very long service life period – nearly 100 years.

283 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Global-local finite element analysis of riveted railway bridge connections for fatigue evaluation

B.M. Imam, T.D. Righiniotis & M.K. Chryssanthopoulos Faculty of Engineering and Physical Sciences, University of Surrey, UK

In the past few years, fatigue of riveted railway bridge connections has received considerable attention. A number of damaged cases have been reported for stringer-to-cross-girder connections both in railway and highway bridges confirming that these details are susceptible to fatigue cracking. Fatigue damage in these connections is caused by deformation induced secondary stresses which were not taken into account during the design of these bridges. This is because these secondary effects are, even today, often very difficult to analyse, requiring sophisticated models. Estimating the fatigue damage in these connections through current assessment codes, in which a single stress descriptor i.e. the nominal stress is traditionally used, is not a straightforward process. Due to the very large number of such connections over the entire UK bridge network as well the connections’ age, there is an urgency to develop novel means of reliably assessing their remaining life. In this paper, a recently developed method, called theTheory of Critical Distances (TCD), is used in order to obtain fatigue damage and remaining life estimates for a critical stringer-to-cross-girder connection. The TCD, rather than relying on a single, remotely applied stress or the stress at the notch, considers the entire stress field ahead of the notch. For the purposes of this investigation, a previously developed and benchmarked detailed FE model of the bridge and the connection is analysed under the passage of the BS 5400 medium traffic trains. Histories of the average, over a volume, principal stress are obtained at various critical locations within the connection. By using the TCD in conjunction with the plain (un-notched) material S-N curve and Miner’s rule, the different components of the connection are ranked according to their estimated total fatigue damage. These damage estimates obtained employing this novel approach are then compared with their more traditional, detail-specific S-N counterpart. Mesh convergence studies are carried out and it is shown that reasonable convergence in fatigue damage is achieved through the use of the TCD. The most critical regions of the connection, where fatigue cracking may be expected are identified as the bottom hole on the stringer leg of the angle, the bottom rivet on the cross-girder leg of the angle and, for high rivet clamping forces, the angle fillet. The results show that the traditional approach tends to underestimate the fatigue damage of the critical regions of the connection, by a maximum factor of 3.5 in the case of low rivet clamping force. These differences for the critical regions were found to decrease for higher clamping force. Although the TCD requires considerable post-processing, especially for three-dimensional bridge applications, its reliance on straightforward material characterisation and its convergent characteristics make it appealing for the fatigue assessment of complex bridge details, such as riveted stringer-to-cross-girder connections. Once the application of this method is validated for bridges, refined fatigue assessment will be possible without the need for extensive testing of different connection details.

284 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Fatigue of riveted metal structures

T. Larsson & O. Lagerqvist Luleå University of Technology, Luleå, Sweden

ABSTRACT: Much have happened the last century concerning loads and speeds on the railroads. Many of the riveted bridges built in the end of the 18th century and the beginning of the 19th century are still in service. Inspections of these bridges show that they have capacity left both concerning fatigue life and the ability to carry heavier loads. Bridge owners have started to show increased interest in their old bridges due to the money that can be saved in keeping these bridges in service. To be able to this it requires better predictions of the remaining life and to make accurate service plans that maintain the safety. Though fatigue is one of the most common causes of failure in metallic structures, this work have focused on determining a detail category to use in evaluation of riveted girders. A similar approach to the one in use in prEN 1993-1-9 to determine the detail categories for welded structures have been applied on riveted girders. The determination of a detail category is derived as the stress range corresponding to 2 × 106 cycles calculated for a 75% confidence level of 95% probability of survival. A distinction where made between regular plate girders and truss girders with built up webs consisting of bars. From the evaluation of their fatigue performance it could be determined that plate girders had a fatigue endurance equal to detail category C 71. However a lower performance where found for truss girders for which the endurance where better described by the detail category C 63. Also an investigation concerning influencing parameters of the fatigue endures where per- formed. The enquiry included parameters such as, material properties, cut off limit, corrosion, hole preparation technique, and clamping force. The mayor findings in these investigations where that no mayor differences where found con- cerning the fatigue endurance between steel and cast iron, evaluation of the preparation technique for producing rivet holes came to the same conclusion, no major difference could be establish. Specimens with holes had lower fatigue endurance than specimens with the holes filled with rivets or bolts. The amount of clamping force obtained by rivets can not be compared to the one obtained by high strength bolts, but still it seems to be enough to increase the fatigue endurance. Investigation of the cut off limit for plate girders indicates that it can be raised from 28.7 MPa to 40 MPa for detail category C 71 if there is no sever corrosion or damage present on the structural elements. The same extrapolation could not be done for truss girders.

285 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Managing fatigue susceptible details on critical railway bridges at CN

J.A. Cavaco Canadian National

ABSTRACT: Concerns about fatigue failures on CN railway bridge structures was first seriously addressed during the late 1980s under the direction of Dr. Robert Sweeney, Bridge Assessment Engineer, as a result of extensive research conducted by Dr. John Fisher at Lehigh University’s, Advanced Technology for Large Structural Systems (ATLSS), Engineering Research Center. During the late 1970’sand on into the early 1980’s,fatigue life studies were conducted for several major railroad bridges located throughout the CN System. Extensive research was undertaken to acquire the most accurate as possible records of past traffic data and a detailed analysis conducted to arrive at an estimated fatigue life expectancy of critical fatigue susceptible details on each of those bridge structures. Since the production of these reports, management at CN has taken constructive steps to introduce measures to mitigate the effects of fatigue distress and ensure the safety of operations over its many aging steel bridge structures. More than 20 years after the reports were first produced, CN’s bridges continued to experience increasing stress with the introduction of heavier and more efficient use of rolling stock, greater tonnage and higher operating speeds. With the realization that potentially, catastrophic bridge failures due to fatigue distress could arise, specific measures to mitigate these effects were identified and undertaken. Besides develop- ing an in-house strain gauge testing program and specific inspection practices over the last 18 years, acoustic emission monitoring was utilized on a considerable number of bridges and continues as an effective and reliable method of checking against critical fatigue prone details. Recently, investigation into developments in Ultrasonic Impact Treatment (UIT) has prompted the use of this promising technology on fatigue susceptible welded details on CN bridge structures. UIT, a Russian developed technology has recently been introduced in North America for use in post weld residual stress treatment and fatigue strength enhancement of welded connections. UIT was first applied to one of CN’s larger bridges in 2002 with the expectation that this treatment would considerably extend its fatigue life expectations. Using acoustic emission testing to annu- ally monitor the effectiveness of treated UIT welded fatigue details has provided confidence and expectations that this treatment will enhance the fatigue life of many bridge structures that would otherwise require premature replacement. Being at the forefront in the exploration and deployment of new and developing technologies and the acquisition of associated data, CN has developed the knowledge base with which to apply risk management strategies to effectively deal with the potential threat of fatigue damage to affected steel railway bridge spans. This paper reviews fatigue related issues afflicting CN steel bridge structures and the measures taken over the last twenty years to successfully address these concerns and ensure the safe movement of trains throughout its system network.

Keywords: fatigue cracks, bridge testing, acoustic emission, ultrasonic impact treatment.

286 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Influence of the fatigue resistance of duplex steel on the bridge design – A span length investigation

T. Rauert, B. Hoffmeister & A. Gieseking Institute of Steel Construction, RWTH Aachen, Germany O. Hechler ArcelorMittal Commercial Sections, Luxembourg

Duplex stainless steels are characterised by their high mechanical and good corrosion resis- tance properties. Therefore they represent a technically and economically attractive alternative to coated carbon steels in terms of bridge construction in aggressive environment. One aim of the research project Bridgeplex (Application of Duplex stainless steel for welded bridge construction in aggressive environment, 2004–2007, contract-No. RFCS-CR-04040), supported by the European Research Fund for Coal and Steel, was to evaluate the ability of the Duplex steel EN 1.4462 to be used in an economic way for this application. Due to the high material strength of Duplex steels a higher structural resistance is given in comparison to standard carbon steels. Hence less amount of material is needed to realise similar span lengths. However, to exploit this benefit of Duplex steel a sufficient fatigue resistance is needed. In the context of the Bridgeplex project the fatigue resistance of the most critical details met in welded bridges has been investigated. The results of the fatigue tests, which are shortly described within the paper, indicate a better or a similar fatigue behaviour of Duplex steel compared to C-steels. For bridges with short spans the fatigue verification normally governs the design of the main structural members. Regarding the favourable results of the fatigue tests, such structures can also be constructed in a proven manner using Duplex steel. Bridges with large spans are characterised by high resistant cross-sections, where the design of the main structural members is governed by stress verification. For the case that the spans are larger than 45 m and the notch categories of the fatigue details are at least 71 N/mm2, no fatigue verification is required for S 235 and S 355. To prove, if such a simplified fatigue verification is also appropriate in case of Duplex steel, a span length investigation has been carried out presented within the paper. The span length for which the fatigue verification becomes negligible is called Lcrit (see Figure 1). Considering a typical bridge cross-section in Duplex steel utilisation levels for critical details, respectively critical span lengths, have been estimated considering span lengths from 20 to 80 m. It is shown that the high structural resistance of Duplex steel can be exploited to realize large spans with less amount of material.

Figure 1. Influence of the critical span length Lcrit on the bridge design.

287 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Fabrication procedure effects on fatigue resistance of steel orthotropic deck welds

H.B. Sim & C.M. Uang University of California, San Diego, La Jolla, CA, USA C. Sikorsky California Department of Transportation, Sacramento, CA, USA

A common practice for the fabrication of steel orthotropic bridge deck with closed ribs in the US is to use 80% Partial-Joint-Penetration groove welds (PJP) for rib-to-deck joints. In practice, it is difficult to control the amount of penetration into the joint components, and weld melt-through inside the thin rib is likely to occur. When weld melt-through occurs, which is difficult to inspect, it is not clear how the resulting geometric discontinuities would have an effect on the fatigue resistance of the rib-to-deck welded joint. A distortion control plan, which involves heat straightening or even pre-cambering, is also used for the fabricated orthotropic deck in order to meet the flatness requirement. It is also unclear how repeated heating would affect the fatigue resistance. Six two-span full-scale orthotropic steel deck panels (10 m long by 3 m wide) were fabricated and tested in order to study the effects of both weld melt-through and distortion control measures on the fatigue resistance of the PJP welded joint. Three of the specimens were only heat straightened, and the other three were pre-cambered to minimize the need for subsequent heat straightening. For each distortion control scheme, one of the three weld conditions [80% PJP weld with no melt- through, 100% PJP weld with evident continuous weld melt-through, and alternating the above two weld conditions every 1 m] was used for each specimen. Up to 8 million cycles of loading, which simulated the expected maximum stress range corresponding to an axle load of 3×HS15 with 15% impact, were applied at the midspans and were out of phase to simulate the effect of a moving truck. Test results showed that six cracks from the weld toe and only one crack from the weld root developed and propagated to the deck plate. The crack from the weld root initiated from a location transitioning from 80% PJP weld to 100% PJP weld. Based on the loading pattern and the limited test data, it may be concluded that discontinuous weld melt-through is not desirable. Rib-to-deck weld cracks were identified in three specimens that required significant heat straightening. On the other hand, the two effectively pre-cambered specimens did not experience PJP weld cracks. This observation suggests that effective pre-cambering is desirable to mitigate the cracking in rib-to-deck PJP welds.

288 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Estimation of low-cycle fatigue strength of steel structural members under earthquake loading

J. Iyama Department of Architecture, Graduate School of Engineering, The University of Tokyo, Japan J.M. Ricles ATLSS Engineering Research Center, Lehigh University, Bethlehem, PA., USA

ABSTRACT: Under cyclic loading causing large plastic deformations (e.g., seismic loading), cracks may develop in the connections of steel structures. These cracks can propagate and eventually cause a low-cycle fatigue failure. Considering a low-cycle fatigue approach, a linear relationship between the number of loading cycles to fracture and the plastic strain range (Coffin-Manson rela- tionship) can be used to predict the number of cycles to failure of structural steel elements subjected to large cyclic plastic deformations due to seismic loading. The Coffin-Manson relationship was established experimentally, where the constants in the relationship are determined by tests. It is difficult to apply this relationship to actual complex steel structures, since there are many members and connections in a structural system and it is impractical to test all the structural elements to determine their constants for the Coffin-Manson relationship. From a micromechanical viewpoint, ductile fracture can be explained by a void initiation, void coalescence, followed by crack extension. Based on this fracture mechanism, some models have been proposed to predict the plastic strain capacity of steel elements at fracture under monotonic and cyclic loading. As this model enables the strain capacity in a micro region to be determined, a finite element analysis with a fine mesh are required to predict the crack length and the remaining strength of a structural element with a complex geometry (e.g., a welded connection). A method to predict the crack length in a structural steel connection or element subject to large plastic cyclic deformations is presented in this paper. In the method, crack length is used as a damage index and fracture is considered to occur when the predicted crack length reaches a critical value. The predicted crack length is calculated based on an established crack extension rule using a micromechanical approach, considering the effect of the stress triaxiality ratio. The methodology requires a finite element analysis to obtain the history of stress and strain. The crack extension rule however simplifies the finite element analysis because the rule requires the stress-strain information only at the surface where the crack is anticipated to develop. To validate the method, a connection specimen tested in the laboratory was analyzed. The specimen was a welded beam-to-column connection that was subjected to inelastic cyclic loading. In the test, a fracture occurred in the bottom beam flange base metal and beam web during the 2nd cycle of 5% drift. The crack length in the beam flanges predicted using the proposed methodology was smaller than the experimental results. This is because the finite element analysis did not consider the crack extension in the groove weld between the beam web and the column flange that occurred during the test, which would alter the stress and strain state in the beam flanges. Consequently, additional analysis was performed with pre-cracks at the interface of the beam web and column flange to simulate the cracks observed in the test specimen at this location. In these additional analyses, the predicted crack length in the beam flanges at fracture was in close agreement with the experimental results. From these analyses, the methodology is considered to be able to predict the crack length and fracture of a complex structural element as long as the stress-strain state at the interested locations is well established.

289 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Variability analysis of fatigue crack growth rates of materials from ancient Portuguese steel bridges

J.A.F.O. Correia & A.M.P. Jesus IDMEC-Pólo FEUP,Porto, Portugal UTAD, Engineering Department, Mechanical Engineering Division, Vila Real, Portugal M.A.V.Figueiredo IDMEC-Pólo FEUP,Porto, Portugal A.S. Ribeiro UTAD, Engineering Department, Mechanical Engineering Division, Vila Real, Portugal A.A. Fernandes IDMEC-Pólo FEUP,Porto, Portugal

In Portugal there is a number of steel riveted railway and highway bridges more than one hundred years old, still in operation, requiring rehabilitation. A consistent residual fatigue life prediction should be based on actual fatigue data from bridge members which is often limited. Authors conducted crack propagation tests on steels from three centenary riveted bridges, namely the Pinhão and Luiz I highway bridges and the Viana highway/railway bridge, and from the fifty years old Trezói railway bridge from Portugal. The crack growth rates are correlated using the Paris’s law. Figure 1 presents the crack propagation data obtained for all tested materials. A tentative is done to derive a unique relation for all data together. A linear regression analysis is carried out, resulting a determination coefficient R2 = 0.87 which is relatively high, taking into account the different origins of the investigated materials. It can be verified that some data from the Luiz I bridge diverges from the average values, presenting higher crack propagation values. One specimen from the Viana bridge exhibits markedly lower crack propagation rates for intermediate K. The constant m from the Paris’s law obtained for all data together is higher than 3.0, being the latter

Figure 1. Crack growth data for all materials.

290 value usually adopted in current design codes of practice. The constant C is significantly lower than usual values recommended in literature for modern steels: 1.2 × 10−13 ≤ C ≤ 5 × 10−13. Figure 2 also includes an upper bound, parallel to the linear regression line, which can be used for design purposes. Only few points of the material from the Luiz I bridge falls above this line; however these points fall within the phase III of crack propagation. Another upper bound was established based on a slope of 3. Clearly, a crack propagation curve based on slope of 3 and C = 5 × 10−3 will produce safe results for all analyzed materials.

ACKNOWLEDGEMENTS

This work was supported by the Portuguese Scientific Foundation (FCT) through the project PTDC/EME-PME/78833/2006.

291 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Assessment of the coupled effect of corrosion-fatigue on the reliability of RC bridges

E. Bastidas-Arteaga & M. Sánchez-Silva Department of Civil and Environmental Engineering, Universidad de los Andes, Bogotá, Colombia Ph. Bressolette & A. Chateauneuf Laboratory of Civil Engineering, Blaise Pascal University, Aubière, France W. Raphael Superior School of Engineering, Saint-Joseph University, Beirut, Lebanon

The durability of Reinforced Concrete (RC) bridges in non-aggressive environments is often sat- isfactory. However, under certain environmental conditions, there are external actions, which can significantly reduce their lifetime. This paper presents an improved model to assess the coupled effect of corrosion and fatigue on the structural reliability of RC bridges. This coupled effect is commonly found in structures located in chloride-contaminated environments and subject to cyclic loading. Separately, corrosion leads to cross-section reduction while fatigue induces the nucleation and propagation of a crack in the bar. When considered together, pitting corrosion nucleates the crack while environmental factors affect the kinematics of crack propagation. The integration of the corrosion-fatigue model into a suitable probabilistic framework is necessary to perform efficient probabilistic lifetime assessments and reliability analysis. The limit state function is defined by the margin between the crack or pit size and the critical crack size, corresponding to bending failure of the structural member. Given the complexity of the problem, closed-form solutions for both the CDF corrosion-fatigue lifetime and the failure probability are not possible due to large degree of nonlinearities, and Monte Carlo simulations are applied to deal with this problem. The implications of the coupled effect of the corrosion-fatigue process on the reliability of a RC bridge girder are illustrated and discussed. In this study, the considered girder, located in two different environmental conditions, is subjected to various cyclic loading frequencies. Three durability design specifications are also taken into consideration for the analysis. Figure 1 depicts the results of the time-dependent reliability analysis. In can be noted that the coupled action of corrosion-fatigue increases substantially the failure probability, in particular, for aggressive environments and for poor durability design specifications. It was also observed that, given the larger contribution of the time to corrosion initiation in the total lifetime, countermeasures must be directed to control this stage of the process. As an example of such countermeasures, the improvement in the durability design specifications reduces the failure probabilities (Fig. 1).

Figure 1. Time-dependent structural reliability for a) tidal and b) atmospheric environment. 292 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

An optimal design of TMD for the improvement of fatigue reliability of steel-composite high-speed railway bridges using target performance approach

S.-J. Kim, S.-C. Kang & H.-M. Koh Seoul National University, Korea W. Park Korea Bridge Design & Engineering Research Center, Korea

This study investigates the effectiveness of Tuned Mass Damper (TMD) in view of the fatigue reliability for a steel-composite high-speed railway bridge. It is shown that the fatigue life of the bridge can be extended by the reduction of resonance-induced vibration using TMD. In parametric analyses, the extent of fatigue reliability improvement and the optimal design factors of TMD are studied by using the procedure of fatigue reliability evaluation based on dynamic analysis. Considering the uncertainties of the train-velocities and the damping ratio of bridge, dynamic analyses are performed repeatedly for the bridge subjected to moving loads by high-speed trains. The uncertainty of the damping ratio is considered as a random variable and the fatigue failure probability is integrated over all the sampling points of the damping ratio. Stress time histories are determined from the time domain analysis at the critical structural components of the bridge. The rain-flow counting method is adopted to obtain a number of stress ranges and cycles from time history responses. Based on these stress ranges, cycles and intended service life, fatigue failure probability are evaluated through the S-N curve-based approach. Through numerical simulations of a steel-composite bridge of 40m span, the effectiveness of TMD on fatigue life of the bridge is examined and the results are presented. Also, the optimal design parameters of TMD including the mass ratio, the natural frequency and the damping ratio are investigated and determined.

Keywords: tuned mass damper, steel-composite high-speed railway bridge, dynamic analysis, S-N curve, fatigue reliability.

293 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Fatigue damage of orthotropic steel bridge decks and its retrofit

T. Shimozato, T.Yabuki & Y. Arizumi University of Ryukyu, Okinawa, Japan Y. Hirabayashi Metropolitan Expressway Company Limited, Tokyo, Japan N. Inaba Nippon Expressway Research Institute Company Limited, Tokyo, Japan S. Ono Japan Construction Method and Machinery Research Institute, Shizuoka, Japan

ABSTRACT: A number of fatigue cracks have been detected in orthotropic steel bridge decks for expressways in Japan. In particular, the crack initiated at the weld root of the longitudinal welding between deck plates and closed trapezoidal longitudinal ribs is one of the most severe cracks because the cracks have the potential to penetrate through the deck plate and to induce cave-in in the road. Therefore, the development of the suitable retrofitting methods has been urged for actual orthotropic steel bridge decks in service. In this study, Fatigue test with a full-sized orthotropic steel deck specimen has been carried out under the similar wheel loading condition of the actual traffic vehicles first. Then stress characteristic and fatigue properties of existing orthotropic steel bridge decks are closely mentioned. Finally the retrofitting method is proposed and evaluated the effectiveness and the fatigue resistance. In this report, the outline of the fatigue test and the obtained results will be presented and discussed for existing and retrofitted orthotropic steel bridge decks.

294 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Fatigue design for highway bridge ancillary structures

Y.C. Park, S. Roy & R. Sause ATLSS Center, Lehigh University, Bethlehem, PA, USA

In last two decades, the report for fatigue cracking on the highway sign and signal structures were increased, which has serious economic impact on the limited infrastructure resources from safety and maintenance point of view. The fatigue performance of welded connections in sign, signal and luminaire structures are unknown.These structures are generally have bridge or cantilever types, and tube to end plate, hand hole and arm to pole connection are the critical details that cause the fatigue crack. Among those critical details, most of fatigue crack happened at tube to end plate connection. The fixity of the boundary condition creates edge perturbation that introduces out-of-plane flexural deformations. Due to thin section size and small section modulus, this deformation introduces large out-of-plane bending stresses through the thickness of tube. The superposition of this out-of-plane flexural stress on to the in-plane membrane stress magnifies the local stress state near the support which governs the fatigue performance of the detail. To evaluate the performance of each detail, hot spot stress based analytical protocol was used to predict the fatigue performance of details. Within number of suggested hot spot stress definition, DNV (Det Norske Veritas) recommended hot spot stress approach was accepted for this research because of its strong theoretical background of shell theory. The analytical predictions were verified against results of fatigue tests on full size specimens. Twelve specimens with various connection details such as socket connection and full penetration connection and tube shapes are tested and investigated. Based on the comparative validation study an analytical protocol was proposed for reliably and consistently assessing fatigue performance of various welded connections in the subject structures. (Fig. 1)

Figure 1. Fatigue test result for specimen Type I (left) and II (right).

295

Fiber reinforced composites in bridges

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

On mechanical performance of different type of FRP beams as reinforcement of pedestrian bridge

G. Boscato & S. Russo University Iuav of Venice, Venice, Italy

The main results and performances deriving by the application of GFRP pultruded profiles in reinforcing the pedestrian bridge deck are shown. The footbridge “Paludo” is a typical historic (end of XIX century) building of Venice, with arc static scheme – 12.7 meters for the length and 3.25 meters for the width – built entirely with iron and wood materials, as shown in Figure 1. The item as been developed both in term of research and real application. In the first case all benefits derived from the application of fiber reinforced beams instead of traditional materials, i.e. wood and steel, was illustrated also to focus the increment of the strength performance in function of the dead load, the interaction with the existing iron deck and wood beams and finally the increment of flexural stiffness assured. In the second case, i.e. the real application and the new way in realizing the reinforcement, all phases regarding the beam–beam connections through bolt in the deck in presence of existing wood beams and iron structure and the in situ test to determine the displacements, are analyzed. A numerical analysis in term of general approach concerning the type of reinforcement of pedestrian bridge deck, balancing the mechanical characteristics of material employed and the performances utilizing very light and strength material as pultruded FRP profiles has been finally proposed in the paper also with comparison with traditional materials.

Figure 1. Transversal and longitudinal sections of pedestrian bridge.

299 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Shaping composite bridges for traffic and the environment

R.A. Daniel Ministry of Transport Public Works and Water Management, Civil Engineering Division, Utrecht, The Netherlands

Fast economical growth introduces new demands to infrastructure in general, and bridges in par- ticular. Carrying traffic loads is not the only task of a bridge today; other demands include doing it in a sustainable way and helping to reduce environmental pollution. This stimulates the demand for new bridge shapes. Thanks to the high strength, low maintenance and other properties, composites can often cope better with such demands than the conventional materials. So far, the shapes of composite bridges usually follow the bridge shapes of other materials, like steel. This was indeed a suitable approach in the introduction stage of composite bridges. The best way to win customer’s confidence was to copy the shapes and sections that proved to be suitable in other materials. Therefore we usually see composite I-beams, channels, box girders, deck grids etc. that resemble the conventional steel or aluminum components. Now that both the confidence and the demands are growing, it is time to review this strategy and to ask which shapes are the most appropriate and offer the best performances in composite bridges. This paper presents some considerations concerning the shape selection in composite bridges. Attention has been paid to the shapes of entire bridges as well as their components, such as girders, crossbeams and decks. In particular, the composite anisotropy and flexible technology offer the shapes of double curvature, smooth transitions, less fatigue sensitivity, high strength in the areas of stress peaks etc. – increasing both aesthetic value and service life. Some new shape ideas for composite bridges have globally been sketched and discussed. This includes the shapes, which help solving diverse environmental problems, like fine dust pollution, CO2 and other emissions caused by the traffic. Transport and traffic – the pillars of the Dutch economy – contribute to this pollution to large extend. Relatively simple precautions have already been taken; the Netherlands belong to the countries of the most rigorous rules and control procedures in the world in this field. Yet, this is not enough. New, challenging solutions in new materials are required. Composites screening or even roofing of traffic lanes, as well as new composite bridge shapes, can help to isolate the air pollution. The author’s suggestions in this field have already been presented and discussed within the composite branch. The intention of this paper is to draw the attention of bridge designers, constructors and managers, in order to initiate more research, design and successful applications.

300 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

A study on the dynamic behavior of a CFRP cable

J. Park Department of Civil and Environmental Engineering, Dankook University, Korea K. J. Hong Department of Civil and Environmental Engineering, Kookmin University, Korea

In the history of structures, the transition from the compressive component such as stone and brick to the tensile steel component leaded to a big change in the form of the structures. Especially, steel cable enabled the design of cable stayed bridge and suspension bridge due to its efficient use of the tensile strength. Currently, many efforts are under way to maximize the span length. One of the most important elements of the long span is the high strength and light weight cable. In this study, we consider the specifications of the NPWS steel cable which was used in the design of the Second Jin-Do Island Bridge, which is a cable stayed bridge connecting Jin-Do Island in Korea. The proposed CFRP cable has the same shape, dimension and strength as the NPWS cable and the only difference is the tensile modulus (86% of the steel cable) For a test setup, the original length of the cable is 5.5 m and each end is subject to an external displacement loading as shown in Figure 1. Figure 2 shows the internal axial force of the cable developed during cyclic loading with 3.5 Hz frequency. This result reveals that CFRP cable has superior to the NPWS cable for dynamic loads to reduce the internal axial force of the cable.

Figure 1. Test setup.

Figure 2. Internal force developed during cyclic loading (f = 3.5 Hz). 301 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Composite ‘Delta Deck’: The promising bridge deck for new and rehabilitated bridges

S.W. Lee & K.J. Hong School of Civil Engineering, Kookmin University, Seoul, Korea

To cope with the problems of deterioration and corrosion in conventional steel and concrete materi- als, high durable and lightweight fiber-reinforced composites are considered one of most promising alternative materials for civil infrastructures. Among many applications, composite decks for bridges seem to be most notable. The composite decks for bridges have significant advantages compared to conventional con- crete decks since they are highly durable and corrosion-free. Much longer service life and lower maintenance cost are expected for bridges with composite decks, which will result in a much lower Life-Cycle Cost (LCC). Due to the light weight of the composite deck, it can reduce dead load by as much as 80% compared to that of the conventional concrete deck. Much slender substructures are possible for the bridges thanks to the light weight of the composite decks. When the composite deck is used for re-decking in a bridge, the capacity of live load on the bridge is upgraded without strengthening its girders or substructures. Furthermore, composite decks can be installed quickly, significantly reducing the duration of construction and traffic-block, so that considerable savings can be achieved. Described in this paper are development of two types of pultruded composite decks, one with tongue-and-groove connection (Fig. 1) and the other with snap-fit connection (Fig. 2). Analyses and tests on these two types of the composite decks to verify their structural performance are conducted and the results are summarized in this paper. Recent applications of the composite decks with tongue-and-groove connection are introduced including the world largest composite deck bridge, ‘Noolcha Bridge’. It also presents recent applications of the composite deck with snap-fit connection to pedestrian bridges.

Figure 1. Composite deck with tongue-and-groove connection. Figure 2. Composite deck with snap-fit connection.

302 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Advanced removable connection for glass fiber reinforced polymer bridges

D.-U. Park, K.-J. Hwang & J. Knippers Institute of Building Structures and Structural Design (ITKE), University of Stuttgart, Stuttgart, Germany

The current design philosophy for the applications of Glass Fiber Reinforced Plastic (GFRP) in civil engineering is merely an imitation of traditional steel construction. The visual appearance of many GFRP bridges is thus very similar to those made of steel or timber. Especially in the case of detachable connections for bridges, there are not enough appropriate options of GFRP connections, such as the traditional bolt connection for steel and wood structures. However, due to the deficient ductility of GFRP, the shear-stress caused by a classical bolt connection, makes it inappropriate for the needed loading capacity. Based on this research problem, the authors are introducing an innovative and detachable connecting system. This new system is based on the principles of form- fit-connection, which is considered to be more appropriate for the material characteristics of GFRP. For the assignment of this study, short term and long term investigations were performed.

Figure 1. Tests and examples of the advanced removable GFRP connection.

303 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Full pultruded FRP profile structures

M.D.G. Pulido PEDELTA, Structural Engineers, Spain

ABSTRACT: Over the last decade there has been significant growth in the use of FRP com- posite materials as construction materials in structural engineering. These materials have proven themselves to be valuable for use in the construction of new buildings and bridges and for the upgrading of existing structures thanks to its lightness, minimum maintenance costs, execution easiness, corrosion resistance, durability, high specific strength, no magnetic interaction, etc. The initial cost due to material supply and design is higher when compared with traditional steel-based solutions, however, considerable savings are made in construction and maintenance. The development and future of advanced composite materials for architectural and civil engi- neering structural applications will depend basically on the development of new structural forms and element-joining techniques. Structural codes would help to spread the use of full advanced composite plastic structures. In the paper, several full pultruded FRP structures are presented: two pedestrian bridges; the Façade of the Conference Center of Badajoz and an outstanding example of an elegant public build- ing integrated into a consolidated urban area. General description, requirements, design criteria, analysis structural, construction and conclusions of all the structures will be described in the full paper. Engineers possess a magnificent potential knowledge regarding the behaviour of materials and structures but aesthetics should also be taken into account as well as structural performance and economy.All this could easily be done by simply making an extra effort during the conceptual design stage. These footbridges demonstrate that advanced composite materials can be easily introduced into most types of structural forms thus taking advantage of their outstanding mechanical and chemical properties.

304 Health monitoring

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Detachable sensor for bridge cable maintenance and safety

S. Sumitro Smart Structures LLC, Illinois, USA H. Hoashi Honshu-Shikoku Bridge Expressway, Naruto, Japan T. Okamoto Keisoku Research Consultant, Hiroshima, Japan M.L. Wang University of Illinois at Chicago, Illinois, USA

Stress of steel reinforcement for reinforced concrete member, strands for prestressing concrete and cables for suspension bridges are the most important parameters for assessing the health of existing bridge structures with cables. Elasto-Magnetic (EM) sensors have been used extensively in measuring actual stresses of strands and cables of many types of constructions such as ground anchors, cable-stayed bridge cables, pre-stressed concrete tendons, spatial dome bracing rebars and steel of reinforced concrete members. It has achieved accuracies of less than 1% for sizes of up to 20 mm and 5% of up to 250 mm in diameter, respectively, for new construction and with solenoid geometry. However, There are several limitations for permanent type O-shape EM sensors to be used at existing bridge structures. Therefore, it is highly required to develop portable and reusable sensors for directly monitoring the actual stress of steel wire, cable and tendons. As a breakthrough in sensory technology, U-shape EM sensors were developed. The sensor is removable and reusable. Small detachable EM sensor has proven to be accurate of within 3% for strand sizes of up to 20 mm in diameter. The application of large U-shape sensor was conducted during the replacement of hanger rope in a diameter of 61 mm at Ohnaruto Bridge (see Fig. 1). The capability and advantage of a detachable EM sensor is confirmed and accepted for bridge cable maintenance and safety.

Figure 1. Stress measurement during hanger rope replacement.

307 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Stock Condition Index analyses response to Bridge Condition Index determinations

P.S. McCarten Opus International Consultants, Napier, New Zealand

ABSTRACT: For Road Controlling Authorities to show good stewardship of bridge assets and compliance with performance targets set for bridge investment programmes condition monitoring and reporting is required. The monitoring and reporting can be based on qualitative low cost inspection programmes using skilled bridge engineering staff or quantitative data collected from high cost sophisticated monitoring programmes. This paper addresses reporting techniques using the former method. Bridge Managers are now applying asset management principles in day-to-day structures man- agement. In recent times it has been shown that preventive maintenance is more cost effective than reactive maintenance requiring any Bridge Manager to be more proactive in monitoring and reporting condition. In order for a Bridge Manager to report the cost effectiveness and efficacy of his bridge maintenance and rehabilitation expenditure programme there needs to be a strong link between bridge condition and the funding investment. In McCarten (2004) a systems method for best bridge maintenance management was outlined and this included a hierarchy of asset preservation performance measures and targets which was calibrated to British Columbia Ministry of Transportation (BCMoT) bridge maintenance manage- ment. In McCarten (2006) a case study of 40 bridges in the BCMoT network using condition data generated from the BCMoT bridge inspection programme from 1996 to 2004 presented a technique for reporting cost effectiveness and efficacy of the bridge maintenance programme and this involved statistical analysis of the Stock Condition Index (SCI). This technique is based on Condition Rating and Indexes which are dependent on the equation weighting factors used for the Bridge Condition Index (BCI). The focus of the research for this paper is to assess the significance of the BCI determination on the correlation between condition and the bridge maintenance and rehabilitation funding invest- ment. The condition data for the 40 bridges used in the McCarten (2006) case study has been input to the BCMoT Condition Rating and Index equations but varying the weighting factors in those equations. Eight BCI determination scenarios were investigated in this study. The results of these analyses have been compared qualitatively to the British Columbia Ministry of Transportation bridge maintenance and rehabilitation funding investment programme over the study period to identify the BCI determination scenario giving the best correlation. It has been concluded that BCI determination scenarios based on weightings assessed using structure risk give the best correlation. With this correlation the analyses shows the annual bridge maintenance and rehabilitation funding investment required to hold the bridge condition to an agreed target level is between 0.3% and 0.5% of the Bridge Replacement Cost. While the study is limited in its size the results are meaningful and give confidence to the findings and techniques applied to data from a well set up bridge inspection and monitoring programme. The method is best suited to the ‘zone of tolerable deterioration’ within the structure life. The importance of using experienced Bridge Engineers to forecast specific funding needs cannot be overstated.

308 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Traffic-induced vibration of bridges: Input force identification

Chin-Hsiung Loh, Ai-Lun Wu & Jian-Huang Weng Department of Civil Engineering, National Taiwan University, Taipei, Taiwan

ABSTRACT: Accurate characterization of input excitation forces acting on a structure dur- ing operation is significant for designing, controlling and diagnosing a structural system. Prior knowledge of accurate input excitation forces can contribute to a greater reliability of numerical simulations as well as reducing the expensive costs for time-consuming experimental tests. Besides, the input force information can provide important message for developing the damage prognosis model of structures. In some circumstances, input excitation forces can be directly measured using load or acceleration transducers. However, for some physical and mechanical systems, direct mea- surements of input excitation forces are difficult to be realized due to very large magnitudes of forces or installation obstacle of load transducers or the complex interaction of loading and struc- tural response. Therefore, an alternate method to reconstruct input excitation forces is needed. One of the methods is to identify the dynamic excitation forces based upon measuring structural responses using an inverse method. In addition to conventional inverse method, the Kalman filter based tracking approach has been studied and developed for the identification of input excitation forces due to its accurate estima- tion in the consideration of measurement noise and modelling error. Singer (1970) augmented the Kalman filter with the target acceleration equation represented by a first-order autoregressive process. Chan et al. (1979) proposed a Kalman filter based tracking scheme with input estimation and used a detector to guard against automatic updating of the simple Kalman filter. Ma (1998) proposed an application of the Kalman filter to determine the impulse loads of a lumped-mass system in a numerical scheme. Liu et al. (2000) used the Kalman filter and least-squares estimator to determine the input excitations of a cantilever plate. The efficiency and robustness of a regular- ization scheme of the proposed algorithm using the Kalman filter and recursive least-squares were proven better than the conventional methods (Ma et al. 2003). An identification method for estimating the time varying excitation force acting on a vibration structural system based on its response measurement is presented in this study.The method employs the simple Kalman filter to establish a regression model between the residual innovation and the input excitation forces. Based on the regression model, a recursive least-squares estimator is proposed to identify the input excitation forces incorporating measurement noise and modeling error. Furthermore, application of the proposed approach is conducted on the field test data of traffic-induced vibration of a bridge, so as to study the characteristics of traffic loadings induced by random vehicles.

Keywords: Kalman filter, recursive least square, force identification.

309 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Turning the Humber Bridge into a smart structure

N.A. Hoult, P.R.A. Fidler & C.R. Middleton Department of Engineering, University of Cambridge, Cambridge, UK P.G. Hill Humber Bridge Board, Hessle, UK

ABSTRACT: As part of a project investigating the potential for using Wireless Sensor Networks (WSNs) to monitor Civil Infrastructure, the Humber Bridge has been chosen for field trials of several WSNs. This paper outlines the development of two networks that were created largely from off-the-shelf components. The first network was installed in the north anchorage of the bridge and consists of 11 MICAz motes, 10 of which have Relative Humidity (RH) and temperature sensors. The RH in the anchorage is maintained below 45% through the use of dehumidifiers in order to minimize the chance of the exposed strands of the main steel suspension cables corroding. There are two existing hardwired RH sensors in each of the two chambers of the north anchorage however their output can only be viewed from within the anchorage. The motes transmit the RH and temperature data wirelessly back to a gateway computer in the anchorage which is connected to the Internet. The data from the WSN can be viewed via a webpage either as numerical results, as plots for a given mote superimposed over the location of that mote on the structure or as plots of the data for all the motes. Each of these data visualization techniques has advantages and disadvantages depending on the analysis being performed. During the first five months of operation, the anchorage WSN has suffered two hardware failures. One of the RH and temperature sensor boards stopped working and malfunctioned again a short time after a new sensor board was installed. In this case the solution was to replace the mote but this demonstrated the requirement for sensor redundancy when the data is critical. The second hardware failure occurred when the clock on the gateway computer began malfunctioning. The clock would keep regular time during most of the day but would at random periods slow down dramatically. This meant that the timestamps on some of the data packets could not be trusted. This failure illustrated the need for redundancy of this critical element since if the gateway malfunctions in a WSN, all the data can potentially be lost or inaccurate. The second WSN will be deployed in the near future on a three-span reinforced concrete slab bridge just to the north of the suspension bridge. Initially this network will monitor two aspects of bridge performance that have been identified during visual inspections. These inspections found transverse cracks on the soffit of the bridge’s centre span which have lengthened with time. What the inspection reports have failed to indicate is whether the cracks have also increased in width, which is generally quite a difficult parameter to measure in any event. Motes with displacement transducers will be installed on the bridge soffit to determine if the crack widths are increasing. Inspections have also noted that the elastomeric bearings have a slight transverse inclination. Motes with inclinometers will be installed on two of the elastomeric bearings to track any changes in their transverse and longitudinal inclination. Both sets of motes will also have RH and temperature sensors to allow for environmental compensation. This WSN will be completely wireless in terms of transmission with the gateway using GPRS to connect to the Internet. Power for the motes will be provided by long-life lithium batteries while the gateway will be powered by a solar panel and battery system. A final consideration is security of both the data and the hardware since a long-term monitoring system could potentially be exposed to malicious attacks.

310 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Application of data fusion in the safety monitoring of Sutong Bridge foundation

Z.J. Chen, L. Bian & T. Xue College of Civil Engineering, Hohai University, Nanjing, Jiangsu, China X.W. Zhang Jiangsu Province Sutong Bridge Construction Commanding Department, Nantong, Jiangsu, China

In the strong tidal river, there are many complex influence factors of pile group foundation bearing behavior, of which many factors, especially the effect degree of pile groups, have not been distin- guished in the prophase of project. So prototype monitoring is taken for the pile group foundation of the cable-stayed bridge with the span of 1088m in Sutong Bridge construction. There are 1531 stress and strain sensors used for long-term monitoring in the two-tower pile group, including vibrating wire concrete embedment strain gages, vibrating wire rebar strain meters, arc weldable strain gages, earth pressure cell, vibrating wire concrete embedment non-stress gages, and low pressure vibrating wire pressure transducer. The main observed contents include the follows: depth of riverbed scour, tide, earth pressure at the bottom of pile, pile axial force in different heights (especially the axial force at the top of pile), three-dimensional stress in the pile cap (especially the stress at the bottom of pile cap), flexural deformation of pile cap, etc. Tracking-observation after installing sensors is taken as soon as the construction of pile foundation. The earliest time of data acquisition is January 28, 2004, and it continues to the period of the operation of Sutong Bridge. The observed frequency varies with the different construction characters and observed purpose. The main task is evaluating the overall bearing behavior by checking out pile group effects and transferring paths of loads, therefore elaborate observed data are needed. But prototype observed data include strong noise which is occurred by tide level, tidal flow, wave, water temperature and accidental factors and would seriously interfere the analysis and forecast of the pile founda- tion bearing capacity. But the noise generated by different factors has different characteristics, so signal-noise separation technique is used to separate the influencing factors. The measured results of the six sensors (three types) on the same section recorded from January 14, 2006 to January 15, 2006 are chosen in representative firstly. Then gross errors are removed and the physical quantities units of data are unified. Finally compactly supported orthogonal db4 is chosen as wavelet basis function, and then three-layer decomposition is done to the data so that random noise is basically eliminated. The stress data of the pile foundation related to the upper load can be gained through increasing the wavelet decomposition scale to six that can further eliminate the influence of tide level. Considering that the credibility and importance vary for the observed data from different sensors, the data of sensors need to be coordinately used with data fusion. Considered all the fac- tors, the credibility level from high to lower is vibrating wire concrete embedment strain gages, vibrating wire rebar strain meters and arc weldable strain gages. These factors are also represented in the observed data sequence. With the data fusion based on optimal weight distribution principle of right, different weights are given according to the level of the data sequences credibility so that the fusion results can be more reasonable.

Keywords: Sutong Bridge, pile group foundation, safety, transfer mechanism, extensive angle monitoring, multi-sensor information fusion.

311 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Remote measurement of crack length in sacrificial test piece by self-reference lock-in thermography

Y. Sakino Joining and Welding Research Institute, Osaka University, Osaka, Japan T. Sakagami Department of Mechanical Engineering, Osaka University, Osaka, Japan Y.-C. Kim Joining and Welding Research Institute, Osaka University, Osaka, Japan

“The sacrificial test piece” is used as a specimen attached to the member of a main structure in order to evaluate the damage before the appearance of a crack in a member of the main structure. The sacri- ficial test piece is designed so that it is damaged earlier than the main members under the same loads because of crack and stress magnification. The damage to the bridge members can be estimated by the observation of the sacrificial test piece. If the fatigue damage parameter can be made clear by the behavior of the sacrificial test piece, the maintenance management of the bridge can be determined. Thin steel plates, which have initial cracks at the center, are used as the sacrificial test pieces in this study. When strains are applied to the main member, these are transmitted from the main member to the thin steel plate and the crack in the thin steel plate will grow as a result. Therefore, the monitoring of fatigue damage parameters on the bridge can be carried out by the observation of the crack growth in the thin steel plate. In this paper, the estimating method for the fatigue damage parameter by crack growth of the thin steel plate is summarized. And it is demonstrated that the thin steel plate with a crack can estimate the fatigue damage parameters with practical accuracy under constant amplitude loading by comparison between the fatigue damage parameters measured by the crack length of the sacrificial test pieces and those calculated by the stress amplitude and loading times. If the crack length of the sacrificial test piece can be measured from a long distance, the damage in a member of bridges can be evaluated still easier and cheaper. We propose measuring method of the crack length from a long distance by a self-reference lock-in thermography. The self-reference lock-in thermography enables to measure the distribution of relative intensity of applied stress under random loading without using any external loading signal. To clarify the applicability of the self-reference lock-in thermography to the measurement of the crack length of the sacrificial test piece from long distance, fatigue tests of the sacrificial test pieces were performed and crack length was measured by proposed method. Distance from the lens to thin steel plate was 2 m. Loading stress range and frequency during measuring were 120 MPa–10 Hz and 60 MPa–3 Hz. In the measurement of 26 mm crack length, 30 MPa–3 Hz and 15 MPa–3 Hz were also performed. Assuming the stress increase ratio is 3, the stress range of 120, 60, 30 and 15 MPa of the sacrificial test piece correspond to that of 40, 20, 10 and 5 MPa of the bridge member by using jig-plate. As a result, location of crack tip can estimated as the largest point of the thermoelastic temperature change in spite of the crack length. And good agreement can be found between the crack length measured by crack gauge and those estimated by the self-reference lock-in thermog- raphy in all range not only in the case of 120 MPa–10 Hz, but also in the case of 60 MPa–3 Hz. The stress range became smaller, the thermoelastic temperature change and the location of crack tip also became small and unclear. But in the case of 30 MPa or more, the self-reference lock-in thermography can measure the clack length of the sacrificial test piece accurately. From these results, the possibility of measurement of crack length in the sacrificial test pieces in the bridge members from long distance by the self-reference lock-in thermography is indicated. 312 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Wind-induced vibrations and countermeasures for cable systems on long-span bridges

I. Yamada, S. Kusuhara, K. Fumoto & N. Toyama Long-span Bridge Engineering Center (LBEC), Honshu-Shikoku Bridge Expressway Company Limited (HSBE), Japan

For the cable systems on the Honshu-Shikoku Bridges, various countermeasures against wind- induced vibrations have been introduced during construction and/or operation (Table 1). For example, on the Tatara Bridge, completed in 1999 as the world’s longest cable-stayed bridge, high-damping rubbers against vortex-induced oscillation were installed at cable-anchoring points of girders and polyethylene-covered cables with indents against rain-induced vibration were prefab- ricated and installed for the first time around the world. On theAkashi Kaikyo Bridge, completed in 1998 as the world’slongest suspension bridge, the original high-damping rubbers were unexpectedly damaged by a strong typhoon and were replaced by helical wires around parallel suspender ropes. On the parallel stay cables of the Hitsuishi/Iwaguro Bridges, completed in 1988 as twin cable- stayed bridges, the vibration countermeasure was originally taken by cross ties and rigid ties. However, several cross ties happened to cut off during operation, because the sub-span vibrations were not necessarily controlled. For a new countermeasure, the helical wire was examined by the wind tunnel tests and the field measurements, and was verified one of the most effective countermeasures. Currently, the cross ties are restored after a minor improvement in the anchoring details and the durability are closely observed after improvement. On the main cable of the Kurushima Kaikyo Bridge, completed in 1999 as three consecutive suspension bridges, the supporting detail of hand ropes for inspection/maintenance was slightly damaged by the wind-induced vibration. In order to identify the cause and mechanism of the vibration, two accelerometers and an anemometer were temporally arranged on the hand rope. The observation revealed that small vibrations occurred in the low wind velocity and that large vibrations occurred in the high wind velocity. The former are the vortex-induced vibrations and the latter might be under the influence of vortexes or streams separated from the main cable. The further analysis and consideration are under way. This paper describes a brief review of vibrations and countermeasures for cable systems on the Honshu-Shikoku Bridges, and also describes the countermeasure for parallel stay cables on the cable-stayed bridge. In addition, the paper describes the vibration measurement of hand ropes on the suspension bridge. Table 1. Cable vibrations and countermeasures on the Honshu-Shikoku bridges.

Bridge name/type Opened Member Cable vibrations Countermeasures

Hitsuishi/Iwaguro Br. 1988 Stay cable Wake galloping, Cross ties (cable-stayed br.) (parallel) Rain-induced vibration and Rigid ties Ikuchi Bridge 1991 Stay cable Vortex-induced oscillation High-damping rubber (cable-stayed br.) (single) damper Tatara Bridge 1999 Stay cable Vortex-induced oscillation, High-damping rubber (cable-stayed br.) (single) Rain-induced vibration and indented cables Akashi Kaikyo Br. 1998 Suspender Vortex -induced oscillation, Helical wires (suspension br.) (parallel) Wake flutter (original: dampers) Kurushima Kaikyo Br. 1999 Suspender Vortex -induced vibration High-damping rubber (suspension br.) (single) damper

313 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Structural monitoring of a bridge prestressed concrete beam under loading

C. Cremona, C. Tessier & V.Le Cam Laboratoire Central des Ponts et Chaussées, Paris, France B. Tonnoir Centre d’Etudes Techniques de l’Equipement Nord-Picardie, Lille, France R. Leconte Centre d’Etudes Techniques de l’Equipement de Lyon, Bron, France V.Barbier Centre d’Etudes Techniques de l’Equipement de l’Est, Nancy, France

TheVIPP (Viaducs à travées Indépendantes à Poutres préfabriquées en béton Précontraint par post- tension) are simple span viaducts made of precast concrete girders prestressed by post-tension. This type of construction was largely developed between the years 1955 to 1970, thanks to its girders launching system which allowed the crossing of non classical obstacles with reasonable height (10 to 25 m above ground-level, range between 30 to 60 m). This technique is not longer used, largely competed by other construction techniques making it possible to carry out more economic and safer redundant structures with respect to rupture. The enthusiasm which reigned at the construction time of the first VIPP generation moreover resulted in many design and exe- cution mistakes, and by the absence of corrective maintenance actions to restore the defective waterproofing of these bridges. The detailed description of the problems encountered in this family of bridges shows that the diagnosis of these structures is essential to conclude a rele- vant assessment of their performance. It is in this context that the network of the Public Works Laboratories of the Ministry of Transport decided to initiate in 2005 a set of experimental and numerical investigations. The Merlebach bridge suffered from all the classical VIPP’s deficien- cies and its replacement offered the opportunity to conserve one the most degraded beams in order to test its strength. Load tests were therefore applied and for each loading case, a set of measurements (acoustic emission, dynamic assessment, curvature measurements, displace- ment measurements, gage measurements) were performed in order to appreciate the structural behaviour. The objectives of these investigations are to correlate the degradation level provided a bending test with a post-mortem analysis and results from various non-destructive methods. Under the effect of an increasing loading, the study thus consisted in applying the following methods:

a) Acoustic emission: this method consists in counting the number of wire ruptures during the test and locating them on the beam cross-sections. Two techniques were considered: the traditional monitoring LCPC-CASC and the new wireless system LCPC-CASC. Acoustic emission sensors were also installed to monitor the concrete cracking; b) Extensometry: several strain gauges on three cross-sections were installed in order to trace the deformation profiles and thus to study of their evolution under loading. Curvatures were also estimated by this traditional technique; c) Curvature measurement: this technique allows getting bending stiffness, locating and quantify- ing damages during tests; 314 d) Dynamic assessment: the objectives of the dynamic tests were to appreciate the temperature influence on the modal characteristics and, after each test of loading, to detect possible structural modifications; The paper presents the principal results obtained from these diagnosis techniques. These inves- tigations make it possible to appreciate the promising character of certain approaches within the framework of the performance assessment of prestressed concrete bridges.

315 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Application of InSAR on ground deformation in the location of Sutong Bridge

Z.J. Chen, N.N. Zhang & X.Y. Li College of Civil Engineering of Hohai University, Nanjing, Jiangsu, China L.Y. Feng Jiangsu Province Sutong Bridge Construction Commanding Department, Nantong, Jiangsu, China

Subsidence is the key monitoring to the foundation of the large bridge. SuTong Bridge is located in wide Yangtze estuary’s tidal reach whose depth of water attains 30–40m. The foundation Subsi- dence observation of the pile group of main pier which is 4 kilometers away from the river bank is affected by weather and especially by tidal level factors, so observation precision of precision optical measurement is limited and has some time-space limitation. Therefore, corner reflectors are installed on the key monitoring parts and with the aid of the permanent scatters characteristics of bridge itself and corner reflector, the method combining Differential Interferometry Synthetic Aperture Radar (D-InSAR) with Permanent Scatters (PS) is applied to monitor the foundation sub- sidence of Sutong Bridge. The 12 scenes SLC data of ENVISATsatellite acquired from 2003(before project construction) to 2007(when main project is basically completed) is used to process “two- pass” mode D-InSAR to gain deformation map of radar Line-of-Sight of pile foundation of main pier of cable-stayed bridge during the process from pouring the platform and the cable tower to hoisting the steel box girder, and gain the vertical deformation component (that is settlement) by conversion. Contrasted with construction working condition of northern cable tower and analyze differential result, it’s known that vertical loads during the pouring of platform and cable tower account for more than 80% of the total, so during this period cumulative amount of subsidence comes to over 100 mm, while vertical loads during the hoisting of steel box girders is relatively fewer, so cumulative amount of subsidence is smaller too. In the process of construction, the obser- vation technologies of low pressure vibrating wire pressure transducers, vibrating wire settlement profilers and precision vibrating wire liquid lever settlement systems are used. All the observation technologies can be mutually verified and have good consistency. The application of D-InSAR technology overcomes the limitation of sensors observation method in the monitoring range, not only implements monitoring to key working condition and position but also reflects the surface deformation of the whole bridge location during the construction objectively and comprehensively. The results indicate coherence of monitoring area is increased by setting corner reflectors and D-InSAR accessorial with PS can become a potential and practical monitoring method of founda- tion subsidence of river-crossing bridge and it can obtain the measured result that is reliable and has the precision of millimeter level. This project will continue to accumulate SAR data, so that PSInSAR calculation will be realized and the high precision settlement of CR and PS in key parts will be extracted.

Keywords: Sutong Bridge, foundation, settlement observation, PS (permanent scatterer), D-InSAR (Differential Interferometry Synthetic Aperture Radar).

316 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Long-term structural health monitoring of the Torino’s pedestrian cabled-stayed bridge

L.M. Giacosa & A. De Stefano Polytechnic of Torino, Torino, Italy

ABSTRACT: Structural health monitoring of an original cable-stayed bridge is discussed. In this paper is highlighted the monitoring methodology adopted and the innovative sensors installed to identify the behaviour of the structure under different condition. It is not a structural health mon- itoring an end of it, but it is a newer approach and prospective to improve the existing methods of civil engineering structures monitoring. In this paper experimental and finite numerical model results are discussed to strengthen methodologies chosen. SHM system provides several types of sensors applied on the structures. Sensors acquisition is simultaneously from all the instruments on the bridge installed, experimental data shall download to a remote acquisition system that elaborat- ing a first pre-processing sending to the users data. Inclinometers and accelerometers sensors are connected among themselves trough a single cable able to provide them energy and also to transmit data to the central unit. Temporary and most important permanent structural health monitoring of an original cable-stayed bridge is discussed. In this paper it is highlighted the monitoring method- ology adopted and the innovative sensors installed to identify the behaviour of the structure under different condition. It is not a structural health monitoring an end of itself, but it is a newer approach and prospective to improve the existing methods of civil engineering structures monitoring. In this paper experimental and finite numerical model results are discussed to strengthen method- ologies chosen. The structure is principally constituted by a flexible steel frame structure with a curvature change along its longitudinal development. As above written the principal and critical points to taken into account and analyze with a specific structural monitoring are cables, pull-down tendon and in general deck behaviour. SHM system provides several type of sensors applied on the structures: making use of the numerical results obtained from designers it was been possible to study properly sensors and technical performance of themselves. Sensors acquisition is simultaneously from all the instruments on the bridge installed, experi- mental data shall download to a remote acquisition system that elaborating a first pre-processing send to the users the data. Inclinometers and accelerometers sensors are connected among themselves trough a single cable able to provided them energy and also to transmit data to the central unit. Fiber Optic Sensor (FOS) used for the cables monitoring is based on FBG fiber optic technology. Active research plan developed by the University of Illinois at Chicago has develop these type of sensors to monitoring the behaviour of tension of the cable. Others type of FOS were installed by the bridge foundation. Fiber Bragg sensors patches were installed directly on the surface of steel frame pillar, connecting the bridge deck and the foundation.

317 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Identifying bridge damage using Brillouin Optical Fiber Sensing

W. Zhang, B. Shi & Y.Q. Zhu Center for Engineering Monitoring with Opto-Electronic Sensing, Nanjing University, Nanjing, China Y.F. Zhang Jiangsu Transportation Research Institute, Nanjing, China

Service loads, environmental and accidental actions may cause damage in bridge structures. In most cases damage will influence the safety or durability of the structure more or less. Therefore, the work of damage identification for a service lifetime prediction, including damage detection, localization and quantification, has received considerable attention for the past two decades. Resulting from the correlation between structural integrity and relative deformation, structural damage can be well interpreted in terms of strain field distribution. The current development of distributed Brillouin Optical Fiber Sensing (BOFS) techniques have provided the measurement availability of strain field distribution for structural damage identification. The theory of Strain Field (SF) method is proposed for identifying bridge damage in this paper. A case study to an existing prestressed concrete box-girder bridge was also presented to verify the feasibility and applicability of this method in practice. The measurements of BOFS, which characterize the strain field distribution continuously both in space and time domains, can be orderly organized in a space-time matrix E.The physical implication of E is a mapping of structural state in a two-dimensional coordinates of both space and time. Damage localization can be implemented by comparing components in damage index with 0. If the condition of greater than Zero is satisfied, it indicates that the measured strain value has exceeded its threshold value, and thus a conclusion can be drawn that damage has occurred. Subsequently, damage quantification can also be implemented by evaluating the value of damage index The greater positive value it possesses, the more severe damage level has been reached. Moreover, The existed damage should be converted to equivalent nominal strain values so as to quantify the absolute damage level. Field test verification to SF method was carried out by loading an existing bridge to its failure state and performing damage identification works on each damage scenarios. The practical ultimate loading was up to 18678 kN, when tensile rebar failure and deflection divergence occurring in the midspan. Also, bending and shear cracks spread all along the web. The sub-matrix featuring the damage scenarios was extracted from the space-time matrix in the first place. And the strain field distribution in space-time was obtained. In order to identify the absolute damage state, the existed crack width in initial structural state was treated as nominal tensile strain by dividing the relevant sensing segment length. Moreover, with regard to crack closure, the strain field value will keep as 200 µε even if which has decreased less than 200 µε because the damage has occurred. And then the components in damage index vector can be identified. In this way damage localization and quantification were ultimately implemented in space-time coordinates.

318 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Innovative treatment of monitoring data for reliability-based structural assessment

Thomas B. Messervey Department of Structural Mechanics, University of Pavia, Italy Dan M. Frangopol Fazlur R. Khan Endowed Chair of Structural Engineering and Architecture, Department of Civil and Environmental Engineering, ATLSS Center, Lehigh University, Bethlehem, PA, USA

ABSTRACT: The planning of monitoring systems and the analysis of monitoring data raises sev- eral important questions. How much data is required and how can it efficiently be managed? What is the potential variability present in the collected information since it is only one observation of a random phenomenon? How does the monitoring data relate to safety provisions and code require- ments? This paper begins to address these questions by applying the statistics of extreme values to simulated monitoring data. Efficient, because only maximum values are recorded within a selected timeframe of interest, once an extreme value distribution is well-defined, a simple transformation predicts the result of that distribution being repeatedly sampled over time due to the asymptotic behavior of extreme distributions. As such, information observed in one timeframe can be related to code-specified return periods. However, one must determine an appropriate timeframe from which to select maximum values and then determine how many timeframes should be observed in order to accurately define the distribution’s parameters. To investigate these questions, a simulation is created that approximates the randomness of live-loads on a bridge structure. Data is generated and maximum values are investigated using different observation timeframes to determine how many observations are required to achieve convergence for both the mean and standard deviation of the maximum values. The simulation is also repeated multiple times to investigate the variability of these parameters at specific benchmarks in time. The results from these simulations show that the mean and standard deviation do converge as the number of observations increases, that the process can be modeled using an extreme value distribution, and that the accuracy of the estimated param- eters improves the longer the process is observed. Tradeoffs between obtaining better parameter accuracy and committing the resources required to collect long-term data are discussed. Parameter variability for a daily observation timeframe is highlighted.

REFERENCES

Ang, A.H-S. & Tang, W.H. 1984. Probability Concepts in Engineering Planning and Design, John Wiley & Sons, II, New York. Frangopol, D.M. & Messervey, T.B. 2007. Lifetime oriented assessment and design optimization concepts under uncertainty: Role of structural health monitoring. Proceedings of the Third International Conference on Lifetime Oriented Design Concepts, Bochum, Germany, November 12–14 (keynote). Frangopol, D.M. & Messervey, T.B. 2008. Effect of SHM on reliability of bridges. Chapter 25 in Monitoring Technologies for Bridge Management: State-of-the-Art, A. A. Mufti and B. Bahkt, eds., Multi-Science Publishing Co., Ltd., England (invited paper, submitted). Messervey, T. & Frangopol, D.M. 2007. Bridge live load effects based on statistics of extremes using on-site load monitoring. Proceedings of the 13th WG 7.5 Working Conference on Reliability and Optimization of Structural Systems, Kobe, Japan. October 13–15, 2006; in Reliability and Optimization of Structural Systems: Assessment, design, and life-cycle performance, Frangopol, D.M., Kawatani, M., and C-W. Kim, eds., Taylor & Francis Group plc, London, 2007, 173–180.

319 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Life-cycle monitoring of the structural configuration of a suspension bridge

Ho-Kyung Kim Department of Civil Engineering, Mokpo National University, Jeonnam, Rep. Korea Ho Lee Department of Civil Engineering, University of California, Irvine, CA, USA Jeong-Hwan Jang TM-Enc., Gyeonggi-Do, Rep. Korea Young-Ho Kim Department of Civil Engineering, Dongkang College, Gwangju, Rep. Korea Sang-Kon Ro Busan Metropolitan City Facilities Management Authorit, Busan, Rep. Korea

The tension of the main cables in a suspension bridge is influencing critically the stiffness of the whole structural system and can be indirectly measured through the survey of its configuration. Since shape-finding analysis methods or results are mostly based on the idealized structural system adopted during design, need is to pay particular attention on the exploitation of field-measured data when establishing or updating the analysis model for maintenance purpose of the bridge in service. This study intends to propose an analytic model for an actual 3-span suspension bridge with main span of 500 m using the measurement resulting from its configuration survey. In order to trace the configuration of the main cable and the structural system at the early stage of design and construction, the shape-finding results and geometry control data are presented for the target bridge. Based on the shape-finding analysis results relative to the dead loads, a linearized finite displacement FE model is proposed and, an attempt to extend the measurement-based shape-finding analysis method from conventional methods by reflecting progressively the measured configuration and temperature is presented. Analysis results reveal that the current configuration of the bridge in service corresponds fairly with the target configuration. This study proposes a model updating method which could provide theoretical basis for the establishment of analysis model for suspension bridges based on various maintenance data.

Figure 1. Summary of applied hape-finding results. 320 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Bridge monitoring while partial demolition under traffic

K. Zilch, E. Penka & M. Hennecke Zilch + Müller Ingenieure GmbH, Munich, Germany U. Willberg Autobahndirektion Südbayern, Munich, Germany Th. Wunderlich & Th. Schäfer Chair of Geodesy, Technische Universität München, Munich, Germany

The Freimann overpass bridge is a central bridge element in the north part of the city of Munich. Severe damage caused by environmental influences has occurred which makes a new construc- tion necessary. A central condition for the new construction project was the maintenance of the traffic flow over the construction during demolition and reconstruction work. As the bridge has a single-component cross-sectional superstructure, the superstructure is being demolished in part. Additional temporary supports are added because of the bearing conditions and the actual bridge condition. This leads to special constructions levels which require monitoring to ensure that safety is maintained. This article contains a report on the monitoring methods and experience involved in the project. Geodetic measurements are based on the reflectorless measurement system, which was presented for IABMAS 06. But additional reflectors are performed to control the possible three dimensional displacement variations of the bridge while cutting the plates and the cross beams. In addition strain measurements are implemented to observe the forces of the temporary supports. With this assembly the rearrangements of the forces during and after the cutting are controlled. Especially the period of the cutting of the cross beams at the supports are observed with this measurements, because most rearrangements are expected then. The primarily reflectorless measurement (IABMAS 06) is executed in a three month interval to control the remaining time of the half-bridge under traffic. The aim is to detect new occurred damages of the bridge, which have influence on the deformations, like the separation of the bottom plate. The comparison and interpretation of the measured values with the before calculated values is very important for the analysis of the bridge state. For this, complex folded plates FE-simulations were made but for quick interpretations also simple beam systems are evaluated. It could be summarised, that the demolition of the Freimann overpass bridge is supported by a comprehensive monitoring program. The purpose of monitoring is to check a damaged supporting element which is supporting traffic and on which comprehensive modifications to the statical sys- tem are being realised in a manner that rules out unsafe situations in as far as possible. The measured values are evaluated through intensive computer examinations. Geodetic and mechanical measure- ments provide excellent results which can be used as a basis for statements relating to structural safety. The system behaves in the manner indicated by computationally-derived prognoses.

321 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Development and implementation of a low-cost, continuous bridge health monitoring system

Yoon-Si Lee, Brent M. Phares & Terry J. Wipf Iowa State University, USA

ABSTRACT: The Iowa State University Bridge Engineering Center developed an easy-to-use, autonomous Structural Health Monitoring (SHM) system. The primary objective of this project was to develop a low-cost SHM system that can be used to continuously monitor typical girder bridges for overloads, vehicle collisions, as well as identify gradual changes in structural behavior. These specific features were established to give bridge owners tools to better manage bridge assets and were accomplished by completing three distinct work tasks: (1) Development of live load structural analysis software, (2) Development of field data collection and analysis software that integrates with select data acquisition hardware, and (3) Demonstration of the developed SHM system. The first task involved developing the live load structural analysis software, BEC Analysis. The analytical methods used in BECAnalysis are the same to the routine stiffness method except that the member stiffness matrix and the fixed-end moments are specific functions of the variation in the member cross section. BEC Analysis is capable of analyzing a bridge girder with various boundary conditions and member geometries under various moving load conditions. One unique feature of BEC Analysis is that it allows users to easily determine maximum results (e.g., maximum moment and/or strain) at any location. In addition, it contains many convenient features that are presented through a simple interface that integrates data entry and analysis with graphical representation. These features make creation and modifications of inputs and execution of analysis quick and easy, thereby, allowing the study of multiple ‘what-if’ scenarios more feasible. In general, one may use BEC Analysis for (1) analyzing girders under moving loads, (2) computing absolute maximums in each span or at a desired location, and (3) generating envelopes of maximum moments and strains. Second, a data collection/analysis/reporting package was developed to autonomously collect, process, and evaluate the measured response of a bridge without user intervention. Significant effort has been given to develop data processing and evaluation algorithms that are based upon strong engineering principles while also taking full advantage of advanced data processing tech- niques. Improved data mining and evaluation procedures allowed the amount of saved data to be significantly reduced and evaluation reports generated in a format that is clear and familiar to bridge owners. Once the development of the SHM system was completed, the system was tested and imple- mented on a highway bridge to demonstrate and verify its general usage. The bridge selected for demonstrating the use of the developed SHM system is a 320 ft × 30 ft, three-span continuous, welded steel girder bridge. The bridge is located in central Iowa in Story County, IA carrying US30 over the Skunk River near Ames, IA. The selected sensors, data acquisition hardware and computer equipment were installed at the bridge site during the fall of 2006. The SHM system, since activated, has performed as expected and has been proven to be a stand-alone, autonomous system capable of continuously monitoring the overall performance of the US30 bridge. Benefits of the developed system include its relative ease of implementation and relative low- cost. The installation of the strain gages and cabling required no training or special equipment other than safety and normal access equipment. Excluding the communication and power equipment and research and development (R&D) costs, the system can be implemented at the cost of $8,000 to $15,000 depending on the number of sensors used. While the developed SHM system can be 322 used on any typical girder bridge, its low-cost feature makes it suitable for implementation on secondary road bridges where traditional instrumentation systems are not often feasible due to time and cost constraints. Furthermore, if the data acquisition, processing and communication system is assembled as a mobile unit, its use can be expanded to monitor multiple bridges in succession.

323 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Structural health monitoring of civil infrastructures using MEMS-based technologies

T. Miyashita & M. Nagai Department of Civil & Environmental Engineering, Nagaoka University of Technology, Niigata, Japan

For monitoring of spatially large-scale civil infrastructures, there are strong needs to install a lot of sensors. Although conventional sensors such as a servo-typed accelerometer have high accuracy, their cost prevents us from monitoring structures widely. In the research field of Structural Health Monitoring (SHM), studies relating to sensor node and/or network for civil infrastructures have been done actively. However, in general, it is difficult to obtain sensor platforms at research phase. And also, commercial products of sensor networks such as MICA MOTE dons not focus on the application into civil infrastructures. Rapid progress in MEMS (Micro Electro Mechanical System) technology produces element technologies such as microprocessor, sensor and wireless module etc. with rich function, high accuracy, low-cost and small-size. As a result, the possibility of realizing distributed sensor network

Figure 1. Newly supplied wireless LAN accelerometer using MEMS-based technologies.

Figure 2. Prototypes of developing sensor nodes using MEMS-based elements.

324 applicable to civil infrastructures becomes high. In particular, senor network using MEMS-based technologies has quite effect on monitoring of spatially large-scale bridge. This paper describes some studies relating to SHM techniques for civil infrastructures especially in bridge using MEMS-based technologies. In the first half of the paper, a new wireless LAN accelerometer based on MEMS-based technology was evaluated and applied to field measurement of bridges. In the latter half of the paper, it is aimed to develop low-cost sensor node using off-the- shelf MEMS element technology for the purpose of making possible to conduct monitoring civil infrastructures widely. Fundamental study on its development is reported.

325 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Error-resilient routing for wireless SHM powered by solar cells

Junhee Ryu, Junu Kim, Injae Yeo,Yongjin Cho & Heonshik Shin Seoul National University, Seoul, Korea

Wireless sensor networks are widely used for Structural Health Monitoring (SHM) systems in bridge environment, because of lower maintenance costs and greater scalability than wired networks. Among many issues to be addressed when using inexpensive wireless sensor nodes, the most critical are communication reliability and network lifetime. Link error-rates in a bridge environment are known to be high since wireless communication must contend with the proximity of large metal and concrete structures. This results in particularly heavy use of error correction techniques, thereby requiring additional energy for computation or retransmission and reducing the operational life of a network. We present an error-resilient routing scheme that maximizes network lifetime using solar- powered wireless networks in SHM. The proposed scheme considers both the actual and predicted availability of energy and the cost of error correction with respect to environmental energy sources. We formulate an energy model which allows for the periodic availability of environmental energy sources, in addition to battery power. This model includes a cost function which combines the energy required for error correction and that obtained from the environment. Based on the model, we propose an error-resilient routing algorithm that allows a node to send data across the link which offers the best trade-off between cost and revenue. This geographically localized routing method is highly scalable because it only uses information about the sending node and its adjacent nodes. We have evaluated the network lifetime, latency and energy utilization achieved by this routing scheme by using sensor network simulator. Simulation results show an increase in both network lifetime and error-resilience (Fig. 1).

Figure 1. Network lifetime against maximum error-rate.

326 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Correlation analysis on long term monitoring data of Donghai Bridge

Limin Sun, Zhihua Min & Danhui Dan State Key Laboratory for Disaster Reduction in Civil Engineering, Tongji University, Shanghai, China

Measurement results of bridge responses usually include effects of environmental and operational conditions such as wind, traffic, boundary condition, and mostly temperature, besides the effects of structural damage. For reliable condition assessment of bridges, it is important to distinguish abnor- mal response resulted from structural condition change and normal response due to environmental fluctuations. In this paper, the correlation analysis was carried out for long term monitoring data of Donghai Bridge (the main navigation channel bridge) in China. Donghai Bridge Linkage is about 32 km long and consists of 2 cable stayed bridges and a large number of spans of continuous and simply supported bridges. The main navigation channel bridge is a cable-stayed bridge with a main spans of 420 m. A long term structural health monitoring system, including 478 sensors distributed in 8 sectors, has been installed in these bridge spans. As a part of the monitoring system, a total number of 163 sensors were installed in the main navigation channel bridge to collect the structural acceleration, displacement, steel deck temperature, concrete deck temperature and so on. The environmental data and the structural response data monitored in 8 months were processed. The vertical frequency decreases when the temperature increases, but the correlation between two variances is no-liner. The statistical results of the continuous identified model parameters, including mean, max, min, standard deviation etc. are listed in table 3, it is concluded that variation of first 12 vertical frequencies may be 1.6∼3.2%, which is less than previous research. The correlations between the vertical modal frequencies and temperature were discussed. It can be concluded that: 1) The temperatures measured at different locations on the bridge deck have good correlation each other, indicating that the temperature sensors in the monitoring system are reliable. 2) It is possible to distinguish the structural response induced by temperature change from the entire response data by correlation analysis. The modal frequency decrease when the air temperature increases. 3) The modal frequencies of Donghai Bridge are affected not only by the temperature, but also by the structural boundary condition. 4) The temperature distribution of the cable-stayed deck is consistent with the assumption bridge design code.

ACKNOWLEDGEMENT

This research is partially supported by the Key Program of the National Natural Science Foundation of China (Grant No. 50538020), the National Key Basic Research and Development Program (973 program, No. 2002CB412709), the Cultivation Fund of the Key Scientific and Technical Inno- vation Project from the Ministry of Education of China (Grant No. 20041207001), the National High-tech R&D Program (863 Program) (Grant No. 2006AA11Z120 and 2006AA11Z109), the Research Program for Transportation of Western China of Ministry of Transportation and Commu- nication, and the Research Programs of Shanghai Committee of Science and Technology (Grant No. 03dz11003, 04dz212041 and 062112007). These supports are greatly appreciated. 327 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Low-latency routing for wireless SHM powered by solar cells

Junu Kim, Yongjin Cho, Junhee Ryu, Injae Yeo & Heonshik Shin Seoul National University, Seoul, Korea

Structural Health Monitoring (SHM) systems composed of inexpensive wireless sensor nodes for civil infrastructure such as bridges and buildings must guarantee life-long network operation and provide low latency simultaneously at high scalability. In this paper, we propose a low-latency geographical routing scheme which increases radio transmission ranges and thus reduces latencies using energy harvesting devices. First, we employ environmental energy sources such as solar cells in addition to batteries to achieve life-long monitoring of civil structures. We present an energy utilization model at a node which controls the node performance in response to available energy and keeps the node from energy starvation. Increasing distances among routing nodes reduces overall latencies; however, it requires more energy and consequently decreases network lifetime. Second, we describe a local routing algorithm which calculates possible radio transmission range of a node and selects next routing node to minimize latency according to their energy utilization models. The selection of the next node uses localized geographical routing method which only uses information about a node and its adjacent nodes so it may maintain high scalability. We consider changing radio transmission range of neighbor nodes when a node selects them as next routing node. We also describe an algorithm that determines the range of neighbor nodes called topological Knowledge Range (KR). A node periodically exchanges information with neighbors in its KR and determines the KR in response to its performance and energy utilization model. The energy utilization model at every algorithm aims to guarantee life-long network operation at low latency. We have evaluated latencies, network lifetime and energy utilization by using sensor network simulator. Simulation results show that our algorithm can reduce latencies significantly while maintaining life-long operation property (Fig. 1).

Figure 1. Average latency for different size of data per message.

328 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Health monitoring for corrosion detection in reinforced concrete bridges

A. Del Grosso & F. Lanata University of Genoa, Genoa, Italy L. Pardi & A. Mercalli Autostrade per l’Italia S.p.A., Italy

Corrosion of steel bars in reinforced concrete bridges is one of the major causes of structural degradation and maintenance cost. Early recognition of corrosion phenomena is therefore a very important aspect in structural health monitoring applications. The paper reports about recent expe- riences conducted as a joint research activity by Autostrade per l’Italia and the Department of Civil Architectural and Environmental Engineering (DICAT)of the University of Genoa in using dynamic response measurements and continuous static deformation monitoring for the identification of corrosion phenomena in simple reinforced concrete beams. Experimental studies in the laboratory as well as computer simulations have been used to test the capability of the above health monitoring approaches in identifying the insurgence of corrosion in reinforcing bars. For the laboratory studies, a set of sample reinforced concrete beams artificially damaged and corroded to known limits has been used. Instrumentation consisted of accelerometers and static/dynamic fiber optic sensing equipments. Static measurements have not been performed on the beams. The preparation of new sets of laboratory beams for the static experiments is presently under way. For the computer simulation studies, 1-D and 2-D finite element models have been used. For the 1-D model the dynamic analysis has been only performed, while for the 2-D model both static and dynamic responses can be obtained. In the paper, special emphasis is given to the signal processing techniques that have been used for damage identification. In particular, for the dynamic measurements, the use of Single Value Decomposition (SVD) and Wavelet Packet Decomposition are discussed and compared. Static deformation measurements have been processed with the Proper Orthogonal decomposition tech- nique. The experiences have shown that in simple beams the effect of corrosion at an initial stage on the bars can be accompanied by secondary phenomena that may hide stiffness degradation. While mechanical damage causes a stiffness reduction in the beam, corrosion damage shows a stiffness reduction in comparison with the undamaged beam, but a greater stiffness in comparison with the beam having the same mechanical damage. This effect has been related to the increase in bond strength for small level of corrosion. However, the two monitoring systems installed on the laboratory beams have shown to be equally reliable and able to detect small changes in frequencies due to damages. The results are in good agreement even if the comparison has been possible only for the first frequency. Even if the identification of damage is clear using all the proposed signal processing techniques, the difference between mechanical and corrosion damages is very hard to detect also because usually both types of damage are present and simultaneous in the structure. The modeling of the corrosion damage has shown to be very complex, also because of the limited studies presented in the literature. Future developments of the research are focused on the improve- ment of the 2-D finite element model for the simulation of the bond strength and on the realization of a more extensive static/dynamic measurement campaign.

329 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Investigation of the relationship between displacement and acceleration in nonlinear dynamics using chaos theory analysis

R.A. Livingston Materials Science & Engineering Department University of Maryland, College Park, USA S. Jin NDE Center, TFRHC/FHWA,Wiss, Janney, Elstner Associate, Inc. McLean, VA, USA

ABSTRACT: By definition Structural Health Monitoring (SHM) systems for civil structures require a network of sensors to measure structural responses, which strictly speaking are time his- tories of displacements. However, many sensor networks actually being installed use accelerometers rather than displacement sensors. This is due to the priority given to evaluating possible damage to the structure as a result of seismic events and reflects the standard practice in seismology of mea- suring accelerations. It is assumed that in the time domain, acceleration data can be converted into displacements simply by double integration requiring two integral constants of integration(Worden, 1990) and that in frequency domain analysis involving the FourierTransform the acceleration power spectrum can be converted to the displacement spectrum simply by scaling it by the inverse of the frequency squared. Both these approaches are based on the implicit assumption that the structural system has linear dynamic behavior. However, previous research has revealed that many types of highway bridges, such as cable-stayed bridges, may exhibit chaotic behavior. The chaos theory analysis of bridge vibrations was originally developed using displacements, and the question is whether this analysis can also be applied to seismometer data. Nonlinear acceleration data cannot be simply double integrated to produce equivalent displace- ment data because of the definition of integrability. This restriction also applies to the application of the Fourier Transform to chaotic data to produce a frequency spectrum. Chaos theory analysis avoids the need for integrals by using an analytical geometry approach in phase space with poly- nomial chaoses which are a combination of deterministic and random components. This approach produces the Lyapunov exponents, which are the characteristic invariants of a chaotic system. A mathematical analysis finds that the Lyapunov exponents can be extracted from acceleration data that have the same signs as the equivalent displacement exponents and differ by only a scale factor. These theoretical results were tested using data from an accelerometer network installed on the Bill Emerson Bridge. A very detailed 3-dimensional nonlinear finite element model of the bridge has been developed using LS-DYNA which can be used to simulate the displacements with very high fidelity. The results show a very good agreement in the normalized Lyapunov exponents between the simulations and the actual acceleration data. Comparison between the acceleration data and the simulated displacements also showed very good agreement, confirming the mathematical analysis. Investigation of the spatial distribution of the vertical accelerations of the bridge deck reveals that the locations showing the strongest chaotic behavior are at the towers. Thus, this research has demonstrated that chaos theory analysis of accelerometer data can be used to establish a base line condition index for structural health monitoring.

330 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

The application of the frequency-shifted feedback laser optical coordinates measurement system for field measurement of bridges in service

S. Umemoto, N. Miyamoto, K. Kubota & T. Okamoto Keisoku Research Consultant Co., Hiroshima, Japan T. Hara & H. Ito Tohoku University, Miyagi, Japan Y. Fujino The University of Tokyo, Tokyo, Japan

As features of the frequency-shifted feedback laser optical coordinates measurement system, due to the measurement principle employed, measurement accuracy is distance-independent and dynamic measurement at a maximum sampling frequency of 1000 Hz is possible. Based on these features, high accuracy remote/non-contact dynamic measurement in Structural Health Monitoring can be expected. High accuracy of 200 µm has already been verified in long-distance displacement mea- surements from a distance of 500 m. In order to verify the applicability of the system to bridges in service, in the present research, the authors first conducted a laboratory experiment with a vibration board (natural frequency: 10 Hz), and then performed dynamic displacement measurements of the girders of bridges in service while vehicles were passing over the bridges. The results confirmed that the system provides high displacement measurement performance under practical conditions, thereby verifying its applicability to bridges in service.

331 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Bridge-Weigh-In-Motion based on strain measurement of vertical stiffeners

E. Yamaguchi,Y. Naitou, K. Matsuo & Y. Matsuki Kyushu Institute of Technology, Kitakyushu, Japan S. Kawamura Kure National College of Technology, Kure, Japan

ABSTRACT: For constructing a good maintenance scheme of an existing bridge, it is essential to know truck weights acting upon it. Therefore, the development of methods estimating truck weights has attracted many researchers in the area of bridge engineering. Needless to say, the methods that do not disturb traffic flow are preferable. This class of method is known as weigh-in-motion. One of such methods is based on the deformation of a bridge and is called Bridge-Weigh-In-Motion (BWIM). The conventional bridge-weigh-in-motion is based on the strain measurements of main girders. However, since they are rather insensitive to traffic loads, they cannot provide some vital information such as vehicle velocity by themselves. Extra measurements are required. The authors have conducted a research on the bridge-weigh-in-motion, concluding that a two- span continuous curved bridge with skew can be used to achieve good truck-weight estimation. In this research, the strains of vertical stiffeners were measured to obtain the vehicle velocity. In the present study, an effort has been made to use those strains of vertical stiffeners so as to evaluate not only the vehicle velocity but also truck weights, applying Integration-BWIM. A good result is obtained in one bridge while large error is observed in the other bridge. The reason behind this has been identified as the influence of a truck in the adjacent lane on the strains of vertical stiffeners. 3-D FEA is then conducted, revealing such an influence varies from vertical stiffener to vertical stiffener. The selection of a vertical stiffener used for Integration- BWIM is therefore important. Further investigation of 3-D FE simulation demonstrates that the proper selection of vertical stiffeners could indeed make Integration-BWIM with the strains of vertical stiffeners valid in practice. Since the strain measurements of main girders are no longer necessary, the method can help reduce the cost of BWIM.

332 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

An FIS and AHP based on line evaluation system on Donghai Bridge

Danhui Dan & Limin Sun State Key Laboratory for Disaster Reduction in Civil Engineering, Tongji, University, Shanghai, China Zhifang Yang Shanghai Municipal Engineering Research Institute(SHMERI) Daqi Xie Donghai Bridge Management Ltd

ABSTRACT: In this paper, a FIS and AHP based online evaluation technique is developed for Donghai bridge health monitoring system since April, 2006. The Analytical Hierarchy Process (AHP) tree-like architectures was induced for graded evaluation of whole 32 km bridge. And the Fuzzy inference system is applied here to realize the evaluation computing in lowest level nodes of the tree-like architectures. Each node goes with several condition indices which can be extracted from raw data real-time. The thresholds of each index were set to an initial value obtained from structure damage and performance evolution analysis of the bridge. After one year base period monitoring, the initial thresholds system was updated with fuzzy cluster and statistic methods by the real life data from base period. The application result shows that the techniques this paper used are reasonable and reliable. It fulfills the online evaluation requirement of Donghai Bridge health monitoring system successfully, and can act as a long term online evaluation system.

1 GENERAL INSTRUCTIONS

In the last two decades there are a lot of integrated bridge health monitoring systems erected all over the world. The third generation SHMS are expected to have a capacity of utilizing the data deeply and in multiple manners, especially, the online health status evaluations are strongly pursued. From bridge owners and designers’ viewpoint, the aim of bridge maintenance and management has a top-priority for BHMS to pursue. So the judgment on bridge health status and safety status aided by BHMS is urgent need now. To migrate the present structural condition assessment approaches into bridge health monitoring system to form an online evaluation mechanism is an urgent and significative thing. In this paper, an AHP and FIS based online bridge evaluation system are developed. The technology present here is realized in the Donghai bridge health monitoring system, and act as an online evaluation subsystem of Donghai Bridge Health Monitoring System.

2 DONGHAI BRIDGE HEALTH MONITORING SYSTEM

Donghai Bridges is a linkage connecting the Luchao Port in Shanghai and the Yangshan Island deep water Port in Zhejiang Province (Fig. 1). Goals of Donghai-Bridge Heath monitoring system is performance monitoring. DHBHMS was designed and erected since April, 2006. Its 478 sensors distributed along the eight segments selected from the whole bridge. A B/S style software is realized to measure the target quantities, to monitor selected features extracted from the raw data real-time, to store and manage the data, more importantly. An online evaluation subsystem is realized, which used the raw data and the features (indices) extracted to give a set of condition evaluations organize 333 Figure 1. Donghai Bridge in Shanghai, China. in hierarchy. The AHP (Analytical Hierarchy Process) was used to organized the whole evaluation hierarchy tree, and the FIS techniques are used to do inference in each knot of the tree. Scores in each knot are transfer to the above level of the tree, finally, a ultimate score can be get and a report can be generated automatically to bridge owner, which help the owner to do a decision-making.

3 AHP AND FIS BASED ONLINE EVALUATION SUBSYSTEM

Figure 2. The AHP based online evaluation tree of DHBHMS.

4 THRESHOLDS SYSTEM EVOLUTION AND UPDATING

The initial thresholds system above gives an apriority division of the feature pattern space. The initial threshold must be updated and improved based on the real monitoring data in the baseline period. Two ways are used here to fulfill above mentioned tasks, one is fuzzy clustering, and the other approach is statistical pattern recognition. In order to validate the performance of updated online evaluation system according above mentioned approach, a compare investigation are conduct between before and after updating.

5 CONCLUSION

In this paper, a FIS and AHP based online evaluation technique is developed. The analytical hierarchy process (AHP) tree-like architectures was induced for graded evaluation of whole 32 km bridge. The Fuzzy inference system is applied here successfully to realize the evaluation computing in lowest level nodes of the tree-like architectures. The initial thresholds of each index set to an initial value obtained from structure damage and performance evolution analysis of the bridge fulfill the need of first year online evaluation in DHBHMS, but it must be updated by real monitoring data 334 based feature patterns and correspond threshold system. Fuzzy clustering approach can help to give a reasonable feature space division. The case study of DHBHMS shows its validity. the technique developed in this paper can use to updating a online FIS system in DHBHMS.The modal parameter based feature pattern like modal flexibility matrix should be the important evaluation item. The application result shows that the techniques this paper used are reasonable and reliable. It fulfills the online evaluation requirement of Donghai Bridge health monitoring system successfully, and can act as a long term online evaluation system.

ACKNOWLEDGEMENT

This paper is supported by Shanghai S&T committee project ‘062112007’. Ministry of Com- munications of P.R.C, ‘West communications Science and technology project 2004’; Ministry of Communications of Hebei Province, P.R.C, ‘Research on Jing Qin highway bridge health monitoring technology’.

335 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Structural health monitoring of complex structural systems using adaptive models

S. Arangio University of Rome “La Sapienza”, Rome, Italy

The theoretical framework for the design of complex structural systems should be based on a comprehensive evaluation of all the performances. Aim of Structural Engineering is not only to achieve an ideally good design and a nominal construction but also to assure the long term exploitation of the system. In this perspective, especially in the case of complex structural systems, the circumstances that may eventually lead to deterioration, damage and unsafe operations may be diagnosed and mitigated in a timely manner, so that costly replacements can be avoided or delayed by effective preventive maintenance. Analyzing this aspects in terms of the expected payoff, comes out that, especially in the case of complex structural systems, the Structural Health Monitoring assumes a primary role in conjunction with the integrity assessment and safety evaluation of structures. The monitoring process should be planned since the design phase and should be carried out during the entire life in order to assess the structural health and performance under in-service conditions. Considering the large quantity of data coming from the monitoring process, an efficient mon- itoring system needs the development of methodologies able to extract useful information from such measured data. In this work, the applicability of adaptive models, like the neural networks, for the analysis and interpretation of this kind of data is investigated. The neural network model is used to propose a hierarchical damage identification strategy for a long suspension bridge. This includes two steps. In the first step various neural network models, referred to the different measurement points, are trained using the response time histories of the structure subjected to wind and traffic excitation in the undamaged situation. Afterward, new inputs corresponding to different damage scenarios have been proposed to the trained network models. Analyzing the increment of the error in the approximation of the signals coming from the various scenarios respect to the error in the approximation of the signal coming from the undamaged situation it is possible to detect the occurrence of damage and to identify the affected section. The second step of the proposed method aims at the identification of the specific damaged element in the section (hanger, cable or transverse) and the quantification of the extent of such damage. In this step a pattern recognition approach is used. A neural network is trained using as inputs the values of the errors in the approximation of the vertical displacement time history, due to traffic excitation, in three points of the identified section, and as output a vector including the possible locations of the damage and their relative extents. After training, the network is tested with some patterns not included in the training set. One can see that the location can be detected in 90% of the cases, whereas the intensity in 66% of the cases. The proposed method can be useful for damage identification of large structural systems instru- mented with on-line monitoring systems because they furnish measurements of the structure in both damaged and undamaged situation and, using the neural network model, no other information on the structural model is needed.

336 Information technology for lifetime management of bridge

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

A Strategy for IT-based lifetime management of bridge

S.-H. Lee, B.-G. Kim, H.-J. Kim & S.-J. Kim School of Civil and Environmental Engineering, Yonsei University, Seoul, Korea

Information fragment is one of the most urgent issues to be solved in lifetime management of bridge. Consistency, reusability, and integrity of infrastructure information must be assured for reliable decision making during the lifetime. In this study, a strategy for information and knowledge management of bridge is provided on the basis of multi-layer concept. The multi-layer concept supports object-oriented and integrated information management based on the classification system of bridge elements. Business activities conducted during the lifetime and information flow between them are identified; information requirements are specified as a set of units of functionality. STEP and XML are employed to develop 3-D shape-based information model and document information model, respectively. The proposed concept and framework were verified by applying it to cable-stayed bridge and steel box girder bridges (see Fig. 1). The two information models such as STEP-based bridge information model and XML-based document information model of structural calculation sheets were introduced as example of standardized information layers. The concluding remarks are as follows. The distributed data in different application data set can match each other by using the same nam- ing rule for structural component objects. The data consistency check between STEP-based bridge information model and XML-based document model was successfully executed. The proposed concept of multi-layered database, therefore, can be effectively used for integrated information operation of lifetime bridge management. Even though GIS systems recently developed show the possibility of integrated operation of civil infrastructure information, the details of information layers for each infrastructure should be defined. This study selected a topological classification system and information models for that functionality.

Figure 1. 3-D shape based bridge information model by using CAD module.

339 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Use of information technology in a regional bridge management contract

R. Kiviluoma WSP Finland Ltd, Helsinki, Finland M. Tervo Finnish Road Administration, Oulu, Finland

ABSTRACT: Need of improved utilization of Information Technology (IT) in infrastructure maintenance has recently risen in a pilot project employing new procurement methods of bridge management in Finland. In the first stage of the project, maintenance of 674 bridges on a selected road region of the Finnish Road Administration are privatized in a sense, that the private service- provider consortium takes care of all works; including bridge inspections, maintenance design, maintenance works and programming of forthcoming works. This paper describes experiences obtained from the utilization of IT during the first two years of the pilot project, and discusses further possibilities on exploitation of IT in bridge management.

1 INTRODUCTION

Majority of bridges in Finland, comprising about 20,000 bridges total, are owned by Finnish Road Administration (Finnra). Finnra is responsible of administrative tasks and procures all works from open markets. Finnra uses various contracting practices in procurement including design- and-built, life-cycle contracts and service agreements. In 2007 Finnra has initiated a pilot project named SILTOPA dealing with maintenance of bridges as regional bases. Bridges at about one-third regional coverage of Finnra’s Oulu Region have been privatized for five years. Alike projects are to be used in other road regions as well, if experiences reveal to be promising. SILTOPA has been started with planning the principles and preparing bidding documents. In that work it, has been necessary to reproduce bridge data and documents in electronic form. Using a specific contractual clause, service provider is encouraged to use and develop advanced technologies including ground-penetrating radar, laser scanning and sensor technologies. This is seen as tool to improve quality of the input used in decision making, especially for those parts and issues on structures, which are not seen by eye. SILTOPA’s first stage 2007–2012 has started with the 1st author working at service provider’s and the 2nd at client’s side. Ground penetrating radar and laser scanning have been applied to several bridges. Sensor technologies are to be implemented by means structural monitoring systems.

2 CONCLUSIONS

– Privatization of a large bridge group needs efficient utilization of IT to facilitate information change between the service provider and the client. 340 – In SILTOPA, Finnra’s bridge management system, comprising a bridge register and expert soft- ware, is referred in contractual use. Performance of the service provider is ranked based on the development of sum of damage points of bridges. – Long-term bridge management contracts may encourage service provider to develop and apply advanced new technologies to give benefit for both parties. IT-related development appears, however, be strategic in nature and needs multiple years, multiple research projects, and probably multiple bridge management contracts as well, to reach full efficiency.

341 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Development of a structural health monitoring system with wireless sensor networks

Hisao Emoto Practical Maintenance Engineering Institute, Yamaguchi University, Japan Ayaho Miyamoto & Kei Kawamura Graduate School of Science & Engineering, Yamaguchi University, Japan

Structural health monitoring involves the maintenance of structures to reduce repair and mainte- nance costs. Health monitoring is always performed by attaching a sensor to the structure. The purpose of this technology is to use the collected data to identify structural damage and deteriora- tion. Furthermore, the health of the entire structure can be diagnosed. The development of a health monitoring system consists of three phases. Phase 1 is the sensing phase, phase 2 is the information phase, and phase 3 is the analytic phase. Several studies have investigated wired sensing with respect to structural health monitor- ing. However, cables and the costs associated with their installation are expensive. Therefore, the application of monitoring systems is limited to important structures. It is impossible to monitor several structures. Consequently, structural health monitoring systems are not popular. On the other hand, the development of wireless sensor network technology is being actively pursued in order to establish a ubiquitous network society. Installing a wireless sensor net- work involves setting up several devices for wireless networking, processing, and sensing in space. A wireless sensor network continually collects physical data, such as temperature, light, sound, acceleration, and magnetism. This technology is expected to be applied to structural health monitoring in order to reduce the costs of cable and installation. An ad hoc network is a network architecture that communicates data to each node by wireless autonomy. Wireless sensor network systems are cheaper than conventional non-wireless systems. These sensors are thus able to configure several sensors over a wide area. Nodes on real structures can always be observed. Wireless sensor networks are currently still under development, and there are problems associated with power resources, network architecture methods, data transfer techniques, node synchroniza- tion, accuracy of sensor devices, and CPU performance. In other words, it is necessary to develop a sensing system that has appropriate specifications for a structural monitoring system. In the present study, a highly versatile basic wireless monitoring system for the structure is developed that is easy to set up and that can monitor static temperature and strain data using MICAz-MOTE, a wireless sensor network device produced by Crossbow Technology, Inc. This system was developed by taking the following two points into consideration:

1) Proposal of a network structure specialized for structural monitoring. 2) The development of a highly accurate strain sensor that can be used in a wireless sensor network device.

In addition, the problems that occur when a wireless sensor network is used in a structural monitoring system were revealed by verifying and testing this system. Consequently, a multi-hops network system of the hybrid-star type is initially proposed as a network structure. The wireless network can be used for wide-area connection and the number of 342 correspondence for each node is less than that of ad-hoc network correspondence. Next, a bridge- circuit was produced to measure the strain on a wireless sensor node. As a result of doing this, it was found that noise from outside became problematic. Finally, this system is expected to be applied to structural health monitoring.

343 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Development of a multi-purpose remote health monitoring system for existing bridges

A. Miyamoto Graduate School of Yamaguchi University, Yamaguchi University, Japan J. Sonoda & K. Kawamura Graduate School of Science & Engineering, Yamaguchi University, Japan

ABSTRACT: In Japan, 40% of civil infrastructures, such as bridges, tunnels, etc were constructed in a period of accelerated economic growth prior to the 1970s, and there is a concern that a large number of aging bridges will need to be repaired or replaced in the near future. Therefore, effective maintenance of bridges is needed. The current maintenance operations for bridges in Japan are primarily on-site visual inspections to evaluate bridge performance and then formulation of a repair plan based on the inspection. However, visual inspections are influenced by the experience of the inspector; however, not only are inspections subjective, but there is also the concern that the decrease of the working population will cause a shortage of maintenance engineers in our country. Given that numerous other countries have similar problems and also need efficient bridge maintenance, bridge health monitoring systems for efficient and quantitative bridge maintenance are being studied in a variety of countries. Bridge health monitoring system via information technology is capable of providing more accurate knowledge of bridge performance characteristics than traditional strategies. This paper describes a multi-purpose remote health monitoring system for existing bridges by using “CompactRIO”, a compact measurement device which can connect various sensors by exchanging modules. In general, the remote health monitoring system for bridge performance evaluation consists of measurement instruments or PCs that collect, process, and storage measurement data near the bridge, and Web server that collects the monitoring data via Internet. In this study, we developed a multi-purpose remote health monitoring system using LabVIEW for rapid programming. And, we adopted CompactRIO for development a measurement environment having telecommunications function. CompactRIO is a compact measurement instrument and it enables to change sensors by exchanging modules, is equipped with a function of Web server. In the result, using the Web publishing tool that is function of LabVIEW, we can make HTML files for publishing measurement screen on Web easily. By saving this file in a CompactRIO for Web server, users can access CompactRIO directly, and monitor from remote location via Web browser. Furthermore, it is verified by experimental results with a bridge model. The system enables to change kinds and number of sensors freely, and to provide remote monitoring environment.

344 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Automated identification of modal properties in a steel bridge instrumented with a dense wireless sensor network

A.T. Zimmerman, R.A. Swartz & J.P. Lynch Department of Civil & Environmental Engineering, University of Michigan, Ann Arbor, MI, USA

Wireless sensing technology has made it possible to instrument large civil structures, such as bridges, with dense networks of wireless sensors. Because many wireless sensing platforms inte- grate data processing capabilities with high-precision analog-to-digital converters, it is possible to autonomously process sensor data without the need for a central data server. By employing parallel processing techniques on typically serial modal analysis methods, a network of sensors can be used to autonomously determine modal properties (modal frequencies, damping ratios, and mode shapes). In this study, three output-only modal identification techniques which have been modified for parallelization by Zimmerman et al. (2008) are implemented for the first time within the Narada wireless sensing platform. These methods are the simple Peak Picking (PP) method, the Random Decrement (RD) method, and the Frequency Domain Decomposition (FDD) method. A network of 16 Narada wireless sensors is then deployed on the Bandemer Park pedestrian bridge in Ann Arbor, MI. Over the span of several vibration tests, each of the distributed modal identification algorithms are utilized to determine the global structural modal frequencies, mode shapes, and damping ratios of the bridge. In civil engineering, modal properties (mode shapes, modal frequencies, and modal damping ratios) have been widely used to assess structural performance, calibrate analytical design models, and (in cases of severe damage) detect and locate damage in the wake of natural disasters such as earthquakes. In this case, however, these modal analysis methods are used simply as an illustration of the possible autonomous damage detection capabilities of an intelligent wireless sensing network. For validation, the results obtained from the automated identification techniques are compared with similar modal analysis methods run off-line using time history data recorded by the network of wireless sensors. It can be seen that these embedded techniques produce modal parameters that are comparable to those obtained using traditional offline analysis techniques.

Figure 1. Embedded results for one experimental test: (a) Peak Picking frequencies, (b) Random Decrement frequencies and damping ratios, and (c) Frequency Domain Decomposition mode shapes.

345 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Downsizing seismic sensing system and its implementation

Y. Mizuno & Y. Fujino The University of Tokyo, Japan CREST, Japan Science and Technology Agent

1 INTRODUCTION

Conventionally, visual inspection plays an integral roll of structure management. In recent years, several bridge collapses were happened and had called attention to the existing inspection methods. For example, the I-35W Bridge in Minneapolis, Minnesota fell down, in which motion pictures were broadcasted world widely (Mn/DOT 2008), although annual inspections have been operated. The collapse implies the needs for the accurate and quantitative bridge management methods in order to prevent similar events. This paper prototyped a networked sensing node in the gap between traditional wired systems and WSN solutions. The prototype is focused on seismic response measurement of buildings, which means that it can be powered by normal outlet, and connected through existing network, although it may require some wireless routing.

2 SYSTEM INTEGRATION

The prototype of the portable sensing system for buildings (ps4BUILD) is shown in Figure 2. It consists of a 3-axis accelerometer, ADC (Analog/ Digital Converter), and an ultra portable PC. The accelerometer adopted in this paper is manufactured by Silicon Designs. The 2442-002 is sensitive to measure at the level of 10−3 G (10−2m/s2). The sensitivity is enough to detect relatively small earthquakes, which people can sense, and they happens once or twice in a month in Tokyo. The ADC USB-6009 is manufactured by National Instruments, and has 14-bit resolution. The combination of these accelerometer and ADC requires only one USB cable for power supply and data acquisition. The PC VGN-UX50 is an ultra portable PC with enough capacity of operating Windows XP Professional.

3 INSTALLATION

As of the end of December 2007, four building have been monitored. Four sensor nodes were installed in each building basically. Two of them are in the bottom floor, which detect ground motions, and the other two are in the higher floor, which measure structural behavior during small earthquakes

4 DATA INTERPRETATION

As retrofitting proceeded, the dominant frequency of strong axis at ERI2 got larger, which means that the stiffness increased. POD (Proper Orthogonal Decomposition) analysis have been performed in order to identify the structural behaviors. The result of retrofitted ERI2 could be interpreted differently. Though the fact that the torsional mode appears in the lower frequency requires more detail investigation, it indicates that the sensing driven approach reveals new findings. 346 5 SUMMARY

The PC-based seismic sensing node “ps4BUILD” has been prototyped. It can detect small earth- quakes that happen once or twice in a month in Tokyo. The installation to the building under retrofitting revealed the process of stiffness increase as retrofit works proceeded. It also implied the sensing driven approach may make new findings beyond existing structure engineering.

347 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

SHM sensor networking with remote powering and interrogation

Michael D. Todd, David Mascarenas, Eric Flynn, Ben Lee, Kaisen Lin, Daniele Musiani, Tajana Rosing, Rajesh Gupta, Samori Kpotufe, Daniel Hsu & Sanjoy Dasgupta University of California San Diego, La Jolla, CA, USA Gyuhae Park, Kevin Farinholt, Matt Nothnagel, Chuck Farrar Engineering Institute, Los Alamos National Laboratory, Los Alamos, NM, USA

ABSTRACT: Post-event (e.g., earthquake) assessment of structural systems such as bridges is a critical link in the path towards returning the system to safe and reliable service. The field of Structural Health Monitoring (SHM) is an integrated paradigm of networked sensing and actuation, data interrogation (signal processing and feature extraction), and statistical assessment (classifi- cation of damage existence, location, and/or type) that attacks structural health assessments in a systematic way. The first “line of defense” in an SHM system is the sensor network architecture itself, which must meet any of a number of challenges that the application (bridge) itself brings. Conventional wired networks have been the dominant paradigm, with individual sensing components typically stand alone and are connected via conductive cabling to a centralized data processing and multiplexing unit. Each sensor is effectively independent of other sensors in the network—each sensor has its own cabling—and controlled synchronized interrogation of the entire network is achieved only through the central unit. More recent research and development has focused on wireless topologies, where centralized control, power, and computing is ceded to local nodes consisting of, generally, a micro- controller, a processor, sensor interface(s), a radio, and (usually) a battery for power. In some recent work, “mobile agent” design has been implemented with wireless networks, where data stay unpro- cessed at each local node, and the integration, processing or fusion code is transported to the data. However, power remains an issue. An entirely different mobile agent approach is possible by allowing the agent itself to be the communications network and power supply. This work considers the possibility of coupling Radio Frequency (RF)-interrogated sensors with a UAVthat is remotely- piloted to move to the sensor target field, identify and locate sensors, power and interrogate each sensor in turn, and perform necessary local computation and storage in accordance with overall task goals. A sensor node with both an analog-to-digital and a capacitive-to-digital interface, anAD5933 impedance measurement chip, an ATmega 128 L microcontroller, and a 2.4 GHz XBee radio for wireless communication was developed to communicate with an RF transmitter/receiver. A super- capacitor is used to store charge, and up to 35 mW of instantaneous power may be delivered. Peak displacement sensors (using capacitive transduction mechanisms) and bolt preload sensors (using impedance measurements) were designed to partner with the node. A helicopter UAV was outfitted with an RF transmitted, and a prototype of this sensor node was deployed on an out-of-service bridge near Truth-or-Consequences, NM, USA. The main advantage of this overall approach is that power does not have to be embedded with the sensing system, but transported to its vicinity (via the UAV) and then wirelessly transmitted to the sensor node. It is anticipated that such a sensor network will have improved reliability and will have inherent advantages when monitoring must be performed over long periods of time in locations that are physically difficult to access or pose post-critical life-safety issues. It is also anticipated that these systems will evolve to hybrid systems that couple local energy harvesting at the node level with the RF energy delivered by UAV.

348 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Development of an advanced inspection system for weathering steel bridges based on digital image recognition

S. Goto Ube Machinery Corporation, Ube, Yamaguchi, Japan T. Aso & A. Miyamoto Graduate School of Science & Engineering, Yamaguchi University, Ube, Japan

Non-painted bridge with weathering steel (weathering steel bridge) is a representative steel bridge that will be used in minimum-maintenance. While dryness and moisture repeating appropriately in the atmosphere, weathering steel can form adherent rust that has the quality for corrosion speed becomes slow. Therefore, if appropriate use methods and maintenance are administered, it will be the superior steel which enable bridges to be used semi-permanently in non-painted. Maintenance work in order to verify the evaluation of rust conditions and the corrosion-proof per- formance could not be missed. The popular NDT methods such as ion transfer resistance method, electrochemical potential method, X-ray diffraction method, spectroscopic identification tech- niques are employed to measure the corrosion behaviors from the point of view of corrosion loss, and chemic, electric and magnetic properties of the rust patina. At present, inspection and evalua- tion methods of rust cannot ignore the influence of the sediment besides rust, and the system that evaluates a lot of inspection results comprehensively is not established. Therefore, there are many cases where the exterior inspection that designates the standard photograph as basis is attached importance, but the visual inspection is inferior to quantitative and objective characteristics, and the dispersion of evaluation by the inspection person is wide. The evaluation of rust conditions is an important process for the maintenance of weathering steel bridges and influences the mainte- nance strategy in the future. The development of the new evaluation method that has high reliable, quantitative, and efficient characteristics is necessary. Recently, digital image recognition has been used experimentally in the automated system for bridge inspection, pavement management, and sewer pipeline inspection. The critical issues of this approach contain imaging-based feature extraction and pattern classification algorithm. The approaches for extracting textural features are very diverse, traditional approaches such as Gray Level Co-occurrence Matrix (GLCM) methods, transform-based approaches based on Fast Fourier Transform (FFT) and Wavelet Transform (WT). The WT analysis has shown better performance in many cases due to its outstanding capabilities in space-frequency decomposition at different scale. Another crucial issue is the pattern classification, numerous classifiers using the extracted features have been employed, including nearest neighbor classifier, Bayes classifier, probabilistic neural network, and learning vector quantization. Recently, the support vector machine (SVM) is becoming a successful method for pattern recognition. In this paper, we propose the advanced inspection system based on image processing and SVM for weathering steel bridges. Three feature extraction; GLCM, FFT and WT characterize a rep- resentative set of 558 images, and SVM is used to classify a set of texture features for various corrosion levels. Comparison of each accuracy shows the classifier trained by Wavelet feature outperformed the others. Investigating the performance of the wavelet texture features in different decomposition levels, the variation appears to be very small in the case of levels 6, 7 and 8. Hence, it can be thought that the optimum decomposition level is 6 for identification. They are high accurate classifications, and the proposed inspection system can be realized efficiently.

349 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

A study on the self-anchored suspension bridge behavior using GPS

H.J. Ham BT Consultants Co., Ltd., Seoul, Korea S.H. Oh Leica-Geosystems Korea, Seoul, Korea I.H. Bae & G.H. Ha New Airport Highway Co., Ltd., Seoul, Korea

To measure deflection of a long-span bridge such as the self-anchored suspension bridge, a laser deflection gauge is used in former days. However, it has defects that expand the error range in case of long measurement distance or foggy weather. So measurement of using GPS is expected to become commercialized in order to measure a super-long span bridge in the future. Accordingly, on this study, a load test was carried out on purpose to proof the availability of the deflection measurement using GPS compared with the shape survey using total station and the deflection measurement using a laser deflection gauge. A 921 tons weight of 12 trucks in the upper road and 2 6-car trains in the lower railroad is loaded on the bridge in four different cases. The results coincide in the structural analysis data. So we have evaluated that measurement of using GPS is more profitable than other devices, as the span of long-span bridge like a suspension or cable-stayed bridge become longer in the future and as the GPS is almost not influenced by the weather conditions such as a fog.

Figure 1. Deflection per load case.

350 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

SHM role in Bridge Life Cycle Analysis (BLCA)

Sreenivas Alampalli New York State Department of Transportation, Albany, NY, USA Mohammed Ettouney Weidlinger Associates, Inc., New York City, NY, USA

The two main objectives of bridge management are ensuring safety of users, while maintaining or reducing costs of ownership. Decision making tools are often used by bridge owners to help in reaching these objectives. One of the often used techniques in asset management is life cycle analysis. In particular, one of the tools that have been utilized extensively is the Bridge Life Cycle Cost Analysis (BLCCA) techniques. We argue that even though the accurate knowledge of BLCCA is necessary, it is not sufficient for accurate decision making regarding life cycle of the asset. There are two other important parameters that are needed for accurate decision making. First is what we will define as Bridge Life Cycle Benefit Analysis (BLCBA). BLCBA is needed in order to decide on the relative cost benefit of any endeavor, including life cycle analysis of the bridge. Moreover, we argue that an accurate estimation of the service life, of life span of the bridge is needed, in order to accurately estimate both BLCCA and BLCBA. We define this as bridge life span analysis BLSA. We now define a more general metric: Bridge Life Cycle Analysis (BLCA) which includes BLCCA, BLCBA, and BLSA. One of the major problems that many bridge officials have confronted while estimating BLCCA (and also BLCBA and BLSA) is the lack of information that is needed for accurate results. We argue that Structural Health Monitoring (SHM) can provide excellent vehicle in estimating BLCA, and all of its subcomponents. In the remainder of this paper we will show specific steps of utilizing SHM to help in estimating accurate BLCA. The paper will then introduce two modes where Structural Health Monitoring (SHM) interact with the BLCA process. Following on of these modes can help decision makers in achieving their safety and cost optimizations goals. Two examples are given for each of the two BLCA- SHM interrelationship modes are given. More studies are needed to further explore the steps of the process. Also, since the applications of the process depends on the type of cost and/or benefit, specific studies are needed to produce simple and detailed guidelines for users and decision makers.

351 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Interoperable information model based on IFC for the cable-stayed bridge monitoring system

J.-H. Yi, H.J. An, H.-J. Kim & S.-H. Lee School of Civil and Environmental Engineering, Yonsei University, Seoul, Korea

To handle the various integrated and related information for the safety management of infrastruc- tures on road networks such as bridges, tunnels and artificial slopes in real time and to make the right decision just in time in emergency, the well-organized and integrated database for monitoring and pre-warning system are inevitably required. The integrated database system based on infor- mation technology is expected to meet the requirement thoroughly. As the first study to satisfy the demand, this paper proposes both product and process model for safety management of steel bridges on road networks. The study has started with analyzing the safety management process including sensing, monitoring, data analysis and decision making process. The analyzed safety management process is re-structured and electrically represented by IDEF0 which is a method designed to model the actions, activities and decisions of a system. Then, the product model of safety management information such as sensing, monitoring, and signal data analysis is built up with EXPRESS language, which is a data description language of Industry Foundation Classes (IFC) model. Measuring items for steel bridges covering strain, displacement, acceleration, and vibration are also identified. The location of measuring point, especially, is electronically expressed on the three dimensional geometry of critical bridge members in order to assist end-user’sintuitional access and perception. Finally it is expected that the information model developed can be linked into the design and structural analysis information model based on three dimensional geometries of steel bridges proposed previously.

352 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

An automatic crack recognition system for concrete bridge inspection by image processing approach

Ayaho Miyamoto Graduate School of Science & Engineering, Yamaguchi University, Ube, Japan

ABSTRACT: Infrastructure systems including numerous concrete structures in the world have to be prudently managed for balancing the safety, economy, and sustainability requirements, then, maintenance of such infrastructure systems has become a major social concern. As a means of inspection for maintenance, visual inspections are generally carried out to draw a sketch of defect. The records of the degree of defect vary depending on the judgment of the inspector. Poor skills may lead to faulty diagnoses. At present, therefore, advanced tools for visual inspection are required. In this study, a prototype system was developed that could ultimately help automatically draw a sketch of defect. The defect sketching support system automatically identifies the defect on concrete surfaces. Once the defect sketching support system has been put to use on structures, drawing a sketch of defect will be made possible by simply using the photographs that are taken using a digital camera. Thus, the efficiency of inspection will be increased, inspection accuracy and reliability enhanced. Computer processing will then enable appropriate evaluation of the condition of the structure. This paper describes the method for measuring defect using the images taken by a digital camera. The authors are currently concentrating their efforts on the study of “cracking", the most important type of defect of concrete structures. This paper proposes a method for measuring the crack width, one of the parameters for assessing deterioration of a concrete structure. The crack width measured using a gauge is compared with the value measured by the system. Then, the effectiveness of the system for measuring crack widths is verified. Cracks 0.2 mm wide or larger are considered to be detrimental to concrete structures. In this study, therefore, a prototype system is developed to accurately calculate the widths of cracks 0.2 mm wide or larger by employing image processing technology. The major results of this study are described as follows:

(1) The difference of brightness in cracked (target) and uncracked (background) areas was obtained in an image of crack of a concrete structure. A method was proposed for calculating crack width from the difference of brightness. (2) No unique crack width can be obtained even where the difference of brightness is uniform, because of the angle of view, brightness of concrete surface, the angle of the crack to the horizontal direction of the image or other factor. Certain correction is therefore necessary in crack width calculation based on the difference of brightness. In order to reflect varying conditions, pseudo cracks were represented by lines on a drawing paper. The effect of the pseudo cracks on the difference of brightness according to the photographing condition was investigated to deduce a correction equation for calculating crack width from the difference of brightness. (3) The crack width calculated by the system was compared with the crack width obtained by manual measurement to verify the effectiveness of the system for crack width calculation. As a result, it was found that crack width could be calculated with few errors at an angle of view of 1.0 m or less. The errors of crack width calculation by the system were allowable and the system proved satisfactorily effective while conventional manual measurement also involves errors.

353

Innovative construction technology

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Erection of asymmetric pylon table and geometry control of Machang cable-stayed bridge

Hyuntai Lim, Moshe Kim & Juwon Seo Institute of Hyundai Engineering and Construction Technology, Korea

The Machang Bridge over Masan Bay was planed to link Masan and Changwon City. The Bridge is divided into three structures; West approach bridge, Main cable stayed bridge and East approach bridge. The Main Bridge is 740 m long and consists of 400 m cable-stayed central span with 2 side spans of 170 m each. It has a composite deck and is supported by reinforced concrete pylon with 164 m height. The construction was commenced in April 2004 and will be finished in 2008 as a privately financed project. The main cable-stayed bridge has 21 m width deck of steel plate girders and reinforced concrete composite boxes at the side span pier table segments. The deck is supported by two reinforced concrete portal frame pylon with vertical legs from which radiate four planes of 30 prefabricated wire strand (PWS) cables in a modified fan formation, making a total of 120 PWS cables. The erection of 50 m pier table segment was erected with asymmetric sequence with floating crane, which was supported by precise erection planning and construction analysis. The other segments at mid and side span will be erected by balanced cantilever erection methods. A general erection cycle of 13 days consists of steel structure erection and composite construction works. A pylon table of composite cable-stayed bridge is installed using temporary bracket or temporary cable generally and a steel girder of pylon table is set up at first and concrete slab is cast-in-situ to decrease self weight. The pylon table is asymmetric and the slab was fabricated in working yard, and was installed by floating crane of 3000 tonf capacity without temporary bracket and temporary cable. During the planning stage for the erection of pier table with asymmetric sequence, the stability of structure was precisely checked. And the erection error at the pier table would not be corrected easily and it will be propagated to the future segment erection. The cantilever erection method of a composite segment is more complicated than that of steel cable-stayed bridge and the working load seriously affect the member forces and geometry. It was reflected in detail in the construction analysis and geometry control of the Machang cable-stayed bridge that the site- condition, construction sequence and structural characteristics changed from the design step for the stability of pylon table erection, the estimation of behavior, the construction error minimized.

Keywords: Asymmetric erection, construction management, cable-stayed bridge.

357 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Experimental study on the stability of temporary support for girder construction

K. Ohdo, S. Takanashi & H. Takahashi National Institute of Occupational Safety and Health, Tokyo, Japan

ABSTRACT: When constructing or reconstructing bridge girders, a temporary structure called a sandle is often used as a support. The sandle is composed of many stacked steel H-beams, each with a width and height of 150 mm, in a double cross. The load on the sandle comprises the vertical load due to the weight of the bridge girders, and the horizontal load due to the launching erection. Although the strength against vertical loads is considered when designing the sandles, the horizontal stability against the horizontal load has largely been judged based on the experience of the construction workers. In recent years, 2-edge girders have been widely used in steel bridges to reduce the cost, but fewer girders means not only a greater girder height but also higher sandles to support them. When sandles are higher than 5 m and conventional construction methods that rely on workers’ experience are used, the horizontal stability of the sandles could constitute a risk as skilled workers are decreasing in number. However, few studies have focused on the horizontal stability of sandles through experiments and analyses. Therefore, in this study we conducted experiments which involved applying vertical and hor- izontal loads to actual sandles stacked in a basic arrangement to obtain fundamental data on the horizontal stability of the sandles. In the experiments, vertical and horizontal loads were applied to sandles measuring in height from 1 m to 4 m for a single sandle, and 3 m and 5 m for the twin sandles. The twin sandles were connected to each other by steel angles and braces. All of these conditions were decided based on the opinions of on-site engineers and considered to be similar to the conditions at actual construction sites. The vertical loads increased in increments of 500 kN from 500 kN to 3000 kN and the horizontal load, which was 5%, 10%, and 20% of each vertical load, was applied to the top of the sandle to examine the maximum vertical load that could be applied to the corresponding horizontal load. The results are summarized as follows: 1) The 2-m single sandles could bear the horizontal load of 10% until the vertical load reached 3000 kN. 2) However, the 3-m single sandles could not bear the horizontal load of 10% until the vertical load reached even 2000 kN, and the 4-m single sandles could not bear the horizontal load of 10% even before the vertical load reached 500 kN. 3) When considering seismic load, the 2-m sandle and 1000 kN vertical load appears to be the limit. 4) On the other hand, the twin sandles connected to each other were stable at the height of even 5 m against seismic load. 5) Therefore, it is concluded that single sandles should be connected to each other when the sandle height exceeds 2 m in consideration of the limits of the combination loads.

358 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Design of Cheong-Poong (steel-concrete hybrid cable-stayed) bridge

D.-H. Yoo & J.-S. Ko ENVICO Consultants Co., Ltd., Seoul, Korea J.-G. Paik Daelim Industrial Co., Ltd., Seoul, Korea

‘Cheong Poong’ Bridge is cable-stayed bridge, which has main span of 327 meters crossing over the lake ‘Chung Ju’, Korea. After careful investigation of site conditions, the bridge was planned with span composition of 57.5 m + 327 m + 57.5 m, where the unbalance of the span composition and the uplift reaction due to the unbalance have to be carefully treated during detail design stages. Hybrid construction, that is, the combination of steel girder with concrete deck slab and concrete girder for the bridge girder was applied and proven to be very efficient in such an unbalanced arrangement, in addition to the extra pier support at side span and cable spacing variation to provide the uplift reaction countermeasure. Under this kind of superstructure composition, superstructure erection in center span was possible with normal truck crane that can travel over the side span superstructure built up by full staging method and special equipment such as derrick crane became to be unnecessary, therefore. Hybrid connection detail has been studied based upon the member forces history acting on the connection throughout erection stages and after completion. Wind tunnel test has been carried out to figure out wind vibration characteristic of the bridge during erection and after completion and from the test results it was decided to attach ‘extension flip’ at the both sides of the superstructure in main span to mitigate the vortex and torsional flutter.

Keywords: Hybrid construction, Cable stayed bridge, Uplift reaction, Cable arrangement, Wind tunnel test.

Figure 1. Graphic simulation of Cheong-Poong bridge.

359 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Influence of initial imperfections on stability of temporary support for bridge girder

H. Takahashi, K. Ohdo & S. Takanashi National Institute of Occupational Safety and Health, Tokyo, Japan

ABSTRACT: In the launching erection method for the construction or reconstruction of a bridge girder, a temporary structure called a sandle is often used. The sandle is composed of multiple stacked steel H-beams, each having a width and height of 150 mm and hereafter referred to as a “sandle member”. The sandle can become unstable by increasing height. However, the sandles are distributed throughout many construction sites because of the following advantages: the structure is simple, the unit of transportation is light, and assembly is simple. The load on the sandle comprises the vertical load due to the weight of the bride girders. The sandle member might become deformed by this loads. Additionally, the sandles are often used repeatedly because the sandles are a temporary structure. Therefore, slightly deformed sandle members may be reused. Until now, the stability of such structures with initial imperfections relied upon the technical expertise and experience of workers. The height of the sandle has been increasing because two-edged girders have widely been used in the steel brides in recent years to reduce costs. Relying on the experience of workers concerning the stability of sandles is more difficult because the height of the sandle has been increasing, and the number of skilled workers is decreasing. Therefore, a method for determining data regarding the stability of the sandle that does not rely solely on the experience of workers is needed. However, very few studies have reported on the stability of sandles. In this study, the influence of the initial imperfections on the stability of sandles is examined by numerical analysis as a fundamental study that confirms the performance of sandles. The analysis sandle models were made with fittings identical to the conditions at general con- struction sites. In the analysis models, the vertical load was added on the top part of the sandle. The initial imperfection of the steel H-beams was expressed by the size of the contact area between the joint of the upper and lower steel H-beams, and the size of this contact area was changed. To change the size of the contact area, plates with a thickness of 2 mm were put between the joint of the top step and the second step of the sandle, and the plate’s area was changed in the analysis. From the results of the analysis, the influence of initial imperfections in each step of the sandle was examined by the load transfer from the top to the bottom of the sandle. The results in this study are summarized as follows. Little dispersion and eccentricity of the load reaction generated by the vertical load exists in each step of sandle when the initial imperfection is located in one position in the upper part of sandle. Therefore, there is little influence of the initial imperfection of the sandle member on the stability of sandle. The influence of the initial imperfection was decreased when it was located at a lower step of the sandle. If the sandle member is managed appropriately such that the extremely deformed sandle member is removed, the stability of sandle will be guaranteed.

360 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Analysis for initial equilibrium condition and erection stages of Sorok (Self-Anchored Suspension) Bridge

Yunki Son & Dongho Yoo ENVICO Consultants Co., Ltd., Seoul, Korea Seungwook Jeong & Taeseop Yoon Daelim Industrial Co., Ltd., Seoul, Korea

For the construction of suspension bridge superstructure, a site engineer must be closely involved in the process of construction to support the construction, who should understand a structural system that is under interaction between cable and stiffening girder. In self-anchored suspension system, horizontal force of main cable is introduced into stiffening girder, therefore construction engineer needs to find out initial equilibrium condition considering deformation of stiffening girder due to the axial compression and the actual construction condition at site for matching the target given in the basic design unlikely anchored suspension system. Erection stage analyses are carried out based upon the initial equilibrium condition analysis result, so that the detail scheme of actual construction may be modified accordingly. This paper presents analysis and construction method of The Sorok (self-anchored suspension) bridge that is currently under construction in The Sorok Island, CheonllaNamdo, Korea and expected to be completed in 2008. To find the optimum initial equilibrium condition of self-anchored suspension bridge, the force of main cable and hanger is developed together with minimizing vertical camber of stiffening girder (Fig. 1). On the course of analysis, the model and its process consider the geometric nonlinearity of the structure that is the one of the most important characteristic of suspension bridge. Erection stage analyses is carried out for both hanger-pulling method and girder jack-down method to find out the more effective one which ensure the erection stability and cost efficiency. Hanger-pulling method is adopted as the erection scheme of suspension system as a result of the study. Various measurement results at site show that analysis results match up well with bridge deformation and profile of the girder during construction and at its final stage and have a good correspondence with the original design profile (Fig. 2).

Figure 1. Moment diagram of initial equilibrium Figure 2. Comparison of girder profile of the Sorok condition. Bridge for each construction stages.

361 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Seosang bridge movable scaffolding system

António Albuquerque Póvoas António Póvoas – Bridge Construction Systems – Portugal, Lda.

ABSTRACT: The great evolution of pre-tensioning technology allowed structural engineers to approach the design of bridges in a different way, by using larger and wider spans, in order to reduce global costs and construction periods. The mobile equipments have been following this tendency, and some of them may now be used to build bridges with larger spans, in less time, with lower global equipment costs. This article presents a modern concept of such equipments used in Seosang Bridge – South Korea., where safety systems have been taken to the extreme. Most of the common mobile equipments still depend a lot on manpower and auxiliary cranes and trucks, and such costs are many times underestimated, what often leads to huge deviations of the expected budget. As bridges normally have a few similar spans, some operations are repetitive and therefore it is possible to convert the manual tasks in operations that can be performed by hydraulic and electrical equipments. In our Movable Scaffolding Systems we are exploring the most the possibilities of converting all tasks in automatic operations performed by electrical or hydraulic equipments. In most of the normal MSS the launching system was already an automatic operation; in ours we developed remote control launching to increase safety. Another one of the tasks where manpower can now be successfully replaced by automatisms is the opening and closing of the formwork. A very challenging cross section is the inversed U shape of Seosang Bridge in South Korea, and in this paper we are pleased to show the automatic solutions designed to solve the corresponding problems. We have designed a formwork solution following several concepts: – It was necessary to have an independent inside formwork that could change wall and slab thickness easily and also could slide and rotate over a bottom grid of beams, where the bottom beams could incline transversally. – It was also necessary to have an independent outside wall formwork that could follow the inclination of the cross section along the bridge and adjust itself to the curve. – The inside formwork should also retract so that all the inside formwork could open when the bottom structure would rotate vertically. – The all formwork should be able to open laterally to allow that the suspended formwork could pass outside of the piers without touching the existing bridge. This model of overhead MSS provides excellent working conditions, both in materials handling equipments and safety platforms. As shown in the picture below, the formwork seems like a warehouse with several monorail hoists and safety platforms. A very important aspect of modern equipments as this one is the extreme care taken in the global safety of the equipment, by implementing the number and quality of working platforms and handrails, giving workers full protection against falling risks.

362 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

The construction of Machang cable-stayed bridge

Moshe Kim & Juwon Seo Institute of Hyundai Engineering and Construction Technology, Korea Jungho Song Hyundai Engineering and Construction Co., Ltd., Korea Hyuntai Lim Institute of Hyundai Engineering and Construction Technology, Korea

The Machang Bridge over Masan Bay was planed to link Masan and Changwon City. The Bridge is divided into three structures; the West Approach Bridge; the Main Bridge (cable-stayed); the East Approach Bridge.TheWestApproach Bridge is continuous, 3 span bridge (130 m + 150 m + 130 m) and consists of composite deck with a single steel box that varies in depth from 5 m to a maximum of 10 m. The Main Bridge is 740 m (170 m + 400 m + 170 m), composite deck cable-stayed bridge. The East Approach Bridge is continuous, 550 m long structure with end spans 65 m and 6 internal spans of 70 m and composite deck with a single steel box that has a constant depth of 3.65 m. The works were commenced in April 2004 and will be finished in 2008 as a privately financed project. The reinforced concrete pylons are H type, some 164 m above sea level. There are two reinforced concrete, hollow, cross beams some 6.5 m deep; one just above the pile cap, again for ship impact protection; and another to support the deck. Above deck level there are three, 2.2 diameter steel struts 18 m apart. The pylon foundations consist of 3.0 diameter RCD 28ea piles up to 50 m long. The 3 segment PC Houses were used to construct the pile cap for minimizing site joint and improving quality. The pile caps have a curtain wall extending to below the lowest water level to hide the piles. The construction methods of the pylon legs were only slip forms in the detail design but in actual method auto climbing forms from EL 119 m to the end considering the complicated structure near cable anchor box. The construction methods were chosen for the working safety above 64 m high sea level, con- struction period and minimizing the interference of ship route. The West Approach Bridge was constructed by the 3 segment large block method using 300 tonf floating crane, the Main Bridge by the balanced cantilever erection method using derrick crane and the EastApproach Bridge by incre- mental launching method. The West Approach Bridge A ballast wagon was used when reinforced concrete slab was constructed to reduce the stress of girder and to prevent crack of slabs.

Keywords: PC House, Pylon, Construction Method, Cable Stayed Bridge.

363 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Construction planning and analysis of six continuous extradosed PSC bridge

Sungho Kim & JuWon Seo Institute of Hyundai Engineering and Construction Technology, Korea Yunbum Lee & Inhwan Seo Hyundai Engineering and Construction Co., Ltd, Korea

The Extradosed Prestressed Segmental Concrete Bridge has been adopted an economical bridge as an alternative of about 150 m main span bridge. This types of bridge is inherited both the aesthetic advantage of cable stayed bridge and the economic strong point of PSC girder bridge. The Mooyeong extradosed PSC bridge will be constructed to connect Cheongho and Samho area near Mokpo city. It is designed with five span continuous extradosed cable supported bridge. The span and pylon height ratio of Mooyong bridge is L/7.3, which is relatively higher than usual extradosed types bridge. The Extradosed PSC bridge is usually required precise construction erection control considering structural change during the erection progress. But compared to the cable stayed bridge, it is assumed constant cable forces during the erection sequence that the result of construction step analysis may not reflect the accurate effects of cable forces variation during the construction. And the construction analysis should consider the variation of loading condition and mismatch of boundary condition from the planning stage assumptions. In this study, the construction planning and erection analysis of the relatively high span and pylon height ratio cases was performed and also the parametric study of cable force modeling and creep-shrinkage assumption. The assumption of anchorage of cable, stressing schedule and effective width of concrete girder segment might affect to the overall result of construction stage analysis result. The erection planning and stage analysis will be feed-backed with monitoring result during construction

Keywords: Extradosed bridge, PSC bridge, Erection Analysis.

364 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Underpass in the street O’Donnell in Madrid

C. Jurado Civil Engineer (PhD in course), Ingecal Ingenieros, S.L., Polytechnic University, Madrid, Spain

ABSTRACT: At the end of 1997 the Town Hall of Madrid, as the consequence of the increasing traffic in the street O’Donnell towards the Madrid airport, decided the construction of an underpass in O’Donnell street, below the underpass existed in Doctor Esquerdo street and above de under- ground line 6, threading between the other two and with a scarcely height, so just that the floor of the Doctor Esquerdo’s tunnel was the roof of O’Donnell’s tunnel. Due to these circumstances, besides the technical difficulty previously mentioned, the tunnel was, in the moment of execution of the work, the deepest (14 m under the surface of the street) and the longest (547 m) of the existing ones, up to this date in Madrid. As additional circumstances of the project and of the execution, there were numerous services existed affected, since the site corresponds to an urban central area: telephony, electricity, supply of water, TV cable, etc.

Figure 1. Plant and sections of the underpass.

365 Among them, it is necessary to mention the existence of two service galleries situated at the two sides of Doctor Esquerdo’s tunnel, made in reinforced concrete with semicircumpherential section at the top, with an exterior width of 3,10 m and a maximum interior height in the key of the arch of 2,65 m. The walls were 0,40 m thickness. Another singular service affected were two waste collectors, which were circulating along O’Donnell street, hand-constructed and affected by the concrete sheetpiling of the new underpass and which was necessary reconstructed, by the same procedure, and in a length of approximately 1,5 kilometres. The contract included also the remodelling of the road connexion between Sainz de Baranda and O’Donnell streets, which made necessary the construction of a bridge of prestessed concrete of two span of 44 m. of total length.

Figure 2. Aerial sight of the O’Donnnell Tunnel.

Figure 3. Bridge over the link to Alcalde Sainz de Baranda.

366 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Construction of P.S.C. Box girder bridge (F.C.M) which pre-compensation method is applied to

Hyunwoo Jee, Jae Heung Shim, Man Keang Min & Hyoung Jae Lee GS Engineering &Construction, Korea

When we design and construct super-structures and piers of P.S.C.box bridges, we have to consider a lot of conditions like creep, shrinkage, thermal load and so on that are influencing the Rahmen- Type P.S.C. box bridges. Especially, in case of the bridges that are constructed by Free Cantilever Method (F.C.M.), because of these conditions, the moments at lower place of pier are so big that we need larger pier sections and more reinforcing bars. In order to avoid these undesirable situations, we apply jacking force to key segments which are joining parts of free cantilevers in P.S.C. box bridge constructed by free cantilever method (this method is named Pre-Compensation Method) and make the moments at lower place of pier considerably small. In this paper, we will introduce an actual case in which pre-compensation method was applied to Sin-Chon Bridge and show economical efficiency of this method.

Figure 1. View of Sinchon bridge.

Figure 2. Location where the pre compensation load is affected and conceptual diagram.

367 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Large scale cyclic tests of precast segmental concrete bridge columns with unbonded post-tensioning tendons

Y.-C. Ou & G.C. Lee Department of Civil, Structural and Environmental Engineering, University at Buffalo, Buffalo, NY, USA P.-H.Wang National Center for Research on Earthquake Engineering, Taipei, Taiwan M.-S. Tsai & K.-C. Chang Department of Civil Engineering, National Taiwan University, Taipei, Taiwan

ABSTRACT: Precast segmental construction of concrete bridge columns is a method of construc- tion in which bridge columns are segmentally prefabricated off site and then assembled on site with post-tensioning. The major advantages of such type of construction as compared to conventional monolithic construction include reduced traffic disruption and environmental impact. The objective of this research is to develop precast segmental columns for use in seismic regions. The proposed columns are designed with bonded mild steel bars that are extended across the segment joints for increasing the columns’ energy dissipation capacity and lateral strength. The post-tensioning ten- dons are placed in the hollow core of the columns and left unbonded with the surrounding concrete, which can decrease the prestress loss due to lateral displacements of the columns. The results of the experimental investigation showed that under lateral cyclic loading the proposed columns exhibited excellent ductility. Moreover, the increases in hysteretic energy dissipation and lateral strength due to the use of the mild steel bars were evident.

368 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Key-segment closing method using artificial heat for partially earth-anchored cable stayed bridges with classical span length

J.H. Won, K.I. Cho, J.H. Yoon & S.H. Kim School of Civil and Environmental Engineering, Yonsei University, Seoul, Korea

In this study, a new type of key-segment closing method by thermal expansion is proposed as a substitute to the process of set-back and reset-back of the FCM (Free Cantilever Method) construc- tion of a partially earth-anchored cable-stayed bridge with a classical main span length, that is, of between 150 and 500 m. The proposed method is artificially to heat up the girder in a cantilever state while the derrick crane holds the key-segment. After the key-segment and the adjacent segments are made to contact due to the longitudinal displacement caused by the heating, a connection of the key-segment is made. Then, the longitudinal restraints that restrict the longitudinal movement at the pylon in the cantilever state are removed. Finally, the artificial heating is removed (Fig. 1). Using the changes of boundary conditions and structural systems, the heating and heat removal process can generate the tensile forces in all of the girders as the cables restrain the contraction of the girder. These forces reflect the advantage of the partially earth-anchored cable system. The effect of the proposed method is analytically investigated for three-span cable-stayed bridge. According to the FCM construction method, a construction sequence analysis is performed to verify the effect of the proposed key-segment closing method. From the results, it is found that the efficiency of partially earth-anchored cable-stayed bridge is enhanced by applying the proposed method since the compressive axial forces in girders are reduced.

Figure 1. Construction sequence by proposed method.

Keywords: key-segment closing method, partially earth-anchored, cable-stayed bridge, artificial heat, construction sequence.

369 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Highway bridges made of circular hollow sections

Ulrike Kuhlmann & Mathias Euler Institute of Structural Design, University of Stuttgart, Germany

Spatial trusses made of Circular Hollow Sections (CHS) have become more and more popular in steel-concrete composite highway bridges over the last 15 years in Europe forming an innovative and aesthetic alternative to the conventional all-steel bridges. The spatial CHS trusses normally com- prise two planes of rising and falling (not crossing) braces and a continuous bottom chord.At the bot- tom chord the connection of the truss members forms a so-called multi-planar KK-joint. The two top chords may be integrated within the concrete slab of the bridge. From the viewpoint of economics and ease of construction the braces are preferred being welded directly on the bottom chord. In order to verify a spatial tubular truss structure against fatigue failure the S−N method cannot be used anymore due to the complex geometry of the welded connections. The more sophisticated hot-spot stress approach has to be applied. The areas of stress concentration close to the truss connections referred to as hot spots are highly dependent on the geometry of the truss members and are crucial for the fatigue strength of the entire truss structure. This paper presents first results of a still on-going research project that aims at developing recom- mendations concerning the fatigue design and construction of welded KK-joints in highway bridges. Since the design of highway bridges including tubular trusses is generally fatigue-controlled, the developed recommendations concerning the design phase, the fatigue verification, including the determination of SCF values for multi-planar KK-joints for highway bridge like geometries, and the fabrication follow an integral view and thus allow the realization of aesthetic composite tubular truss bridges with welded KK-joints considering the high level of requirements for highway bridge design.

ACKNOWLEDGEMENT

The research project has been financed by the German Federal Highway Research Institute Bundesanstalt fuer Strassenwesen (BASt). We gratefully acknowledge the received support.

Figure 1. Composite bridge with tubular trusses and KK-joints (a), cast steel type (b), welded type (c).

370 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Monitoring of early stage prestress change of long span steam-cured concrete box girder with pretension method

Kyu-Yong Choi, Geun-Yong Song & Dong-Ok Kang Samsung Corporation, Korea Su-Won Cha Ulsan University, Korea

In this study, the early-stage change of prestressing force was measured by using effective stress- meters and strain gauges, which was installed on a steam cured 50m-long span pretensioned box girder. From the measured values from effective stress-meters and embedded strain gauges, the instantaneous response at detensioning strands as well as global change of concrete stress could be checked by the consistent measurement during the whole fabrication stages. This result verified the safety and stability of the girder. A systematic monitoring was conducted to a girder through the whole process from fabrication to erection. It was focused on the structural behavior, especially the introduction of enough pre- stressed force and the long-term stability under the repeated extreme loads induced when a new girder on the carrier was passing on the monitored girder.

Figure 1. strain changes of SL-2, SL-9 and ST-1 and results of effective stress-meters at main locations.

Figure 2. static loading test for the check of long-term stability. 371 Finally, it has been shown that the stable compressive stress could be sufficiently maintained in manufacture steps. This comes from the special features of steam-curing process that is the combination of introduced pre-stress for the longitudinal and transverse direction and induced concrete stress during steam-curing.

372 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Effect of deformation of spans on serviceability of composite highway bridge

Z. Manko Institute of Civil Engineering, Wroclaw University of Technology, Wroclaw, Poland

ABSTRACT: The serviceability and performance (behaviour) of a multispan composite (steel + concrete) highway bridge over the Warta river in Slugocinek on Poznan – Warsaw highway A-2 were determined in the context of the bridge’s long service and the deformations discovered in all its spans. This was done on the basis of careful inspections of the bridge and field measurements of deflections in all the seven spans in the two roadways as well as by calculations. The effect of the considered types of bearings and ways of resting the spans and of some non-mechanical factors (e.g. shrinkage, temperature, etc.) on the magnitude of deflections and internal forces in such bridge structures made apparently continuous by the deck plate slab were determined. A description of the structure is provided, the way in which the displacements of the spans were measured is shown, the results obtained from the calculations and the measurements are analyzed and the main conclusions are discussed.

An assessment of the durability of a building structure, and a bridge in particular, is a highly complex task performed on many planes. Therefore, as a rule, multifarious activities such as periodic surveys, inventory control and strength analyses that employ mathematical apparatus are conducted. The conclusions drawn from them serve as a basis for determining the serviceability and durability of the bridge structure with probability depending on the method of assessment. The paper discusses the static work of the new multispan composite bridge with its made- continuous deck slab under the service (operational) load for different geometric-constructional parameters in comparison with the work and performance of the spans of such a bridge with many simple-supported spans. Effect different configurations are to be analyzed. Results of measure- ments of the deflections of the spans made on the bridge and typical defects illustrating its current configurations are included. The effect of particular static schemes adopted for the composite spans and of non-mechanical operational effects is described. It is shown that when the latter had been taken into account in the design, the anxieties about the bridge’sload capacity and serviceability turned out to be unfounded. Also the influence of some parameters on the behaviour of numerical models of the spans of this bridge under an actual load (geometry, loads, etc.) for different schemes of their support, depending on the system and type of the bearings, is described. Making the reinforced concrete deck slab in composite bridges continuous has become quite a common practice in bridge construction in Poland, adopted when repairing old bridges or when building new multispan bridges. The main aim is to reduce the number of expansion joints, especially on highways. Various implemented methods of making the deck plate slab continuous, particularly in composite bridges can be found in some papers. Considering the relatively short time which has passed since these solutions were implemented, no definite opinion on the merits of the particular solutions can be formed yet. In most cases there is no seepage within the expansion gaps between the made-continuous spans. At the same time there is more seepage and the supports’ caps and the steel girders’ faces are wet in the places where expansion joints were employed at the contact between the sections which were made continuous.

373

Integrated assessment – Practical application of probabilistic methods

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Guideline for probabilistic assessment of deteriorated bridges

Jørn Lauridsen The Danish Road Directorate, Ministry of Transport, Copenhagen, Denmark Finn M. Jensen COWI A/S, Copenhagen, Denmark Svend Engelund COWI A/S, Aalborg, Denmark

The Danish Road Directorate (DRD) has within the last 5–10 years applied probability-based approaches to determine the reliability of bridges which fails an initial deterministic assessment. This approach has in many cases ensured that costly strengthening or replacement of existing serviceable bridges could be avoided or postponed. The result has been monetary savings and avoidance or minimization of traffic inconvenience. In 2004 the DRD issued a Guideline Document “Reliability-Based Classification of the Load Carrying Capacity of Existing Bridges”, which in detail describes how a probability-based assessment of the load bearing capacity can be performed in accordance with the requirements for the safety level prescribed by the DRD. The guideline is developed for “as-new” bridges, i.e. bridges with no or minor deterioration. A bridge is subject to a number of destructive mechanisms. The deterioration of the bridge will ultimately lead to a situation where the reliability of the bridge with respect to serviceability or collapse is no longer acceptable. Normally, using the LRFD (Load and Resistance Factor Design) methodology it may be demonstrated that the reliability of the bridge is acceptable. However, the LRFD methodology does not allow new information gained from inspections and measurements to be taken into account in a rational manner. Therefore, the assessment of the reliability of existing structures must be performed using probabilistic models and methods. In order to disseminate the approach within Denmark a guideline for probability-based management has been developed. Based upon the experience obtained from previous prac- tical probability-based management plans a supplementary guideline to “Reliability-based classification of the load carrying capacity of bridges” has been completed, see http://www. vejdirektoratet.dk/publikationer/VDrap291/index.htm. The guideline focus on probability-based maintenance management of deteriorated bridges, such as the inspection and sampling required to model current deterioration, modelling of deterioration rates to be used in future predictions, comparative evaluation of possible actions and so forth – and includes policies and possibilities for maintenance and for continuous updating and validation of the probability-based management plans. As a primary result from the guideline, the DRD expect monetary savings and more degrees of freedom within the overall bridge management.

377 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Optimization of special inspections of concrete bridges

Svend Engelund COWI A/S, Aalborg, Denmark Mette Sloth COWI A/S, Copenhagen, Denmark

ABSTRACT: The Danish Road Directorate (DRD) administers approximately 2500 minor high- way bridges of which a majority is between 30 and 50 years old and of which 95% are concrete bridges. Many of these bridges require major repair or replacement within the next 0–20 years. In order to ensure that the safety of a given bridge is acceptable, inspections and measurements are performed. On the basis of the results of the inspections and measurements, an optimal strategy for repair and maintenance must be developed. The optimal strategy is the strategy which minimizes the cost of repair and maintenance while ensuring that the reliability of the bridge is acceptable. The number of bridges to be inspected and the limited budgets for repairs has forced the DRD to optimize and streamline the inspection procedure and the corresponding repair strategies. The DRD has initiated a program for optimization of special inspections in the form of a computer program and a corresponding guideline for planning, evaluating and reporting of special inspec- tions. The program will be used to develop cost-effective inspection plans for special inspections of reinforced concrete structures subject to deterioration due to corrosion of the reinforcement. In future the program is to form the basis for a tool for maintenance planning of concrete structures subject to corrosion. Results of special inspections and measurements typically describe the condition very locally (point wise). Planning of special inspections must, however, be performed on the basis of the condition of a surface. Reliability-based assessment of the condition of a bridge has traditionally been based on a point wise evaluation of the inspection results and has then been extracted on basis of experiences to describe the whole surface (area). The program to be developed is reliability- based and it quantifies the spatial variability based on methods developed recently and described in a number of papers and journals – see e.g. Malioka, V.and Faber, M.H. Modelling of the Spatial variability for Concrete Structures, Proceedings IABMAS’04, 2nd International Conference on Bridge Maintenance, Safety and Management, Kyoto, Japan, October 18–22, 2004.

378 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Probability based assessment of railway bridges in Denmark – Previous applications, current state and future possibilities

J.S. Jensen & D.F. Wisniewski COWI A/S, Kongens Lyngby, Denmark O.B. Ulstrup Rail Net Denmark, Frederica, Denmark

ABSTRACT: Probabilistic methods have already shown their potentials in saving many existing bridges in Denmark from unnecessary strengthening or replacement. This paper discus several topics related to the probability based assessment of railway bridges. It presents some application examples, describe current state of Danish legislation regarding the use of these methods and discuss future possibilities of using these methods in every day assessment of railway bridges, as it is intended by the authors of the “Guideline for Load and Resistance Assessment of Existing European Railway Bridges – advices on the use of advanced methods” prepared within Sustainable Bridges project.

1 INTRODUCTION

The Danish Railway Administration “Rail Net Denmark” is responsible for the maintenance man- agement of whole Danish railway network comprising of 3240 km of railway lines, 2375 ordinary, short to medium span bridges and 7 large bridges. To keep all bridges in service and allowed for uninterrupted operation of all the railway lines Rail Net Denmark is continuously trying to apply and promote state-of-the-art solutions and methodologies for the efficient maintenance manage- ment of bridges. The Rail Net Denmark and COWI as one of its main consultants are also trying to apply the state-of-the-art methodology for the load capacity evaluation of existing bridges in order to allow for passage of heavier and faster trains. This methodology consists of several assess- ment steps with increasing level of accuracy and complexity. The highest step combines the use of detailed investigations (special inspection, some laboratory or in-situ test, etc.) and enhanced assessment methods (non-linear or plastic structural analysis, probabilistic assessment, etc.). Probabilistic methods combined with enhanced structural analysis have already shown their potentials in saving many existing bridges in Denmark from unnecessary strengthening or replace- ment. An excellent example is a railway line from Copenhagen to Padborg which had to be upgraded for higher traffic loads. The current guideline for the assessment of Danish railway bridges (BN1-59-2,2006) allows for using probabilistic methods in the assessment. However, so far, these methods are not as commonly applied as in case of highway bridges due to the lack of compre- hensive recommendations or guidelines focused on the use of these methods for evaluation of railway bridges. The “Guideline for Load and Resistance Assessment of Existing European Rail- way Bridges – advices on the use of advanced methods” (SB-LRA, 2007) prepared within the Sustainable Bridges Project, by the Work Package 4, where COWI is a task leader, is bridging this gap and hopefully will result that application of probabilistic methods will become the standard assessment tool for the evaluation of existing railway bridges in Denmark, which fails the traditional assessment. 379 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Probability based assessment of motorway bridges in Denmark

J. Bjerrum The Danish Road Directorate, Ministry of Transport, Copenhagen, Denmark A. O’Connor, C. Pedersen & I. Enevoldsen RAMBOLL, Virum, Denmark

This paper describes practical applications of the Danish guideline for probability-based assessment of motorway bridges. The guideline, the first in the world of its kind, describes how probability- based assessment of bridges can be performed in accordance with the requirements for the safety level prescribed by the Danish Roads Directorate (DRD). The guideline specifies principles for modelling uncertainties including treatment of model uncertainties. The requirement at the ultimate limit state for the structural safety is specified with reference to failure types and failure conse- quences. The guideline, in conjunction with codes of practice, provides the DRD with the legal justification necessary for application of probability-based approaches in Denmark. A number of practical examples where the guideline has proven beneficial in increasing the structures load rating are presented. Figure 1 presents two of the structures considered in the paper. Figure 1(a) presents sister prestressed beam and slab bridges located at Klovtofte, outside Copenhagen. Constructed in the early 1970’sthe bridges are constructed from 31 precast beams (29 internal + 2 edge) with an in-situ slab. The beam spans are 10.7, 24.1, 24.1 and 10.7 m respectively. The reinforced concrete deck is 36.1 m wide (excluding cantilevers, 2 × 0.5 m) and 160 mm deep. The bridges are of integral form being continuous over the internal supports. Deterministic evaluation of required sectional resistances, in shear and flexure, indicated that the bridges have sufficient capacity to achieve a Class 100 rating everywhere except in shear in the inverted T-beams which were the principal structural elements in the main spans. Figure 1(b) presents Avdebo bridge which is reinforced concrete arch bridge built in 1932. The structural which is skewed at 56.6◦ has a clear span of 23.0 m and height 3.2 m. The arch bridge carries two lanes of traffic over a river. The width of the structure is 9 m. The arch thickness varies from 0.3 m at the crown to 0.6 m at the base. The wing walls on either side of the arch have length 5.9 m, giving a total structure length of 36.3 m. The wingwalls are supported

Figure 1. Practical examples.

380 by a single rib on either side of the bridge, where the bridge has 4 ribs in total. Some renovation work was carried out on the bridge in 1986. This work consisted of replacing the eastern edge beam and both barriers. The results of deterministic assessment of the structure for the normal and conditional passage types suggested insufficient bending capacity in the arch footings. Recognising the inherent conservatism of the deterministic assessment it was decided to per- form a probability based assessment of the structures to determine their formal level of safety in accordance with the relevant safety standards. This probabilistic assessment involves modelling the variables of load and resistance as random and identifying the formal probability of failure of the structures based on the FORM technique. This ‘site-specific’ failure probability is then com- pared with minimum allowable value permitted, to ensure the relevant safety criteria are met. It is important to stress that at no stage is the safety of the structure compromised, rather the process derives a bridge specific code, removing the conservatism inherent in deterministic codes, which must necessarily generalise to be widely applicable. The requirements at the ultimate limit state for the structural safety were specified with reference to failure types and failure consequences, i.e. safety class with requirements for the formal annual probability of failure pf . For the structures under consideration in this paper the assessment criteria was defined by a requirement of ductile failure without remaining capacity, with a corresponding minimum acceptable value of β = 4.75 according to the aforementioned Danish guideline for probability-based assessment. The results of the probabilistic assessment yielded minimum safety indices β>4.75 for all of the structures considered, in the normal passage cases. As such the structures were deemed safe without any requirement for rehabilitation or strengthening. The results represent a significant saving for the bridge owner both in terms of the direct replacement cost and of the indirect costs, i.e. user delay costs, which would have been incurred in replacing the structure. It is important to stress that at no stage has the safety of the structure been compromised, rather a more realistic and bridge specific safety assessment has been performed free from the generalisations of deterministic codes. The Danish Roads Directorate now pursues reliability based assessment as a matter of course for all structures which have failed a deterministic assessment. The results of this policy have provided significant savings in both the direct and indirect costs associated with bridge rehabilitation or replacement. Table 1 below lists the direct monetary benefits, in excess of $30,000,000(USD), accrued in some recent cases where probability based assessments have been employed. While the guideline is intended to be applied in probability-based assessment of Danish bridges it is anticipated that it could equally be applied in other countries. It may be downloaded from the website of the Danish Roads Directorate at www.vd.dk.

Table 1. Examples of DRD savings from probability based assessments.

Bridge Result of deterministic analysis Probability-based assessment Cost saving $(USD)

Vilsund Max W = 40 t Max W = 100 t 3,500,000 Skovdiget Lifetime ∼ 0 years Lifetime > 15 years 13,200,000 Storstroem Lifetime ∼ 0 years Lifetime > 10 years 2,500,000 Klovtofte Max W = 50 t Max W = 100 t 2,200,000 407-0028 Max W = 60 t Max W = 150 t 175,000 30-0124 Max W = 45 t Max W = 100 t 175,000 Norreso Max W = 50 t Max W = 100 t 600,000 Rødbyhavn Max W = 70 t Max W = 100 t 600,000 Åkalve Bro Max W = 80 t Max W = 100 t 1,750,000 Nystedvej Bro Max W = 80 t Max W = 100 t 2,200,000 Avdebo Bro Max W = 80 t Max W = 100 t 3,300,000 TOTAL >30,000,000

381 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Probability based assessment of a large riveted truss railway bridge

A. O’Connor, C. Pedersen & I. Enevoldsen RAMBOLL, Virum, Denmark L. Gustavsson & J. Hammarbäck Banverket (Swedish Rail Administration), Sweden

This paper describes the techniques employed in the probabilistic assessment of a steel arch rail- way bridge in Sweden. The probability-based classification of the structure serves as an example of how probability-based assessment of railway bridges can be applied to avoid unnecessary repair/rehabilitation and/or to optimise those repairs where they are indeed necessary. Modelling of the critical limit states are presented for both the elements and the riveted joints of the structure. The statistical techniques employed in modelling the train loads are presented. The overall aim of the analysis was to achieve a higher load rating for element/joints of the structure than those result- ing from the deterministic assessment. Ultimately the cost benefits to bridge owners/managers of performing a probabilistic assessment are apparent from the results. Bergeforsen railway bridge is a single track bridge which was constructed in 1923. The bridge is situated on Swedish rail’s Sundsvall-Härnösand line, approximately 350 km north of Stockholm. The superstructure of the bridge is composed of riveted trusses with spans of 42 + 84 + 42 = 168 m as illustrated in Figure 1. Simply supported side approach spans of 22.5 m and 11.6 m give a total bridge length of 202.1 m. The superstructure of 42 + 84 + 42 m is supported at 4 longitudinal locations, 1-fixed + 3-roller, which essentially results in the bridge working as a continuous beam over 3 spans. Deterministic assessment of the bridge was performed according to the Swedish Assessment code, BVH 583.11, using the train load model BV-3 (i.e. a trainload model with 25 tonne axles and 8 tonne/m line load). The results of the deterministic assessment demonstrated that while the structure had sufficient capacity with respect to the Serviceability and Fatigue Limit States, that at

Figure 1. Bergeforsen Railway Bridge.

382 Figure 2. Connection for which β<4.8. a number of locations the structure failed to demonstrate the necessary Ultimate Limit State (ULS) capacity and as such some form of strengthening was required. As this conclusion would prove extremely costly, rather than immediately accepting the consequences of a traditional decision making process, in which this failure in a deterministic assessment would lead to a requirement for strengthening/ rehabilitation of the bridge, Banverket’s management strategy focused on an alternative decision process which would permit a probabilistic assessment of the structure to be performed to see if the necessary capacity could be demonstrated. Where sufficient capacity could not be demonstrated by the probabilistic approach it was intended that a probability based decision process would be implemented which would provide a rigorous/robust decision methodology and consequently provide information on the optimal maintenance strategy. The legal justification for this process was obtained from BVH 583.11 where it is stated that Banverket permits probability- based assessments. The requirement in the ultimate limit state for the structural safety was specified with reference to required safety class. BFS 1993:58 1993 specified that the highest safety class was needed for both the elements and the connections and as such a value of β = 4.8 was specified. In the majority of cases analyzed the probabilistic assessments were able to demonstrate sufficient capacity, i.e. β>4.8 for the elements and joints considered. However, in the case of 2 joints, probabilistic assessment could not succeed in avoiding some level of strengthening. Figure 2 illustrates one of the joints, circled, for which the calculated β = 4.51, which, as it was < 4.8 could not be deemed to satisfy the minimum safety requirements. Two sample proposals for strengthening the joint were considered, both of which involved removal of certain rivet groups from the joint and their replacement with high strength bolts. Option A, labelled in Figure 2 considered replacement of the central rivets with 27 mm diameter bolts, while option B, also labelled in Figure 2, considered replacement of the bottom two rows of rivets with bolts. In both cases the probabilistic model which was established for safety analysis of the joint was used to recompute the revised structural safety on the basis of the proposed joint strengthening methodology. The results of these re-assessments were, Option A β = 6.05 > 4.8 and Option B β = 7.80 >4.8. The significance of these results lies in the way in which they demonstrate the ability of the prob- abilistic model to assess the relative efficiency of proposed repair methods and as such to be used not only in safety assessment of structures to demonstrate increased load carrying capacity, relative to deterministic assessment, but also to be used as a tool in optimising maintenance/rehabilitation planning for structures which are determined to require some level of repair/rehabilitation to satisfy specified minimum safety criteria. The results of the assessment represented a considerable eco- nomic saving for the bridge owner through the avoidance of unnecessary repairs and the optimisation of repairs where they were indeed required.

383

Integrating health monitoring and lifecycle management of bridge and highways

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Integrating human, natural and engineered systems and associated paradigms for infrastructure asset management

Franklin Moon, Patrick Gurian, Franco Montalto & A. Emin Aktan Drexel University, Philadelphia, USA

EXTENDED ABSTRACT

The intertwined system-of-systems character of infrastructures have become far too complex for civil or other engineers or scientists to fully understand, model, simulate, perform scenario analysis and to positively perturb and control these systems. The 1996 NSF workshop on integrated research for civil infrastructure delineated that infrastructures are intertwined systems with human, natural and engineered elements. Since then, we have made little progress in fully understanding and char- acterizing infrastructure systems. Shinozuka (2003), Haimes (2006 and 2007), deNeufville (1999 and 2003), and many others have contributed to the state-of-the-art in modeling infrastructures at different levels of resolution; however, a taxonomy of modeling approaches for human, natural and engineered elements of infrastructures, and methods of their integration for the simulation of entire infrastructures does not yet exist. To serve as a starting point, we attempted a classification of known modeling techniques in engineering and science as shown in Table 1. Modeling infrastructures as multi-domain systems reliably is currently a formidable challenge. Our inability to construct models that could be calibrated and validated has significant negative financial impacts. For example, federal transportation acts since the U.S. Inter-modal SurfaceTrans- portation Efficiency Act of 1991 have required that infrastructure investments, including renewal, should be based on a demand analysis. Demand projections are currently based on empirical mod- els, and there is considerable evidence of a lack of their reliability. Many infrastructure investments

Table 1. Tentative classification of modeling approaches.

Physics based models Non physics-based

Continua Models Data-Based (Rational) Models • Differential Equations: • Statistical Theory of Elasticity ANN Field and Wave Eqns Signal/Pattern Analysis Idealized Diff. Eqns. (Bernouilli, Vlasov, etc) Rule-Based (Meta-Models) Geometric Lumped-Parameter Models: • Macro-economic Models • Smeared – Macro or Element Level Models • Semantic Models • Microscopic FEM • Ontology • Mixed Smeared-FEM • Semiotic Models • Modal Models • Empirical Numerical Models • Probabilistic Models Network Models Histograms Standard Distributions Monte-Carlo Event-based (Bayesian) Time-based (Markov) Symptom-based • Agent Models: Rules + MonteCarlo

387 such as bridges and highways often become functionally obsolete due to operational saturation in just one or two decades while their projected lifecycles are 50–75 years. To construct infrastructure models at a sufficiently fine resolution such that many of the critical mechanisms and parameters as well as connectivity and interdependence between various systems may be explicitly represented and simulated is a pressing need. Hansman, et al. (2006) articulated the need to approach infras- tructures as multi-domain systems, but recognized that this is a new field of study that has to be established. Further, there is the issue of which disciplines would be the best coordinators of such a field. As advocated by Bordogna since early 1990’s, we propose that civil engineers, with an appreciation of how human, natural and engineered systems interact, should be coordinating multi-disciplinary teams for integrative research to develop this new field of multi-domain sys- tems engineering. We also emphasize that such a new field of multi-domain systems engineering will require new tools for observing, measuring, modeling, identifying and simulating, and, man- ners of controlling infrastructure systems, based on metrics that should be defined for measuring and enhancing their multi-dimensional performance. Finally, we emphasize that the challenges of “observing and measuring infrastructures” is not unlike that is faced by deep sea researchers in the Planet Ocean and other major NSF science programs. Both the challenges and the importance of modeling the multi-domain systems at various resolu- tions, using both physics-based and non physics-based mixed modeling approaches, a multi-layered framework is being proposed as shown in Fig. 1. This Figure depicts the highway transportation system in terms of various human, natural and engineered layers, which are interconnected and interacting with each other. This system permits a limited set of inputs, such as politics, policy, planning, financing, etc. that are all intertwined and highly constrained. These inputs perturb and interact with the system in various manners which we can neither FULLY UNDERSTAND nor CONTROL, resulting in system performances in various categories. We have no means of determining which measures to take for optimum lifecycle benefit/cost and we do not have per- formance metrics for various layers of the system which affect outcomes. If we could model this multi-domain system using mixed modeling approaches at various resolutions, we may perform “system-identification,” i.e. identify the system and establish optimum manners of perturbing and controlling its performances and health. To construct such a model we have to observe and measure the system, map and model it, perturb it in a controlled manner, and then calibrate the model. The challenge is in the geometric and temporal scales of the system, and in the complexity of interac- tions between various human, natural and engineered elements. Just as it took Galileo’sinvention of the telescope for observing and modeling the solar system, we need multi-scale sensing-imaging- communication and computing systems and infrastructure sub-systems serving as field laboratories (or observation platforms) for observing and modeling infrastructure systems. This paper outlines emerging paradigms relevant to infrastructure management and discuss a framework for their eventual integration.

388 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Structure and infrastructure health monitoring as a key enabling paradigm for integrated asset management

Franklin L. Moon & A. Emin Aktan Drexel University, Philadelphia, USA Frank Jalinoos & Hamid Ghasemi US Federal Highway Administration

EXTENDED ABSTRACT: The overarching goal of this paper is to provide an overview of Asset Management (AM) and to discuss the roles of various technologies and related paradigms within AM. To place this discussion in its proper context, it is first necessary to more clearly define transportation infrastructure as more than simply physical assets. For the sake of this paper, we will define transportation infrastructures as systems made up of interacting engineered (including field constructed, shop fabricated, and factory manufactured components), natural (soil, water, weather, climate, etc.) and human (users, owners, industries, politicians, etc.) sub-systems. Unlike some other infrastructures, transportation infrastructures are almost exclusively engineered and managed by civil and environmental engineers. Recently there has been growing concern over the manner and effectiveness of the management approach employed for these infrastructures. Whether one considers the increasingly poor performance, related to congestion (USDOT 2006), safety (NHTSA 2004), or condition (www.asce.org/reportcard/2005) to name a few aspects, the pervasive budget shortfalls many owners are facing, or the increased demand for more accountability with the use of public funds by US citizenry, it is clear that a revolution related to the management of transportation infrastructure is long overdue. Towards that end, there is broad consensus that this revolution will take shape through adapting the paradigm of AM. Amidst the buzz surrounding AM, a clear understanding of what it is and what it is not may be difficult to distill. It is imperative that AM not be viewed as a solution to the challenges that currently face our infrastructure. Rather this paradigm should be viewed as a tool that holds promise to improve decision-making through the proper identification of trade-offs. That is, AM will not increase funding but rather aid in identifying the most appropriate allocation of funds – in the absence of unlimited resources, such decisions will always result in funding certain assets at the expense of others. The goal then is to seek more optimum investments where the benefits of the new investment outweigh the losses incurred by reducing funds in other areas. An illustrative, if tragic, example of this important distinction between seeking solutions and seeking optimum trade-offs was offered by economist Thomas Sowell (1995). In 1989 an infant was fatally injured during a minor airplane crash after being ripped from the arms of her mother. The knee-jerk reaction indicative of solution seeking was to require all infants to have their own seats and to be securely fastened in their seats during take-off and landing. Although such an approach seems quite logical, its flaws are clearly seen if one examines the trade-offs. First, such a solution would clearly increase the cost of the flying public and thus, in some cases, deter flying and promote driving, which is a more dangerous form of travel. In fact, hearings held by the U.S. House of Representatives (1990) concluded that within a decade’s time, this solution could be expected to result in nine additional infant deaths due to car accidents and save only one infant from dying in a plane. Moreover this solution would have cost the travelling public nearly $3 billion. While not completely parallel with the discussion at hand, this example is illustrative of both the benefits and the requirements of AM. First and foremost, AM aims to provide information about trade-offs to decision-makers in order to avoid the knee-jerk solutions that are far from optimum

389 and may in fact be counter-productive. Although such a holistic approach to the management of infrastructure has been long desired, it has been allusive, and currently decision-makers are still largely faced with finding solutions (e.g. “worst first”) without a clear understanding of the trade- offs they are making. For example, many transportation engineers define AM at the project level (i.e. management of a bridge or a stretch of pavement). A common solution applied is then to invest in preventive maintenance prior to or at the onset of deterioration. The reasoning behind this solution is that the decrease in the condition of a newly constructed facility is expected to progress linearly until the onset of deterioration, and then decrease exponentially if preventive maintenance is deferred. For the time being, let us neglect the inherent problems associated with assuming a time-dependent deterioration progression and identifying when deterioration has begun. Similar to the knee-jerk solution to the infant safety issue, this approach neglects the trade-offs that impact the transportation system’s performance as it is a one-size-fit’s-all approach based solely on benefits of a small component of the total system. To effectively identify trade-offs associated with investment decisions (which represents the underpinning of AM), two components are required: (1) the definition of objectives of the infras- tructure owner and metrics correlated with these objectives, and (2) the ability to monitor and forecast the identified metrics to support the identification of trade-offs. In the example given, the objectives were to maximize safety and minimize cost, and the metrics used to quantify the trade-offs were infant deaths and consumer costs in dollars. These metrics were then forecasted based on statistical databases to identify the trade-offs and provide decision-makers with a clear understanding of the impact of their choices. The challenges we face as we strive to crystallize and demonstrate this paradigm that is essen- tial for innovating the engineering and management of complex infrastructures in general, and highway infrastructure in particular, are significant. The rewards we expect are also significant. For example, the Netherlands Directorate-General for Public Works and Water Management indi- cated savings of about 25% following their transition to integrated AM. According to de Neufville and Odoni (2003), operational savings of 30% may be expected when applied systems analysis is employed for an airport. These savings are too significant, and the challenges we face in prop- erly operating, preserving and protecting the highway infrastructure are too prevailing for us to continue our research, education and practices without leveraging such powerful paradigms. This requires a multi-disciplinary systems engineering approach, requiring civil engineers with expertise in coordinated multi-disciplinary education, research and practice.

390 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Integration of health monitoring in asset management in a life-cycle perspective

Thomas B. Messervey Department of Structural Mechanics, University of Pavia, Italy Dan M. Frangopol Fazlur R. Khan Endowed Chair of Structural Engineering and Architecture, Department of Civil and Environmental Engineering, ATLSS Center, Lehigh University, Bethlehem, PA, USA

ABSTRACT: The management of civil infrastructure involves a unique composition of inter- ested parties. Researchers, transportation officials, infrastructure managers, and elected officials (to name a few) may compete for resources or have conflicting interests although they share the same goal of safe and efficient structures. The contrast between the rapid pace of development within the field of SHM and the time required to implement changes in the design, construction, and management of civil infrastructure further complicated this scenario. To ensure the best use of limited resources and to promote the timely development of SHM applications, coordinated and synchronized actions are required. Such adoptions in concert of common metrics, method- ologies, and means of communication can ensure synergistic benefits between interested parties. An introspective look into the development and current state-of-the-art of design and manage- ment practices is conducted to better understand what these common metrics and methodologies should be. Although the pressing needs of the aging infrastructure problem and a growing interest in durability-based design present a great need for new innovations, the integration of monitor- ing technologies will likely be incremental. As such, how these technologies can benefit existing methods while serving as a catalyst for the development of new methods is of interest. Based on previous work of the authors and some recent developments, this paper examines these issues and highlights how new technologies can successfully be utilized as a catalyst to improve the design, assessment, and management of civil infrastructure.

REFERENCES

American Association of State Highway and Transportation Officials (AASHTO) Highway Subcom- mittee on Bridges and Structures, 2005. Grand Challenges: A Strategic Plan. Available online at: http://www.transportation.org/?siteid=34 Frangopol, D.M. 2007, Perspective from US Researchers on Bridge Long-Term Performance Program, FHWA/NSF Workshop on Future Directions for Long-Term Monitoring, Assessment, and Management, Las Vegas, Nevada, USA. Frangopol, D.M. & Liu, M. 2007. Maintenance and management of civil infrastructure based on condition, safety, optimization, and life-cycle cost. Structure and Infrastructure Engineering, Taylor & Francis, 3(1), 29–41. Frangopol, D.M. & Messervey, T.B. 2007a. Lifetime oriented assessment and design optimization concepts under uncertainty: Role of structural health monitoring. Proceedings of the Third International Conference on Lifetime-Oriented Design Concepts, Bochum, Germany, November 12–24 (keynote paper); in Lifetime- oriented Design Concepts, 2007, Stangenberg, F., Bruhns, O.T., Hartmann, D., and Meschke. G., eds., Aedificatio Publishers Freiburg, 2007, 133–145 (keynote paper).

391 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Challenges for asset management in W-Europe

Leo (H.E.) Klatter Ministry of Transport, Public Works, and Water Management, Center for Public Works, The Netherlands

Western European countries have to manage an ageing bridge stock that has to withstand increased traffic loads. A specific feature of bridges is the long service life time of order a hundred years. The probability of large changes in use is apparent over such a time span. In addition insight in behavior of materials and structural design develop over time and can lead to a change in insight in their performance and durability. This development leads to improvement of the design practice of new bridges. The question is how to judge the older bridges in the light of this knowledge. Especially as new insight in loads, material behavior and structural design are translated in formal design codes. Older bridges often do not comply with these design codes. This is not merely a theoretical problem because traffic loads have increased substantially in practice while at the same time the strength of the bridges has been reduced by deterioration. A major survey of all “elderly” bridges has been performed recently in the Netherlands. This survey showed structural reliability of a total of 1,180 concrete bridges in the highways network had to be assessed further. This about 30% of the total number of bridges. These structures are first assessed with simplified models based on current design codes. The research program is end of 2007 in halfway these assessments. One of the problems is the large number of structures with incomplete construction data of information. For 40 % of the structures in the detailed design data like reinforcement drawings are missing or there are no construction data at all. The research got media attention and an answer to questions why bridges are not up to standard, apparently just as a sudden was needed. For the bridge stock in the Netherlands such a historic perspective understandable for the public has been draw up. A short version has been described in the paper as an illustration. Assessment of structural reliability is often performed after problems due to increased traffic loads already present. It is preferable to foresee these problems. A prognosis of effects of traffic development on structural reliability is necessary for this. A proactive approach is needed, looking forward to future use of the structures. Prominent non visible risks, such as insufficient capacity to carry the heavy traffic, call for a different approach to inspections. The risks should be taken as the starting point for the inspections instead. Structural reliability, or lack of it, is one of the important risks. Asset management is regarded as a valuable methodology to cope with these challenges. An international team consisting of experts from (semi)governmental agencies or institutes and universities from England, Germany, Switzerland and France is assigned with the review on the research program in the Netherlands. It is interesting to see from the discussion in this review team how this sort of problems is dealt with in neighboring countries A similar development of loading the bridges to the limit is recognized. However the Netherlands seem to be ahead in development of traffic load related problems. In most cases structural non-compliant bridges are treated as incidents and not as a general development with increasing demands from traffic. In such a case tailor made solutions are sufficient. By the time a large fraction of the structures is involved an integral approach to the problem, soundly founded in regulations is inevitable. In the Netherlands this is definitely the case.

392 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Effective bridge management using ABMS

Hiroyuki Kawamura, Kazuhiko Kudo, Motoi Soma & Hidetaka Kawaragi Road Management Division, Aomori Prefectural Government, Japan Makoto Kaneuji Civil Engineering Management Division, Kajima Corporation, Tokyo, Japan

EXTENDED ABSTRACT: We have developed the Aomori Bridge Management System (ABMS) in 2004 and 2005, and began its bridge management in 2006 by using ABMS. ABMS is consisted of five steps; STEP1: Establish Basic Strategy STEP2: Establish Individual Bridge Strategy (Micro Management) STEP3: Establish Long-term Budgetary Plan (Macro Management) STEP4: Establish Mid-term Bridge Management Plan (Micro Management) STEP5: Evaluation and Feedback In STEP2, we select applicable Management Scenarios for every bridge by taking various factors such as roles in the road network into account, and calculate Life-Cycle Cost for each selected Man- agement Scenario based on the inspection data and deterioration prediction. In STEP3, we execute Budget Simulation to find the best and feasible Long-term Budgetary Plan for bridge preservation under the given budget constraint by changing the combination of Management Scenarios for all bridges. Through the Budget Simulation, we found that we can decrease a great amount of LCC by chang- ing the management policy toward the Preventive Maintenance and making intensive investment for the bridge preservation within five years. By showing the budget simulation study results, we were able get approval for the Long-term Bridge Management Plan with the necessary mid-term budget, and we established the Five-year Bridge Management Plan according to the approved Long-term Budgetary Plan. In the execution process of bridge management, most of the judgments have to be made by in-house engineers of the Asset Management Teams in each District Office. Therefore, we had several meetings with them to have the same understanding on the Five-year Bridge Management Plan. We also organized seminars and workshops to educate and train engineers of consulting firms and construction companies as well as in-house engineers. With an aim to bring up a culture to take care of the bridges in the vicinity by ourselves, we awarded a construction company a contract to do the whole Routine Maintenance works including annual inspection, cleaning, minor repair works on the bridges in the district. We believe that the Routine Maintenance is most effective and efficient to extend the bridge lives, and want to ask the construction company to play a role of home doctor of the bridges in the district. After one year operation of ABMS, we accomplished the bridge preservation works according to the Five-year Bridge Management Plan within the budget. Weare in the first stage of the execution of ABMS and we still need to make efforts to get useful information for the better bridge management such as deterioration process, deterioration prediction and effective maintenance measures.

393 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Development of BMS and possibility of performance based contracting using BMS

Makoto Kaneuji Civil Engineering Management Division, Kajima Corporation, Tokyo, Japan Hiroyuki Kawamura & Motoi Soma Road Management Division, Aomori Prefectural Government, Japan Eiichi Watanabe Kyoto University & Chairperson, Regional Planning Institute of Osaka, Osaka, Japan

ExtendedABSTRACT: There are very many road bridges in Japan, and they are exposed to severe environment and starting to deteriorate. Since the process of finding the best Bridge Management Plan under the budgetary constraint is complicated, a great amount of effort has been made to develop Bridge Management System. Most of these BMS are useful in the planning process, but not many are useful in the execution process of bridge management, such as maintenance works, repair works, rehabilitation and replacement works. Aomori Bridge Management System (ABMS) has unique and useful functions such as “Knowl- edge Base Deterioration Prediction Method”, “Management Scenario Concept” and “Budget Simulation”, and is consisted of five steps;

STEP1: Establish Basic Strategy STEP2: Establish Individual Bridge Strategy (Micro Management) STEP3: Establish Long-term Budgetary Plan (Macro Management) STEP4: Establish Mid-term Bridge Management Plan (Micro Management) STEP5: Evaluation and Feedback

In order to execute deterioration prediction and LCC calculation in STEP2, we need to indicate the management levels to execute maintenance measures. It is possible to have many different combinations of management levels and maintenance measures for each bridge, but it is more practical to limit the number of combinations from the management point of view. Therefore, we have established a concept of Management Scenario which has a combination of maintenance measures to be executed at designated management levels. After selecting applicable Management Scenario for every bridge by taking various factors such as roles in the road network into account, we calculate Life-Cycle Cost for selected Management Scenarios. In STEP3, we execute Budget Simulation to find the best and feasible Long-term Budgetary Plan for bridge preservation under the given budget constraint by changing the combination of Management Scenarios for all bridges. After the process of Budget Simulation, the maintenance scenarios of all bridge are determined, and we can establish the Mid-term Management Plan based on the determined Management Scenarios and the Long-term Budgetary Plan. In our BMS, when the long-term budgetary plan is finalized, Management Scenarios are determined for all bridges. Since performance requirements of all bridge components are defined for each Management Scenario, it is possible to indicate performance requirement for every bridge component. Therefore, there is a possibility to make a Performance Based Contracting by showing the Five-year Bridge Management Plan obtained by our BMS to the contractor.

394 The Performance Based Contracting has been widely adopted by many States’ Department of Transportation in the United States for highway maintenance works, because it is effective, efficient and best suited for the maintenance work. The Performance Based Contracting itself is not popular in Japan yet, but it will soon be recognized as the best suited contracting method for the maintenance works as the amount of maintenance work increases in the near future.

395 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

A study on LCC prediction for bridge management taking future uncertainty into account

Koichi Mitsunari, Yuji Takahashi & Yoshiteru Otani Civil Engineering Design Division, Kajima Corporation, Tokyo, Japan

Extended ABSTRACT: The averaged age of Japanese existing road bridges is approximately 30–40 years. A large number of bridges that exceed their service lives are to be repaired or replaced within 20 years to come. In order to avoid the concentration of repair and replacement, we need planned maintenance and asset management. Bridges require special maintenance and management because of the large amount of compli- cated deterioration incidents encountered during their lifespan. Bridge Management Systems have to reflect these deterioration effect practically thus enabling us to predict future deterioration rates precisely for asset management. Accurate prediction of future deterioration rates is quite important for estimating the future maintenance cost. Individual bridge deterioration incidents however differ significantly according to their environmental condition, material qualities, etc. Future deteriora- tion speeds therefore should be predicted with their mean values and some range of fluctuations. By evaluating this range, we could estimate future risks caused by the uncertainty of deterioration phenomenon. This paper introduces some practical method to improve the existing Bridge Management System employing actual bridge inspection data. First, we examine the deterioration trend by inspec- tion data. Secondly we propose a revising method of deterioration curves prepared for individual deterioration mechanisms in advance, by analyzing deterioration trends from inspection data to improve deterioration predictions. Finally, we demonstrate the LCC estimation with some range of fluctuations of deterioration speeds by evaluating the parameters from inspection data. Applying Modification Index, the ratio of the duration time of inspection data to that of original deterioration curves, as an indicator, the difference between the prediction by the original deterio- ration curves and inspection data were evaluated. In most cases in this study, actual deterioration speeds estimated by inspection data are slower than those of predictions. Employing actual inspec- tion data, revising method of deterioration curves based on Modification Indexes shall improve deterioration prediction. LCC prediction with some range of fluctuations of deterioration speeds could contribute decision making of maintenance strategies. The existing Bridge Management Sys- tem will be developed employing these functions thus contributing the precision of the prediction of deterioration rates and LCC.

396 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Load testing and analysis of bridges missing critical documentation

John Prader, Jeff Weidner, Hesham Hassanain, Franklin Moon, Emin Aktan Drexel Intelligent Infrastructure Institute, Drexel University, Philadelphia, Pennsylvania, USA Frank Jalinoos, Burt Buchanan, Hamid Ghasemi Federal Highway Administration

Extended ABSTRACT: The writers had an opportunity for research at West Virginia for formulat- ing and demonstrating a methodology that can reliably establish the safe load carrying capacity of aged bridges missing critical documentation. Two bridges, including a two span slab and a single span arch bridge were selected as test specimens and were rated at less than 1 usingAASHTO LRFR guidelines and AASHTO recommended material properties. A distributed sample of concrete cores was obtained from each structure to check for uniformity. After taking the cores it was found that the Tams Slab had different concrete mixes for the different spans, while Amigo Arch was uni- form. Rebar samples were taken from both bridges at the locations of exposure by WVDOT staff for material testing and size determination. Additionally, FHWA completed pachometer scanning to determine approximate rebar spacing. Steel rebar coupons were also obtained and tested. The material test results were used to update the preliminary AASHTO load ratings. After updating the ratings the Tams Slab Bridge still rated below 1 for flexure but above 1 for shear. However, the Amigo Arch Bridge rated above 1 for flexure and shear. To be confident in rating values the writers proposed load testing each structure to establish the actual load capacity. The structures were load tested by trucks up to diagnostic load levels and then up to proof load levels. The final proof loads for the Tams Slab and Amigo Arch bridges were 280 kips and 205 kips respectively. Both proof load tests were successful in demonstrating the bridges’ ability to carry a very high load without visible damage or distress. The Tams Slab Bridge showed steel strain levels of 30 microstrain (tension) while the concrete strains on the underside of the slab reached around 20 microstrain (tension). The maximum concrete strains were observed in the parapets and were around 35 microstrain in compression. The maximum displacement observed in Span 1 was 0.019 in (0.50 mm) while the maximum displacement in Span 2 was 0.027 in (0.68 mm). The readings are very linear for Span 2 through each load step. However, when the load was increased in Span 1 from 143 kips (636 kN) to 280 kips (1245 kN), there was a stiffening of the bridge which resulted in the readings becoming non-linear between the final two load steps. The non-linearity was most prevalent towards the downstream side of Span 1. The Amigo Arch Bridge experienced maximum steel rebar strains of 20 microstrain when the loads were placed near the midspan of the arch. The corresponding concrete surface strains reached a maximum of 25 microstrain in compression. These readings occurred at the 1/4 span of the arch when the load was placed at the midspan. The displacements of the bridge varied between 0.002 in (0.20 mm) to 0.043 in (1.10 mm). The largest displacements were recorded at the midspan of the structure. Both bridges exhibited small displacements and stresses under proof loads. However, because of cracking, spalling, and some level of support softness, the measured responses differed from analysis. To justify, interpret and validate the experimental results, 3D finite element models were constructed and calibrated to simulate the actual measured behavior, revealing how deterioration and distress affected bridge capacity. Each structure was then rated using the field calibrated FE models and found to have rating factors exceeding four due to significant reserve capacity inherent in such cast-in-place RC bridges. The resulting load ratings were adequate for WVDOT to remove the postings on the bridges.

397 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Integrating health monitoring in asset management

Hitoshi Furuta, Hiroshi Hattori, Takuya Ohama & Kohei Yoshida Kansai University, Takatsuki, Japan Dan M. Frangopol Lehigh University, Bethlehem, USA

1 INTRODUCTION

In this study, an attempt is made to develop an optimal maintenance planning system that can account for the uncertainties involved in the prediction of the deterioration curve. Furthermore, it is also attempted to introduce the seismic risk into the maintenance planning with the aid of the structural health monitoring system. In the proposed system, an optimal maintenance scheduling is obtained by using Genetic Algorithm (GA) and the system can update the maintenance schedule by using the data given from the structural health monitoring. By updating the maintenance schedule, the seismic risk can be reduced. Several numerical examples are presented to demonstrate the applicability and efficiency of the proposed method.

2 NUMERICAL EXAMPLE

A numerical examination for a bridge is presented. The uncertainty of the deterioration curve is expressed by the normal distribution. In this numerical example, it is assumed that the health moni- toring system can update the deterioration curve every year. The result of the proposed system with health monitoring system is compared with the systems without health monitoring. The simulation is done 100 times and the result is shown in Table 1. Table 1 shows that the proposed method with health monitoring can provide less seismic risk than that without health monitoring, by using equation (8). This result shows the proposed system can adapt to the uncertainties of the deterioration by introducing the health monitoring system. Since the maintenance cost of the proposed system is almost the same, the effectiveness of the proposed system is verified.

3 CONCLUSIONS

In this paper, an attempt was made to develop a maintenance scheduling system that can consider the maintenance cost and seismic risk at the same time by introducing the health monitoring system and GA. Through the numerical simulation, the effectiveness of the proposed system is verified. The proposed system can reduce the seismic risk without increasing the maintenance cost by introducing the health monitoring system. Table 1. Seismic risk. Monitoring

Seismic Risk Without With

Max 180477 (Yen) 155767 (Yen) Min 169817 (Yen) 155272 (Yen) Average 176872 (Yen) 155469 (Yen)

398 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Integrative research supporting decision making for bridges

F. Necati Catbas University of Central Florida, Orlando, FL, USA Dan M. Frangopol Lehigh University, Bethlehem, PA, USA A. Emin Aktan Drexel University, Philadelphia, PA, USA

ABSTRACT: There has been a large body of research in the area of bridge health monitoring and objective condition assessment for bridge management. However, there is still an urgent need to better integrate technologies, methods and decision making for the reliable safety evaluation, enhanced maintenance and operation of bridges. The condition status of the bridges in the USA is still not at acceptable levels. The 2007 cost of improving the Nation’s roads and bridges to the levels needed is estimated to be $155 billion, according to the American Association of State Highway and Transportation Officials. The backlog is increasing as a result of approximately $75 billion in annual spending by federal, state, and local governments combined falls short of levels needed just to maintain the status quo. With limited funds and recent bridge failures, it is even more important to understand the actual condition of the bridges by means of novel yet available technologies. However, technological advances alone are simply not sufficient. Research on health monitoring or objective condition assessment presently remains detached from decision making. To connect the linkage between these, we need champion-owners, and, over the longer-term, organizational and societal reform. With all these pieces coming together, we can expect to develop better strategies and prioritization for maintenance and replacement decision making. The writers have been conducting research that supports the decision making process. In this paper, the writers will review their work over the last decade in the areas that are critical for bridge management and decision making. The review will include basic definitions, representative studies related to health monitoring of major bridges as well as population of bridges; structural identification; damage assessment; use of information technologies; probabilistic analysis for reliability estimation and deterioration modeling for bridges and bridge management based on safety and life cycle costs. Lessons learned from these studies will be presented along with examples and a vision for future work that is need for best practices for bridge decision making.

399 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Development of affordable GPS displacement monitoring system

M. Saeki Science University of Tokyo, Noda-shi, Chiba, Japan K. Oguni & M. Hori The University of Tokyo, Bunkyo-ku, Tokyo, Japan

ABSTRACT: For the assessment of civil structures, deformation is useful information. There- fore, the authors have been trying to develop a system which monitors displacements densely by combining the technologies of GPS and wireless sensor network. In this system, the sensor nodes with an L1-receiver measures L1 carrier phases and send their measurements to a central server machine by wireless communications. The central server machine estimates displacement of the sensor nodes with high accuracy by carrying out relative positioning. In order to distribute the sensor nodes densely at a reasonable cost, the sensor nodes are required to be 1) of low power consumption, 2) small and light and 3) inexpensive. To satisfy these requirements, a small patch antenna used for a vehicle navigation system is connected to an inexpensive L1-receiver in the sensor nodes. The patch antenna is not used for measuring L1 carrier phases in general because the measurements are highly contaminated by multi-path noises and the signal-to-noise ratio is low. Therefore, to obtain accurate displacements of sensor nodes using a small patch antenna, it might need observing data for a long time. On the other hand, the sensor node has other limitation on battery. The size of a battery which is connected to the sensor node is small in order to satisfy the requirements mentioned above. Observing long data results in high energy consumption and it shortens a life of sensor node due to the limitation of small battery. To overcome this problem, we have developed an efficient algorithm for estimating displacement based on noisy measurements with short data length. In the proposed algorithm, a time series of residual DD (Double Differenced) carrier phases is modeled as a linear function of time. While this modeling is only applicable to the static problem, it improves the condition number of coefficient matrix of normal equations. As a result, the success rate and the accuracy of displacement are improved especially in case of short data length. The linear function of residual DD carrier phases is estimated through the least squares method. The weight values are evaluated according to signal-to-noise ratio output from an L1-receiver. In order to evaluate the performance of the present system, filed experiments are conducted in an ideal condition and realistic condition. First, we carry out an experiment on the top floor of a building as an ideal condition. In this experiment, two patch antennas are fixed on the concrete floor and other two patch antennas are fixed on the tripods in order to investigate the difference of effects of multi-path noises on the estimation. In case patch antennas are fixed on the concrete floor, the success rate is almost 100 percent and the accuracy is less than 1 cm when the data length is three minutes. In case the patch antennas are fixed on the tripods, the success rate is about 96 percent and the accuracy is about 1 cm in horizontal and 2 cm in vertical. In the latter case, the measurements are largely affected by multi-path noises. Another experiment is carried out in a realistic condition in which the present system is installed to an embankment. In this experiment, several sensor nodes are deployed on the surface of embankment, and a sensor node is moved by 20 cm during the observation. This displacement is measured by the present system.

400 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Methods for measuring structural deflection and applications to bridge deck performance monitoring

J.M.W. Brownjohn Vibration Engineering Section, University of Sheffield, UK X. Meng Centre for Geospatial Science, University of Nottingham, UK

ABSTRACT: Deformation measurements play a vital role in the assessment of bridge perfor- mance, and many SHM systems have aimed to track movements of the superstructure using a wide range of techniques. These include many types of hydraulic and optical systems, recovery from velocity, acceleration or strain signals and latterly GPS. The paper reviews these techniques, with special reference to long span bridge deflection moni- toring and presents the authors experience on several examples, with the conclusion that there is no clear winner among the various techniques: they all have their disadvantages and each case needs to be considered in terms of required resolution, cost, acquisition rates and compatibility.

1 INTRODUCTION

Bridge decks are among the most flexible forms of civil infrastructure, particular for long-span sus- pension and cable-stayed bridges. As such, serviceability in terms of deflection may govern design. Deflections are an integral part of in-wind performance of bridges both statically and dynamically, and mechanisms giving rise to deflections are particularly important to aerodynamic and aero- elastic instability. Naturally, design against unsatisfactory in-wind performance requires analytical or physical models, both of which need to be referenced to full-scale evaluations. Even when satis- factory from a structural point of view, deflections can be unsatisfactory from an operational point of view, also requiring evaluation at full-scale.

2 DISPLACEMENT MEASUREMENT TECHNOLOGIES

The technologies reviewed in this paper include laser and LED devices, image tracking via CCD arrays, optical marker tracking , close range non-contacting devices, GPS, total sta- tion/electronic distance measurement (EDM), pneumatic systems, microwave interferometry, contacting displacement measurements and use of acceleration, velocity, strain or rotation signals. The characteristics of these methods are described and optimal strategies suggested for bridge deflection measurement. Many references to applications and descriptions of the technology are provided.

401 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Experimental identification of multiple oscillation frequencies using GPS

P.A. Psimoulis Geodesy Laboratory, Department of Civil Engineering, Patras University, Patras, Greece S. Pytharouli Geodesy Laboratory, Department of Civil Engineering, Patras University, Patras, Greece currently at Dept. Of Civil Engineering, University of Strathclyde, Glasgow, Scotland S. Stiros Geodesy Laboratory, Department of Civil Engineering, Patras University, Patras, Greece

ABSTRACT: A serious limitation in the use of GPS in the monitoring of important structures like bridges, and especially in the identification of their modal frequencies, is whether this geodetic sensor can identify more than one modal frequency. In order to contribute in the solution of this problem, we made a large number of experiments with a linear oscillator of predetermined characteristics and with three degrees of freedom and max displacements up to 3.4 cm. A GPS receiver was mounted on the sliding mass most remote from the generator and was recording in kinematic mode simultaneously with another receiver in nearby, stable position, both sampling at 20 Hz rate. Experiments were made with different constant excitation frequencies, from 0.05 to 4 Hz. Recorded displacements were analyzed using spectral analysis techniques based on least-squares analysis, suitable for processing of short and non-equi-spaced time series. In the steady part of the oscillations with constant excitation frequencies, computed frequencies were practically equal with the input ones. For the transient part of the oscillations, excitation frequencies as well as modal frequencies were accurately identified, but the amplitude of the latter was smaller, as expected. However, in the case of very small oscillation amplitudes, high (>2 Hz) modal frequencies differ by up to 6% from the real values. These results are very promising for they were obtained from signals a few seconds long and with very small amplitude. The conclusion is therefore that GPS can identify multiple frequencies, including transient and modal ones in normal structures, especially in flexible ones in which the oscillation amplitude is larger than those tested and leads to a higher signal-to noise value.

402 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Real-time dynamic monitoring with GPS and georobot during Sutong Bridge construction

S.X. Huang & B.C. Yang School of Geodesy and Geomatics, Wuhan University, Wuhan, Hubei, China

ABSTRACT: Sutong Bridge is the largest cable-stayed bridge in the world and the construction process is complicated. It is located in lower reaches of Yangtze River and the effects of monsoons and diurnal cycles of temperature change are severe, which will bring great difficulties for construc- tion control. In expectance of significant deformations of the pylon towers and cantilevers caused by wind loads and thermal effects, a deformation monitoring system using GPS and georobot is proposed. Major goals are detecting the deformations of pylon towers and cantilevers in real-time and analyzing the dynamics of the bridge so as to ensure safety during construction. The chal- lenges include the required high accuracy and reliability of the monitoring system which should work automatically and consecutively for nearly one year in all weather conditions. Conventional measurement techniques for deformation monitoring can’t meet the requirement of large structures like Sutong Bridge. This paper introduces the general situation of the bridge briefly and outlines the difficulties of construction in the second part. Then it gives a specific description of the monitoring system, which consists of two subsystems based on GPS and georobot respectively, and its measurement scheme. Emphasis is laid on the processing and analysis of the time series data. Coordinate transform is investigated and the coordinate system of the two subsystem is unified. With two subsystems work independently, results of one subsystem could serve as check for another so as to improve the reliability. Correlation analysis is used to determine the relationships between displacements of the bridge and influential factors, such as thermal effects and wind loads. The results indicate that bridge pylon towers are affected by thermal effects with strong inverse correlation and their behaviors lag behind, while the impact of wind loads is nonlinear. Spectrum analysis is performed to determine the fundamental frequencies of different construction situations and abnormal environmental conditions. Data collected by GPS proved to be good for this analysis and changes in the fundamental frequencies could indicate the health status. The dynamic deformation monitoring system worked consecutively during superstructure con- struction of Sutong Bridge. Long-term results provided excellent means for understanding the complex structure and ensured the safety during construction.

403 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

The safety assessment method of existing large span steel structural members

Xiao Liu & Yongfeng Luo College of Civil Engineering, Tongji University, Shanghai, China

ABSTRACT: Large span structures are widely used in modern buildings because of their novel forms or shapes, light self-weight and lower construction cost. The large steel structures are built as the national economy and the people’s livelihood buildings, so the safety during their service periods is widely considered. The reasonable safety assessment method of existing large span steel structures is one of the national civil engineering disaster prevention subjects to be solved. However, the current safety assessment methods are mostly focused on existing concrete structures. Few papers or research references about existing large span steel structures are found. Based on the systematic science method and according to the reliability index of the structural members under ultimate state specified in the reference, a new safety assessment principle of existing steel structural members is proposed. The load model and the resistance model are discussed in the paper. The safety grades of members of existing large span structures are ascertained by the reliability programme. The safety assessment method of steel members of existing large span structures is proposed.

Keywords: the existing large span steel structure; load model; resistance model; safety grades of members; reliability; safety assessment

REFERENCES

Unified standard for design of building structures (GBJ68-84). The architectural industry press of China, Beijing Unified standard for reliability design of building structures (GB 50068-2001). The architectural industry press of China, Beijing The design basis of structures—the assessment of existing structures (ISO/CD 13822). Translated by the inspect and appraisal station of Xi’an Architectural University of Science and Technology, 2001 Andong Yu, Runxiu Ye. The safety and reliability of architectural structures. The science and technology literature press in Shanghai, 1985 Standard for appraiser of reliability of civil buildings (GB 50292-1999). The architectural industry press of China, Beijing Jihua Li, Zhongmin Lin. The reliability limit state design of architectural structures. The architectural industry press of China, Beijing, 1990

404 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

A mechanical model of steel frames with joint damages

Yongfeng Luo & Huaijin Song College of Civil Engineering, Tongji University, Shanghai, China

The mechanical behavior of steel frames will be in nonlinear state under strong earthquakes. The stiffness and the strength of the structures may degrade. This degradation has significant effects on seismic performances of the structures in later service stage. Based on the fracture mechanics, the behavior of the bolt-seam joints with cracks in welding seams of the steel frames are analyzed in this paper. The effective length and the effective depth of a crack are derived. The relationship between bending moment (M) and rotation (θ) of the damaged rigid connection is obtained. The stiffness matrix of a beam with damaged joints is also derived. Numerical analysis is also conducted to ascertain the effects of joint stiffness on the internal forces of the damaged steel frames. The results show that the joint damages have significant effects on the internal forces of the frames. Therefore, joint damages must be taken into account to ascertain the load bearing capacity of steel frames.

Keywords: fracture mechanics, rotational stiffness, stiffness matrix, load bearing capacity

REFERENCES

Chen Huifa & Zhou Suiping. 2001. Stability Design of Steel Frame. Shanghai: World Book Publishing Company, Christopher JE & Bjorhovde R. 1999. Semi-rigid frame design methods for practicing engineering. Engineering Journal, (1) Ding Suidong & Sun Limin. 1997. Fracture Mechanics. Beijing: China Machine Press, Dugdale D S. 1960. Yeilding of steel containing slit [J]. Journal of the Mechanics and Physics of Solids, 8:100∼104. V.V. Bertero, J.C. Anderson & H. Krawinkler. 1994. Performance of Steel Building Structures during the Northridge Earthquake, Report NO. UCB/EERC—94/09, Earthquake Engineering Research Center, Univ. of California, Berkeley. Zhou Xuejun & Zhang Xianglong. 2003. The researches and applications for steel frames with semi-rigid connections[J]. Journal of Shandong Institute of Architecture and Engineering, 18(2):85–88

405 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Statistic analysis of a prototype structural health monitoring system for the Nanpu Bridge in Shanghai, P. R. China

R. Wang Department of Building Engineering, Tongji University, Shanghai, China X. Meng Institute of Engineering Surveying and Space Geodesy, The University of Nottingham, Nottingham, UK Y. Luo Department of Building Engineering, Tongji University, Shanghai, China L. Yao Department of Surveying and Geoinformatics, Tongji University, Shanghai, China W. Huang Intelligent Transportation Systems Research Centre, Southeast University, Nanjing, China

ABSTRACT: In this paper, we present a statistic model and preliminary analysis results of a prototype Structural Health Monitoring (SHM) system for the Nanpu Bridge in Shanghai, China. The proposed SHM system consists of two key components, i.e. data acquisition system and the statistic model that form basis elements of a damage identification and assessment system. In the prototype SHM system, the Real-Time Kinematic Global Positioning System (RTK GPS) was employed to acquire in-situ ambient dynamic responses of the civil structures and huge amount of raw data was obtained during field data acquisition. Three dimensional coordinate time series were obtained from these raw data and used in this paper for the creation of the statistic model and the analysis of the dynamics system of the civil structures. In a recent practice conducted in September 2006, 14 dual frequency GPS receivers were employed to acquire in-situ ambient dynamic responses of the Nanpu Bridge for four consecutive days. There are 12 monitoring points that were localized along the Nanpu Bridge and two other GPS receivers were installed on two well measured stationary locations as the reference stations. The data processing was carried out to obtain the coordinate time series of these 12 monitoring points. From identification viewpoint, a general dynamic system can be expressed as a state space representation and various statistic models. The statistic models can be used to predict the abrupt changes of dynamics characteristics. These GPS coordinate time series can be used to extract the statistic models of the SHM and various statistic analyses can be carried out which form the main contents of this paper.

406 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

The analysis of GPS single epoch positioning algorithm based on the deformation monitoring

L. Yao & P.Yao Department of Surveying and Geoinformatics, Tongji University, Shanghai, China X. Meng Institute of Engineering Surveying and Space Geodesy, The University of Nottingham, Nottingham, UK

ABSTRACT: GPS has the advantage of long-term stability, high automation in data collection, all weather observation and precise time synchronization, and the accuracy of short distance deforma- tion monitoring can reach millimetre level. This makes such a system be a viable tool in monitoring the large structures, such as dam, bridge, high building etc. with a high precision. In the appli- cations of dynamic monitoring of various slender structures, the coordinates of every observation epoch should be calculated. Most commercial software packages have benn implemented with an On-The-Fly (OTF) algorithm in their data processing engines. However, this solution needs a period of initialization, and it has difficult to apply in the circumstance when the GPS satellites’ signal interruption and cycle slips occur frequently which is typical in bridge deformation moni- toring. Therefore, a further study of GPS single epoch positioning algorithm is becoming a critical necessity. It is well known that the measurements in one observation epoch are not adequate in the computation of coordinate time series of a monitoring site. So other constraints shall be utilised in the coordinate estimation. Example constraints include integer constraint of ambiguity, coordinate constraint of a platform attached, application of P code pseudorange, etc. Considering structural monitoring applications, relatively high accuracy approximate coordinates of a monitoring site can be easily obtained. For example, the accuracy of approximate coordinates is tens of millimetres level in dam and slope monitoring, and the amplitudes of instant deflection of high buildings and large-size bridges are not very large, around 0.5m. It provides some advantageous conditions for the resolution of single epoch positions. Recently, most GPS single epoch positioning algorithms comprise different types of search methods based on the Least-squares algorithm. In these search methods coordinate constraint and integer constraint of ambiguity are used to establish a search space, and then V T PV values are calculated with the Least-squares algorithm. The coordinate, which has a minimum V T PV value and passes a Ratio-test, is determined as the final result. This paper analyzes and summarizes dual-frequency GPS single epoch positioning algorithms in the context of deformation monitoring. Firstly, key GPS single epoch positioning algorithms and their feasibility are presented. Then three single epoch algorithms which are used in the data processing of deformation monitoring are introduced. Finally, three real-life examples are used to demonstrate the feasibility and practicability of each algorithm. Meanwhile, the defects of two well-known algorithms are analyzed and ameliorated.

407 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Deformation analysis of the supporting towers of the Nanpu Bridge from GPS measurements

L. Yao,Y. Xie & Y. She Department of Surveying and Geoinformatics, Tongji University, Shanghai, China X. Meng Institute of Engineering Surveying and Space Geodesy, University of Nottingham, Nottingham, UK

ABSTRACT: This paper presents a research of deformation analysis of the supporting towers of Nanpu Cable-stayed Bridge in Shanghai, China. This bridge has two towers of 154 m high and was the third longest cable-stayed bridge in the world with a total length of 8346 m. GPS data of several sessions have been collected from 21 to 24 September 2006. Each session lasted 2 hours. During the trail, 14 high grade geodetic GPS receivers were used and 1 GPS receiver was installed on each of the two supporting towers, observing at a data rate of 10 Hz. Coordinate system transformation and data pre-treatment were conducted and spectrum analysis was used to obtain the power spectral density that illustrates the response of the structure activated by the random dynamic loading. It is found that the vertical vibration frequency of supporting towers is stable with a range from 0.039 to 0.044 Hz and the horizontal frequency changes dramatically. According to the result of cubic spline fitting, the largest amplitude of the supporting towers occurred in the first session on 22 September 2006. The vertical amplitude value of one of the supporting tower 0.0521 m and the horizontal deformation is 0.0579 m, whilst the vertical amplitude value of another supporting tower is 0.0437 m and the horizontal displacement is 0.0549 m. Average amplitude value of two points is 0.034 m both in vertical and horizontal directions. In addition, correlation analysis is employed to obtain the deformation correlation information about two towers and it is found that the horizontal correlation is slightly more obvious than the vertical direction although horizontal correlation is also weak. According to the analytical result, we identified the deformation characters of the supporting towers and concluded that GPS can be used as a trustworthy tool to characterize the deformation of large bridge.

408 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Deformation monitoring and analysis of high pylon of Su-Tong Bridge in construction

Yue Dongjie, Wang Chaoling & Li Hongxiang College of Civil Engineering, Hohai University, NanJing, P.R.China

ABSTRACT: Su-Tong Yangtze River Bridge is located in the southeast part of Jiangsu Province and connects Nantong and Suzhou city in China. The total length is 8206 m. The main bridge is a double-pylon and double-cable-plane steel box girder cable-stayed bridge with a world record breaking 1088 m main span. The bridge has more than 300 m high invertedY concrete pylons, which includes down, middle, upper piles. There are some key technique problems in construction as other bridges. For example, the construction surveying and controlling is one of key problems. Because of the influence of foundation displacement, the cable pylon deadweight, concrete shrinkage, elastic compression, creep, temperature, temperature difference by sunlight radiation, wind power, wind direction, mechanical vibration, and etc. in construction, the deformations of cable pylon will occur, which can far exceed the demand of measurement precision and even affect the safety of the structure, and so they must be considered in construction surveying and controlling. Therefore, the offset and deformation law of the cable pylon must be monitored at any time, and then the deformation situation can be mastered and verified in time under different work conditions, so that effective measures are taken to adjust the layout of cable pylon, and to ensure the safety and construction quality of cable pylon. In this paper, firstly the deformation reason and consequence are analyzed. Secondly, combining with the characteristics of high cable pylon of Su-Tong bridge, which pylon is 300 m high, the monitoring plan is elaborated, includes the choice of monitoring method, precision estimation, monitoring point distribution, monitoring time, observations and so on. In order to collect monitoring data of points setting on pylon quickly and accurately, the software based on TCA2003 is developed, which can realize monitoring automatically and repetitively. Finally the monitoring data is analyzed, the deformation law of pylon in construction is summarized. Monitoring result shows that the displacement of cable pylon in direction of vertical to the bridge axis matches practical case, and is consistent with the design which provided pre-displacement in the bracing demolition of cable pylon. The conclusion has the very good reference value to confirm the design, the instruction for following construction as well as the scientific research and so on.

409 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

The statistical investigation on one year GPS monitoring data from Donghai Bridge Health Monitoring System (DHBHMS)

D. Dan & L. Sun State Key Laboratory for Disaster Reduction in Civil Engineering, Tongji University, Shanghai, China X. Meng Institute of Engineering Surveying and Space Geodesy, The University of Nottingham, Nottingham, UK D. Xie Donghai Bridge Management Ltd, Shanghai

ABSTRACT: In this paper, the design and the implementation of GPS monitoring sub-system of the DHBHMS are introduced firstly, including the initial design motivation of GPS monitoring items, the special considerations on measurement point displacement and the data process scheme, and the advanced GPS coordinate resolution algorithms and its validity. After a test run lasting one year and 6 months, the GPS data collected from two cable stayed bridge are investigated in a viewpoint of statistical way. The distribution type test of the real life GPS data is verified and the cumulative distribution function is developed. The comparison between GPS and acceleration integrated displacement is made. Some useful information of the structural behaviors from a view- point of civil engineering is introduced by the authors. GPS is a suitable technique for geometry monitoring of large span bridges. Both static control measurement and real-time kinetic approaches are investigated in the bridge monitoring applications. In recent years, many investigations on GPS are performed in China. The applications in bridge monitoring systems are also ongoing in China. Donghai Bridge health monitoring system (DHBHMS) has a well designed GPS subsystem with 9 GPS monitoring points and other relative independent data acquisition, communication, and control and process systems. The authors were chiefly in charge of the whole system design implementa- tion, and the data analysis project. This paper is only limited to the introduction and validation of the long terms GPS subsystem and the extraction of useful information from a viewpoint of civil engineering. The reference station is lying in the island of Dawugui. The monitoring station layout on these two bridge are illustrated in Figures 1(b) and 1(c). Taking the reference station into account, 9 GPS receiver stations are all set up in open sea circumstance. The GPS subsystem is composed of three parts, i.e. a monitoring unit, a data transmission and console unit, and a data process and management unit. The GPS receivers are Trimble 5700, and the highest update rate is 10 Hz. An extended triple-differential Kalman filter model using a combined value to eliminate the ionospheric delay was developed to overcome some shortcoming in traditional RTK algorithms. This system is proved to be feasible with the analysis and comparison to the monitoring records.

410 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Nonlinear dynamic responses of large span hybrid structures under multi-dimensional seismic excitation

Yun Huang & Yongfeng Luo College of Civil Engineering, Tongji University, Shanghai, China

Large span hybrid structures are widely used in modern steel structures for their remarkable spanning capability and beautiful appearance. They usually consist of beams, struts or similar components and flexible tension rods or cables. Considering the probability of slackness of tension cables under earthquake, it is necessary and important to learn the dynamic responses of large span hybrid structures under seismic excitation. The response of a large span hybrid steel structure subjected to multi-dimensional seismic exci- tation is analyzed by use of the finite-element program ANSYS. The nonlinear behavior of cable elements is considered. The optimization of the cable force is carried out and the vibration modes of the structure are obtained for understanding the earthquake-resistance performance of the beam- cable structure. The dynamic performance of the structure under different pre-stress conditions is obtained.

Keywords: large-span hybrid structure, seismic excitation, optimization, cable.

411 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Research into the use of GNSS to monitor the deflections of suspension bridges, and the role of the FIG in deformation monitoring of bridges

G.W. Roberts & X. Meng Institute of Engineering Surveying and Space Geodesy, The University of Nottingham, Nottingham, UK C.J. Brown School of Engineering and Design, Brunel University of West London, Uxbridge, UK

ABSTRACT: The International Federation of Surveyors (FIG) is an umbrella body covering all aspects of surveying. It is split into various commissions and working groups and task forces, and organizes meetings and conferences to bring like minded people together. One Working Group, entitled “WG 6.4 Engineering Surveys for Construction Works and Structural Engineering” focuses on various aspects of surveying for engineering and construction. The Working Group has the following policy issues: • Promoting the use of adapted survey techniques in industry & Engineering; • Promoting a multidisciplinary collaboration between survey engineers, civil engineers, struc- tural & mechanical engineers; • Promoting the understanding of fibre optic sensors, e.g. interferometric sensors, Brillouin and Raman scattering and Bragg gratings; • Study the use of embedded sensor arrays and the role of advanced surveying techniques for structural monitoring; • Creating an awareness of surveyors through a task force ‘Fibre optic sensors’ of the rapidly emerging technology of fibre optic sensors as “non-geodetic” sensors to measure deformations (strain) and temperatures in civil engineering structures In addition to which, it has the following specific projects: • Precise methods and equipment for staking out during construction and structural works; • QC and documentation for as build compared to as designed; • Precise methods and equipment for Engineering surveys for visualisation and photo match; • Precise methods and equipment for remote surveys. (Terrestrial laser scanners etc.) • Dynamic Monitoring of Buildings and Structures; • Offshore construction surveys. The following paper details the role of the FIG in promoting research work carried out into the use of GPS and GNSS for deformation and deflection monitoring of structures, notably bridges. This is carried out through various working groups and task forces. In addition to this, the paper outlines work that has been carried out for more than a decade by the Universities of Nottingham and Brunel in using GPS and GNSS to monitor the deflections of a number of bridges, namely the Wilford Suspension Bridge in Nottingham, the London Millennium Bridge, The Humber Bridge and the Forth Road Bridge.

412 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Deflection monitoring of bridges: A case study of the Forth Road Bridge

X. Meng & G.W. Roberts Institute of Engineering Surveying and Space Geodesy, The University of Nottingham, Nottingham, UK C.J. Brown School of Engineering and Design, Brunel University of West London, Uxbridge, UK

ABSTRACT: The Forth Road Bridge spans the Firth of Forth, linking Fife and the North with Edinburgh and the central belt, and provides a vital strategic link to the rest of the UK for industry and tourism in the north and east of Scotland. It was opened to traffic on 4 September 1964. At that time it was the largest suspension bridge in Europe, fourth largest in the world, with a main span of 1005 m and the overall length of 2.5 km. Traffic has steadily increased over this bridge, from 4 million vehicles in 1964 to about 12 million in 2004 (www.feta.gov.uk). In addition, the heaviest commercial vehicles weighed 24 tons but the current limit is set to 44 tons. When the bridge was opened, it brought to an end an 800 year history of ferry-boat service across the river at the Queensferry in Scotland. A feasibility study into the use of GPS to measure the magnitude of the movements and the vibration frequencies of the Forth Road Bridge was carried out during February 2005. The objectives of the study were to validate the effectiveness of GPS to measure the deflections of the bridge, as well as the vibrations of the structure. During the trial state of the art dual frequency code/carrier GPS receivers with an output data rate of 10 Hz were employed. Measurements were taken over a period of 46 hours at 5 locations on the bridge deck and two atop the southern towers. In addition, two reference stations were located on the observation platform adjacent to The Forth Estuary Transport Authority (FETA) building. Altogether, 11½ million 3D data points resulted. The resulting coordinates from each GPS receiver located upon the bridge were given in 3D and with a precise time; therefore all are synchronized and output absolute positional information. “On-The-Fly” kinematic GPS processing in a post-processing manner was utilized and subsequent data filtering leaded to millimeter precision “movements” measuring bridge deformations as they varied with time. Hence, some dynamic characteristics can also be determined. A complete analysis of all the data points is beyond the scope of this feasibility study. Whilst further processing and analysis is still underway, the results presented focus on specific occurrences upon the bridge. In the following sessions the authors will firstly introduce the field trial, including instrumentation, types of receivers used and data acquisition arrangement. It is followed by the presentations of obtained results obtained through processing some sample data sets and relevant analysis of the structural dynamics. The paper is concluded with a summary to the trials and some useful recommendations to the future work.

413 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Recent progress in GNSS-based long bridge deformation monitoring

X. Meng, G.W. Roberts & A.H. Dodson Institute of Engineering Surveying and Space Geodesy, The University of Nottingham, Nottingham, UK L. Xu & Z. Wan School of Civil Engineering, Wuhan University, Wuhan, China

ABSTRACT: Invented more than three decades ago, Global Positioning System (GPS) has revo- lutionized people’slives and the latest high grade GPS receivers can provide instant 3D positioning accuracy of a few centimeters at a sampling rate of 50 Hz or higher. This achievement makes GPS an indispensable modern “utility” and its new applications are continuously to be exploited. At the moment, there are only two independent operational satellite positioning and navigation systems which are called generically as Global Navigation Satellite Systems or GNSS. These systems are US’s GPS and Russia’s GLObal NAvigation Satellite System (GLONASS). Currently, there are 29 GPS satellites distributed in 6 orbits of 55◦ inclination to the equator. After many ups and downs due the collapse of former Soviet Union the current GLONASS constellation consists of 11 active satellites which are distributed in three orbit planes of 64.8◦ to the equator. It is expected the GLONASS will resume its full operational capacity (FOC) by 2013 with a constellation of total 24 nominal satellites. Research reveals even with current incomplete GLONASS satellite constellation the addition of these satellites into the combined positioning computation the positioning accuracy in the ver- tical direction (the worst coordinate of a 3D coordinate triplet) can be significantly improved. A combined GPS and GLONASS positioning solution with a mean SD of 4 mm is possible in real time. This character will greatly benefit GNSS positioning in difficult areas such as city canyon, or other partially obstructed observation environments, for instance, GPS for monitoring a suspen- sion bridge where dense cables, supporting towers or even the passing high-rise vehicle can pose problems for obtaining high quality coordinate time series. The launch of the first Galileo satellite, a future European navigation system, was made on 28 December 2005 and it is expected this new navigation system, which will consist of 30 satellites orbiting in three orbital planes of 56◦ inclination to the equator, will be FOC in 2013, hopefully! These three navigation systems will be interoperable. China is also expending its regional navigation system Beidou (or Big Dipper) into a global coverage called Compass. Compass system consists of 30 MEO (Medium Earth Orbit) satellites in three orbits as other navigation systems but includes 5 extra GEO (Geostationary Earth Orbit) satellites as an augmentation system. It is anticipated that Compass will be full operational in the middle of 2010s. With the completion of these navigation systems, the total observable satellites for navigation and positioning can reach more than 118. This will make the GNSS based positioning more robust, reliable and accurate in most outdoor activities. Global Positioning System (GPS) has been used for structural deformation monitoring of civil engineering infrastructures for more than a decade now. It has gradually been accepted by the civil engineering community as a viable tool for global structural health monitoring (GSHM) of large civil structures with many unsolved issues such as what are the best achievable positioning accuracy, the fastest sampling rate, and the cheapest yet accurate GPS hardware and software. The author of this paper attempts to answer these questions and also gives a comprehensive review of GPS for structural health monitoring in the last decade.

414 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Investigation on the severe corroded steel girder bridge, Hakkeibashi-Bridge

H. Furuta Department of Informatics, Kansai University, Osaka, Japan M. Kawatani Department of Civil Engineering, Kobe University, Kobe, Japan T.Yamaguchi & I.H. Kim Department of Civil Engineering, Osaka City University, Osaka, Japan M. Soma Department of Road and Bridge, Aomori Prefectural Government, Aomori, Japan

ABSTRACT: Hakkeibashi-Bridge consists of 4 simple supported girder bridges. Three of them are steel bridges which has 7 steel I girders and RC slab, and the rest one is RC slab bridge. This bridge has been erected atAomori prefecture about 50 years ago and removed due to severe corroded damages and deterioration of the slab in 2007. The objectives of this study are to investigate the corrosion of the steel girders in detail and to make the residual load carrying capacity of the girder clear by using some parts of girders which were taken from Hakkeibashi-Bridge.

1 INTRODUCTION

Hakkeibasi-Bridge has been constructed about 50 years ago at Gonohe, Aomori Prefecture. Aomori is located at the north side of Japan, so there is much snow in winter season. This bridge consists of 4 simple supported bridges, three of them are simple supported steel girder bridges with RC slab and the rest one is a PC girder bridge. Each span length is about 9 m. Severe corrosion at the girder ends of the steel girders and deterioration of the RC slab have been observed at the detail inspection. As a result, it is determined to demolish this bridge and to construct a new bridge. In this research, many detail investigations have been carried out in order to make it clear the residual loading capacity, deterioration process/corrosion process by using some portions of the girders which are removed from Hakkeibashi-Bridge. First of all, the material test for the test specimens obtained from the corroded portion of Hakkeibashi-Bridge, has been executed in order to investigate its deterioration due to corrosion. Secondary, the thickness of the steel plate has been measured in precious and the state of the paint film on the girders has been observed. Based on these results, FE analysis of the steel girder has been carried out in order to evaluate the residual load carrying capacity after severe corrosion.

2 CONCLUDING REMARKS

Obtained concluding remarks are as follows; 1) Maximum decrease of the thickness of the girder for about 50 years is about 9 mm under the condition that the girders have been repainted twice during 50 years. 415 2) It is considered that the causes of severe corrosion are location of the bridge, that is in north side of Japan, lack of water-proof layer between the pavement and the RC slab, and damages of the RC slab. 3) In case of Hakkeibashi-Bridge, the load carrying capacity subjected to bending decreased due to less thickness of the flange plate by corrosion. But the carrying capacity subjected to shear force is not varied because the web panel of the girder has less corrosion.

416 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

System of partial safety factors in reliability-based bridge assessment

I. Paik Kyungwon University, Kyunggi-do, Korea D. Kim Namdo University, Chollanam-do, Korea S. Shin Inha University, Incheon, Korea

Research work for developing a new reliability based assessment code for concrete bridges has been being carried out in Korea. An important issue in the reliability based assessment code is to define a system of the partial safety factors for resistances and load effects. The paper investigates the definitions and effects of those factors available in various codes of different countries and proposes a reasonable format for Korean assessment code. The factors are defined according to the target reliability index in the current context of the study. A sample reliability analysis is presented to evaluate the effects of those partial factors and is going to be further extended to determine partial safety factors of both sides of resistance and load in accordance with the target reliability index for assessment. Also, proper sets of partial resistance factor system are further studied to be proposed with both material and member force type resistance factors for determining the assessment resistance. A new format for determining the assessment resistance R is proposed as follows:

The assessment resistance R can be determined from the factored resistance φRdRn(φmXm), multiplied by the condition factor φcond, and the resistance adjustment factors U. The φm and φRd are respectively the partial safety factor for material and member strength in accordance with the bridge design code.

ACKNOWLEDGEMENTS

This work is a part of a research project supported by Korea Ministry of Construction & Trans- portation (MOCT) through Infra-Structures Assessment Research Center (ISARC). The authors wish to express their gratitude for the financial support.

417 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Bridge management system for national highway network in Korea

H.Y. Kim Korea Institute of Construction Technology, Goyang, Gyeonggi-Do, Republic of Korea

As of December 2007, 23,805 bridges are operated on the highway network systems in Korea. The Korean Ministry of Land, Transport and Maritime Affairs (former Ministry of Construction and Transportation) is responsible for the management of approximately 4,800 bridges on the national highway network. Currently, the management of the bridges on the national highway network is carried out through 18 District Offices of the Ministry. In 1994, the Ministry conducted safety inspection for the entire bridges on the national high- way network. The inspection reports indicated that over 300 bridges need immediate repairs and strengthening. At that time, the Ministry did not have any analytical tool to make more informed decision making for the bridge management. To meet the Ministry’s needs, a series of research projects to develop a network level bridge management system had been conducted by the Korea Institute of Construction Technology. This system was named as KOBMS. The major tasks in the development stage of KOBMS were the construction of database, develop- ment of BMS software, and construction of computer network. The components of BMS software are the inspection module, priority ranking module, load rating module, life cycle cost analysis module, and reporting module. The database includes the inventory data, structural data, drawings, scanned photographs, and inspection and maintenance records. Each bridge has 230 common data fields but the number of data fields depends on the type of superstructure. Priority ranking modules were developed and implemented into KOBMS. The priority ranking modules are key modules of KOBMS. For the entire bridges managed by the Ministry, bridge projects with high priority are assigned by this module. The bridge improvement projects consid- ered in the system are the replacement and rehabilitation. In the course of development of KOBMS, the most difficult task was to obtain the load rating information for the entire bridge inventory. Innovative feature of KOBMS is load rating capability using the structural data. In 1999, to pro- vide the load rating information for entire bridge inventory, a simplified load rating method was developed and implemented into KOBMS. The concept of influence line was utilized to compute the load effect due to a live load. Using this module, approximately 2,300 bridges were rated. The load rating information has effectively been used for the management of bridges on the national highway network. In 1999, to compare the alternatives of actions in the decision making process for the bridge management, a life cycle cost analysis module was developed and implemented into KOBMS. The life cycle cost analysis module is composed of the preprocessor, agency cost model, user cost model, deterioration model, and optimization module. A network level bridge management system was successfully developed and implemented for the management of bridges on the national highway network in Korea. Based on the priority ranking lists provided by KOBMS, over 1,000 bridges on the national highway network have been replaced or rehabilitated since 1995. The objective of this paper is to summarize the history of KOBMS exclusively developed for the management of bridges on the national highway network and to provide a general description of the system.

418 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Autonomous bridge inspection and monitoring based on the robotic systems

J.S. Lee Hanyang University, Ansan, Korea Sungkon Kim Seoul National University of Technology, Seoul, Korea I. Hwang Hanyang University, Ansan, Korea J.F. Choo Seoul National University, Seoul, Korea

It has been seen that bridge structures occupy a large proportion of the transportation network and, since bridges constitute the very vulnerable part of civil infrastructures affecting directly public transportation and safety, attention has been focused on the development of a total bridge man- agement system since early of 1990s in Korea. A Bridge Management System (BMS) attempts to include inspection, evaluation, estimation and rehabilitation of bridges in a systematized orga- nization, even which integrates structural health monitoring systems installed in bridges. A core part of technology to constituent a successful BMS system is robustness of the data processing from field inspection and bridge itself. Recent remarkable change to this area would be attempts to apply emerging technologies such as Information Technology (IT) and Robot Technology (RT) for upgrading out inspection technologies and practices. This paper addresses recent research efforts to develop multi-disciplinary robot devices for bridge inspection and monitoring domain. BIRDI (Bridge Inspection Research Development Interface) which is a research group supported by national funds are also introduced.

419 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Development of inspection robot to PSC box bridge using digital image processing

Jongkwon Kim Dongil Fatec Corporation, Korea Byeongju Lee Expressway & Transportation Technology Institute, Korea Dongjin Park Robot & Design Corporation, Korea Jaein Shin & Changho Park Expressway & Transportation Technology Institute, Korea

The purpose of this study is to develop a new inspection robot system to overcome many difficulties and to provide automated works for the inspection of PSC Box bridges. Digital image processing technology and Robotics are mainly involved in this system. The studies performed for developing the system can be categorized into three parts. 1)A study on development of digital image processing system. This image processing system is developed to be able to detect main cracks in the inside surface of PSC Box bridge by automatically putting all photographs obtained by cameras. 2) Development of remote controlled robot system with caterpillars and the radio communication. The caterpillars system carrying the cameras is manufactured to make the images of the inside surface of PSC Box captured about all position of bridges. The photographs can be remotely obtained by the controlled robot. 3) A study on construction of databases of the digital images. The pictures are stored in a main computer and used in order to check the changes of surface crack propagations periodically. The system which is expected to make the work of the bridge inspections simple and effective comprises of several modules which include a digital camera sun it, a remote controlled robot, a carrying caterpillar system and a radio communication, and a global positioning system.

420 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Quantification models of bridge condition and performance

Kwang-Joo Lee, Seong-Hyun Park & Jung-Sik Kong Civil, Environment and Architecture Engineering, Korea University, Seoul, Korea Kyung-Hoon Park Hybrid Structure Research Division, Korea Institute of Construction Technology, Korea Chang-Ho Park Korea Expressway & Transportation Technology Institute, Technical Support Center, Korea

In Korea, the number of bridges at the national highway built during 1997 to 2007 is 4,400. That is about 73% of the whole inventory of highway bridges. The use of bridge management systems (BMSs) has increased to establish a systematic management process and to perform efficient maintenance interventions. To predict the condition state of bridges, condition profile models based on experts’opinions have been suggested and used broadly because of the lack of quantitative inspection and maintenance records in the past. The weakness of these conventional systems is that experts’ opinions are very subjective. Moreover, the procedure to compute the condition profile of bridges has not been standardized. In this study, the present status of bridges in Korea has been investigated and bridge condition profiles have been evaluated. Database in HBMS has been used for this analysis even though the data has been accumulating for only 7 years. It cannot be said the reliability of obtained regression functions is high because of the quality of data, however, it can be improved as the period of data collection increases. The purpose of this study is to provide a method to produce bridge condition profiles even though we do not have enough data at this moment. Simple regression analysis and more complicated linear and nonlinear multiple regression analyses have been performed to obtain condition profiles of rigid-frame bridges at Korean Highway (Fig 1.). Especially, multiple nonlinear regression analysis gives considerable results and we can see that it could be an important technique for bridge assessment.

Figure 1. Regression surface of RA slab.

421 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Structural behaviors of Seohae cable-stayed bridge affected by temperature

Sang Gyu Kang & Jae Bong Kwon Seohae Grand Bridge Maintenance Office, Korea Expressway Corporation, Korea Il Keun Lee Expressway & Transportation Research Institute, Korea Expressway Corporation, Korea Gyu Hak Lee Seohae Grand Bridge Maintenance Office, Korea Expressway Corporation, Korea

Seohae Grand Bridge, as a Cable-stayed bridge, completed in December 2000 is the main route of Seohae expressway, a main artery of Korea linking north and south. Its ADT(Average Daily Traffic) is about 70,000 vpd in weekdays and 90,000 vpd on weekends. The purposes of its self-measuring and monitoring system are to make sure the serviceability and safety, to offer data for condition estimating and maintenance of the bridge, to help constructions and designs of other similar bridges and to save costs with efficient maintenance. This system was started in 2001 as a leading group in the second generation monitoring system, got through a stabilization stage and now operates in systematic ways. Offering information for designs, constructions and maintenances of recent cable-stayed bridges planned and constructed in South Korea, this paper includes the research on structural behaviors of cable-stayed bridge and variations of natural frequencies according to temperature, a representative of ambient factors, during 5 years of monitoring since 2001. Recently, many researches on finding the deficiency of structure through the real time mea- surement for prevention from accidents have been performed in many countries. However, the deficiencies that can be estimated are few in substance and the effects of temperature are con- siderable. The data obtained from SHMS during 5 years show that the structural behaviors of Cable-stayed bridge are dominantly affected by temperature, and temperature has an effect on the variations of natural frequencies. Therefore, for reasonable assessment of structural health through using the responses of structure, researches on the quantitative evaluation of changes of structural behaviors due to temperature and the rational exclusion of its effect must be performed in advance.

422 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Influences of diffusion coefficient and verification of validity on prediction of chloride induced deterioration of concrete bridges

H. Tsuruta & H. Furuta Kansai University, Suita, Osaka, Japan I. Iwaki Nihon University, Koriyama, Fukushima, Japan A. Kamiharako Hirosaki University, Hirosaki, Aomori, Japan M. Soma Aomori Prefecture, Aomori, Japan M. Suzuki Tohoku University, Sendai, Miyagi, Japan

ABSTRACT: It is important to set the parameter appropriately for prediction of chloride induced deterioration of concrete bridge. In this paper, authors researched the relationships between the diffusion coefficient of concrete and the distance from coast to concrete bridge from many liter- atures. The prediction of chloride induced deterioration of concrete bridges was carried out with the solution of the diffusion equation and Monte Carlo method to make clear the reason for setting parameters’ value. In the prediction, the chloride ion content on the surface concrete, apparent diffusion coefficient of chloride ions, cover depth of concrete and corrosion rate of steel reinforcement etc were used as parameters and were considered with probability theory.As a result, it was found that the influences of diffusion coefficient on prediction of deterioration were different by location of the bridge, side of the Sea of Japan or side of the Pacific Ocean. And the validity on prediction of deterioration was tried to verify by getting repairing information of the actual PC bridge in northern Japan. This bridge was 32 m long and had been used for 30 years. And the grade of chloride induced deterioration was so severe partially. So the diffusion coefficient of chloride ion was calculated by using actual data from concrete core specimens which extracted from actual bridge girder, the validity on our prediction method of deterioration was verified by comparing the calculated result with the state of deterioration in actual bridge. Then it was confirmed. In this paper, the following results were obtained.

(1) There are big differences depend on the quality and the number of data in the result of deterioration prediction. It is necessary to collect and sort many data in order to improve them. (2) The distribution of data is different by the difference of area and distance from coast. It is confirmed that the difference of the distribution of data occurs the big differences for result of deterioration prediction. (3) At the result of deterioration prediction using data from database and limited real data, it is confirmed that prediction by using database enables to good prediction approximately. (4) In this paper, because the standard deviation of chloride ion content on the concrete surface was too big and the result varied widely, the supposed value was used for analysis. It is necessary to reconsider about setting and estimation of standard deviation hereafter.

423

Life cycle costing

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Life time assessment of steel bridges via monitoring and testing

U. Peil, M. Frenz & I. Schendel Institute of Steel Structures, Technical University Braunschweig, Germany

The prediction of a realistic lifetime and the prolongation of the service life of a structure are important factors to reduce costs. The commonly used theoretical predictions of fatigue in steel- structures have often not proven to be accurate. The commonly used prediction model consists of a load model, a system-transfer model and a damage model. The results of these sequentially coupled models are usually unreliable, especially as the influence of the uncertain load and damage models controls the reliability of the result. A method based on monitoring strategies which avoids these problems is presented in this paper. To avoid the errors of the load and the system transfer model the random strains at the critical points are monitored. Using a multi step Markov-process, including the past and the expected trends of the future traffic, synthetic time series of the local strains are generated which includes the real statistics of the process and cluster effects induced by truck convoys. To avoid the damage model, an artificial generated time history is used as an input for the digitally controlled test rig, in which a specimen of the actual constructional detail, or even a part of it, is tested. In the whole concept two further main focuses can be identified. On the one hand it’s the growth of cracks which was already considered in the probabilistic calculation of the critical points, on the other hand it’s the crack growth rate and its prediction. In least cases the predicted first crack causes the end of the utilization of the structure. In fact crack growth which leads to system redistributions can occur. The stresses in other parts of the structure are increasing. Following the former described method, the crack growth rate must be known exactly in order to generate appropriate artificial time series for the next critical detail in the structure. Therefore specimens are used to determine the crack growth rate, which is more accurate than calculated results. Following this concept, the knowledge of the critical details in the structure is of particular importance. In old structures the choice of the few details at which the monitoring is necessary can be done in a deterministic way. In contrast, new structures normally cover a broad range of critical details with an equally distributed failure level. Monitoring of all these points is far too expensive. Semi-automatic software on the basis of an FE analysis is presented here to determine the critical details in a probabilistic way.The damage computation takes place thereby on the basis of the stretch calculated at the tip of a crack or flaw.

427 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Life cycle cost evaluation of neutralized reinforced concrete bridges subjected to earthquake

Y.C. Sung, C.K. Su, C.C. Hsu & M.C. Lai Department of Civil Engineering, National Taipei University of Technology, Taiwan K.Y. Liu National Center for Research on Earthquake Engineering, Taiwan K.C. Chang Department of Civil Engineering, National Taiwan University, Taiwan

For the Reinforced Concrete (RC) bridges, the deterioration of surface concrete caused by neu- tralization often leads to corrosion of the steel reinforcement. As neutralization progresses, the corrosion could become serious enough to endanger the structural performance of the bridges. Our previous study has established some essential mathematical expressions and important parameters to predict the neutralization effect of existing RC bridges inTaiwan, such as the diffusive coefficients of neutralized concrete, the corrosive speed and corroded depth of the reinforcements, time for initial corrosion of the reinforcements, and time for cracking of cover concrete were be able to be evaluated. As a consequence, the performance degradation of the structure can be determined quantitatively. This paper will further make the study on the life cycle cost evaluation of the neutralized bridges subjected to earthquake. The results obtained will benefit the proposing of an optimum maintenance plan for the bridges.

Keywords: Neutralization, Push-over analysis, Bridge management.

428 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Degradation, repair methods and real service life of soil steel composite bridges in Sweden

H-Å. Mattsson & H. Sundquist Department of Civil and Architectural Engineering, Royal Institute of Technology (KTH), Stockholm, Sweden

The total bridge stock owned by the Swedish Road Administration (SRA) consisted, in the autumn 2006, of 15 300 bridges with an area of 4 460 000 m2 of which 2 400 are culverts with a total bridge area of 200 000 m2. Of these culverts about 2 270 are made of corrugated steel. Culverts of corrugated steel were introduced in Sweden in the mid 1950s. A culvert is a low cost bridge which is, most of the times, quick and easy to construct. A typical steel culvert in Sweden has a span of 2–5 meter and a bridge area of 40–120 m2. It has been noted that the most severe damages of the steel culverts, caused by corrosion, were to be found in water environment. Most corrosion was found inside the steel culvert close to the waterline. It was also noted that more corrosion occurred at the end of the steel culvert compared to the middle. The most common method to repair and strengthen culverts of steel in Sweden is to apply shotcrete inside the culvert. Steel culverts constructed during the 1960s consisted typically of 3–5 mm thick steel with a thin layer of galvanized Zink. The expected life span was not an issue at that time. During the 1970s and 1980s it becomes clearer, because of corrosion, that the life span of a SSC bridge in water was limited. In Swedish regulations three different technical life spans for a bridge is described: 40 years, 80 years and 120 years, SRA (1994). Technical life span is the time the bridge should meet its required function, including “normal maintenance”. SRA’s regulations evolve over time regarding rust protection of SSC bridges in water. Findings and results from research and experiences are incorporated when SRA update their regulations. Survival analysis has been used for a long time in areas like medical research and health sciences, and can also be used for a population of bridges. The time elapsing between enrolment in the study and the event (demolition) is referred to as the bridge’s survival time. The statistical treatment of survival times is known as survival analysis. From a set of observed survival times and censored times from a sample of individual bridges one can estimate the proportion of the population of such bridges which would survive a given length in time. One common method to use is the Kaplan-Meier procedure. The survival analysis is based on the SSC bridge population in water, 1 833 still in use of which 224 have been major repaired with shotcrete and 98 have been demolished (dead). Major repaired culverts have been treated as demolished. Minor repair is not calculated since the SSC bridge is “still alive”. The median survival time for SSC bridges in water can’t be calculated since the survival curve doesn’t drop below 0.5. The cumulative survival proportion of 0.56 at 50 years is close to the median of 0.5 at 50 years for TLK 40. The cumulative survival proportion is 0.95 at 27 years which is some 30% below the minimum of 0.95 at 40 years for TLK 40. The major reason for this is that SSC bridges constructed during the 1960s and 1970s didn’t have enough protection against corrosion. The real life span for modern SSC bridges will be known about 50 to 100 years from now. A well structured BMS enables following-up the advantages and disadvantages of different structural solutions in ordinary bridges and also the use of survival analysis to find out the real life span and LCC cost including investment and repair for bridges.

429 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Study on function extension of an existing PC rigid frame bridge during its life cycle

Xiao-Xiang Li, Xue-Fei Shi, Xin Ruan & Tian-Yi Ying Department of Bridge Engineering, Tongji University, Shanghai, China

On the background of an existing expressway PC rigid-frame bridge widening and connecting project, integral and local structural performances in both old and new bridges were analyzed from perspectives of many factors such as connecting configuration design, prestressing design of new bridge, concrete shrinkage and creep characteristics in new bridge, and connecting time of the new and old bridges. Eventually, some countermeasures to control integral and local stress state in the new structure were put foreword. Pole system finite element model was established to simulate the process of structural performance change before and after connecting the new and old bridges, and correlative influence factors were analyzed quantitatively. Local spatial stress distributions at connecting belt under loads like concrete shrinkage effect, manifold wheel loads and new bridge’s support settlements were studied through the 3-D solid finite element model. The analysis results show that stress redistribution between new and old bridges occurs in some extent due to different shrinkage and creep characteristics. The trend and degree of redistribution are decided by connecting time, properties of concrete materials and prestressing design of new bridge. Meanwhile, 2.7 MPa normal stress difference appears among different webs of box girders in both new and old bridges. The longitudinal shear effect at connecting belt due to the difference of shrinkage deformation is also significant. Therefore, new box girder made of new concrete material with low shrinkage and creep is recommended. Combined with the construction schedule, six months to one year after the new bridge built to connect the two bridges is advised. Optimization on the prestressing design of new bridge is also needed. Results also show that the lateral stress at connecting belt is very sensitive to the wheel load and new bridge support settlements. Stress in the lateral direction subject to wheel load reaches its maximum at 2.6 MPa of the connecting area near side pier, while maximum lateral tension stress in the connecting belt attains 4.5 MPa caused by pier settlements. Consequently, the new bridge support settlements must be strictly controlled as well as transversal reinforcing steel bars must be rationally distributed at connecting flange according to the force behavior of the corresponding position, to guarantee the security of contact surface. The results are applied to guide the design of the bridge widening and connecting, and well effect is obtained. The research achievements also provide some references on the optimization of design and construction scheme during the process of bridge widening in the future.

430 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Residual life assessment of steel girder bridges

Rama Kant Gupta Ircon International Limited, District Center, Saket, New Delhi, India

ABSTRACT: In the present age of resource constraint, ‘residual life assessment’not only ensures best possible utilization of the assets, but also ensures avoidance of traffic dislocation which oth- erwise is likely to arise on account of decision being taken about rebuilding/regirdering of bridges, ofcourse, on adhoc basis not supported on solid technical ground. Further more, in triangulated girder steel bridges, different members face different fatigue even after passing of the same load. If the situation permits to replace more fatigued members, then it is possible to add appreciable life to the bridge at bare minimum cost. While working as Executive Director, Bridge & Structures at Research Designs & Standard Organization of Indian Railways, author of this paper received two proposals from the zonal railways wherein the concerned bridges were sanctioned for regirdering on account of completing its codal life, but, found to be otherwise in satisfactory condition. Chief Bridge Engineer of the concerned zonal railways wanted to explore the possibility of residual life, if any, of those bridges so that incurrence of expenditure, if possible, can be postponed. This was so since, even after completion of the codal life, both the bridges were seems to be in good condition. There were two options in determining the residual life. One was to rely on the standard S-N curve and based on the past traffic record, first assess the fatigue already built in and then after, assess the residual life. Another option was to first draw S-N curve from the sample taken from the bridge components where fatigue has already crept in, by testing of the samples taken from different bridge components for further cycles of failure at different loads and then after, assess the residual life. To have more accurate result, 2nd option was tried, since in first option, it was difficult to collect the past traffic records passed over the bridge. To assess the residual life, field instrumentation of the various bridge members were done to ascertain number of cycles the various bridge components are facing per day on the present day traffic pattern. Samples were also taken for ascertaining its metallurgical composition and for fatigue testing in order to plot the S-N curve. After observing the load cycles based on field instrumentation and experimental work, it was found that minimum residual service life considering the most critical bridge component is 32 years for the bridge already completed 94 years of life and 40 years for the another bridge, which has completed 78 years of life. Accordingly, re-girdering work of those two bridges was postponed. The aforesaid study raised the confidence level of the field engineers by knowing the residual life. Simultaneously, traffic dislocation problem was avoided. Over and above, unwarranted expenditure to the tune of INR 400 millions was deferred.

431 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Application of a new metal spraying system for steel bridges Part4. Reference product service life prediction for the system

T. Kondo Institute of Technologists, Saitama, Japan S. Okuno Nitto Engineering Co. Ltd., Osaka, Japan A. Yamazaki Nippon Paint Co. Ltd., Osaka, Japan H. Matsuno Dai-Nippon Toryo Co. Ltd., Osaka, Japan

Japan often experiences high humidity and hot temperatures, especially during the summer. Therefore, metallic corrosion prevention techniques are an important requirement for the steel construction projects. With this in mind, a new metal spraying system was developed. An outline of this recently developed metal spraying system and its performance on zinc- aluminum complex sprayed steel bridges was previously reported in IABMAS ’04. In that paper, the efficacy of the system in corrosion prevention of steel bridges was discussed. Later the corrosion prevention mechanism based on analysis results from actual metal sprayed steel structures was discussed in IABAMAS ’06. The durability of the sprayed film based on its corrosion prevention mechanism has also been discussed. The newly developed zinc-aluminum sprayed film retained corrosion inhibition effects over an extended period. In this paper, the referential service life of the metal spraying system is discussed, based upon investigations of actual conditions for steel structures with metal sprayed film. Since 1990, the metal spraying system has been applied to many steel structures in Japan. An outside investigation of selected 9 structures was been conducted in August 2007. As a result of the investigation, none of the structures investigated demonstrated any need for immediate repairs. However, sections of some structures had slightly corroded. There are difficulties in applying the system, for example, narrow parts or ones exposed to humidity. Further, 3 structures selected from the above 9 were conducted reference product service life prediction based upon the investigation results. Consequently, the life of this product is related to the location of the structure. If investigations will be repeated after the completion of construction, it will then be possible to estimate more exacting product service life numbers. However, it is understood that the above results are enable sufficient to estimate LCC for the present time.

432 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Resource allocation for seismic retrofit of highway network

U.J. Na & M. Shinozuka University of California, Irvine, CA, USA P. Franchetti, E. Da Lozzo & C. Modena Universita degli Studi di Padova, Italy

Rapid increase in socio-economic activities in many urban area was usually accompanied by an interactive development of highway transportation network in the form of new construction of beltways, radial arteries, and other improvements to accommodate attendant increase in traffic flow particularly arising from sprawling expansion of suburban communities. As the expansion process further proceeds and matures, however, situations become more common in which the public highway expenditure must be more targeted for not only routine maintenance but also for rehabilitation of the existing functionally or physically aging core of highway network from which the expansion originated. This paper demonstrates a preliminary study that would lead to an optimal financial resource allocation for rehabilitation of a regional highway or road network. The paper idealizes the network as a topological system and considers the effect of rehabilitation. The effect is measured in terms of enhanced socio-economic serviceability resulting from the increased traffic flow capacity of the network under operational conditions or from reduced functional degradation of the network under disaster conditions. Either can be estimated depending on the type and extent of rehabilitation, which can take different forms such as pavement improvement, adding more lanes, or even improved road signs, and seismic retrofit of bridges if located in a seismically active region. In this paper, highway networks, serving urban areas subjected to high seismic hazard (specified by a probabilistic hazard curve at each site over the service area) are analyzed so as to identify a specific set of constituent bridges to be seismically rehabilitated (typically using steel jacketing of bridge columns) under a given budget to maximize the benefit. The analysis uses Caltrans’ freeway system serving Los Angels and Orange County as testbed. The study involving Caltrans’ system represents a sequel to a keynote paper presented by Shinozuka at IAMBAS ’06 in Port, Portugal where the analysis of benefit-cost ratio was carried out from a view point of Caltrans as stake holder. Thus, major benefit of rehabilitation accrues from cost avoided by seismic retrofit that reduces the sum of the repair and societal cost, the latter associated with drivers’ delay and opportunity loss. This paper presents the seismic retrofit priority of bridges based on the transportation network analysis. In this study, drivers delay which happens immediately after earthquake due to transportation network disruption is used to determine the prioritization.

433 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Probabilistic cost model for bridge integrated project delivery and management

M.G. Huang, B.G. Kim & S.H. Lee School of Civil and Environmental Engineering, Yonsei University, Seoul, Korea Y.H. Park Korea Highway Corporation, Dongtan-myeon, Hwasung-si, Korea

There is an increasing need for effective tools for the interoperable and integrated management of life cycle cost data of civil structures. Asset management and life cycle cost analysis of civil infrastructure have been widely accepted as useful and powerful tools for the decision making of civil infrastructure. Long Term Bridge Performance (LTBP) program, which will instrument, mon- itor, and evaluate a large number of bridges in USA in order to capture performance over 20-year period of time, has initialized recently. Both concepts and the program require for comprehensively collections and integrated management of life cycle cost data and information. Meanwhile, the life cycle cost management should be interoperable in order to reduce the cost of rework due to inadequate interoperability throughout bridge life cycle. Regarding the cost of inad- equate interoperability, US National Institute of Standards and Technology (NIST) (Gallaher et al. 2004) reported that $15.8 billion in annual interoperability costs were conservatively quantified for the capital facilities industry in 2002. Lee & Jeong (2006) also suggested to using standardized information model based on open standards for maximizing the reusability of data and information. In this paper, bridge cost model, which includes an element cost model and a common data structure for life cycle cost data management of highway bridges is proposed. The element cost model is based on bridge components and life cycle activities, thus it is capable of integrating with 3D configuration model. The common data structure is ISO/STEP standard based so that the cost data can be shared and transferred among the shareholder for the integrated bridge project delivery. By using these two models, the life cycle cost data could be systematically collected, stored, and analysis for optimal funding and life cycle decision making. Cost of reworks due to inadequate interoperability can be reduced. The application to Seohae Grand Bridge, which is currently the largest cable stayed bridge in Korea, illustrated the applicability of the proposed method in estimating bridge life cycle cost in a probabilistic way. From the case study of the steel plate girder bridge, it is found that the cost based on probabilistic theory is much larger than that of deterministic theory. This is because the probabilistic approach is bridge structure safety based. It is a trade-off problem between safety and cost. Future studies are required on such areas as cost variant factors, sensitivity analysis, web-based service environment etc.

REFERENCES

Gallaher, M.P., O’Connor, A.C., Dettbarn, J.L. & Gilday, L.T. 2004. Cost Analysis of Inadequate Interoperability in the U.S. Capital Facilities Industry. NIST. USA. Lee, S.-H. & Jeong, Y.-S. 2006. A system integration framework through development of ISO 10303-based product model for steel bridges. Automation in Construction 15: 212–228. Lee, S.-H., Kim, B.-G. & Jeong, Y.-S. 2005. A new strategy for the information management of bridge maintenance. Proceeding of 9th International Conference on Inspection, Appraisal, Repairs & Maintenance of Structures: 305–313, Changsha, China.

434 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Reliability analysis for bridge piles

A.S. Nowak, M. Kozikowski & T. Lutomirski University of Nebraska, Lincoln, Nebraska, USA J. Larsen Nebraska Department of Roads, Lincoln, Nebraska, USA

The new AASHTO LRFD code (2007) provisions for design of bridge superstructures were devel- oped based on the limit state philosophy. Load and resistance factors were determined by following the reliability-based calibration procedure. However, the resistance factors for piles were based mostly on engineering judgment and there is a need for verification of the corresponding design provisions. Therefore, the objectives of the present study include review of the available pile capac- ity analysis methods, development of an accurate and user-friendly formula for pile capacity, and calculation of the resistance factor for piles. A total of 48 samples were collected from the Nebraska Department of Roads (NDOR) files. For all of the piles considered, the capacity was determined using the following four methods: Pile Driv- ing Analyzer (PDA), CAPWAP, Nebraska Modified ENR Formula, Wave Equation, GRLWEAP, and AASHTO Gates Formula. The CAPWAP method is used in NDOR practice to predict the ultimate pile capacity. To compare with the other methods, the CAPWAP ultimate pile capacities were divided by the capacities calculated using other methods. The resulting ratios were considered as random variables, and for a more efficient comparison, their cumulative distribution functions (CDF) were plotted on the normal probability paper. The new formula has been developed by trial and error. The best fit to CAPWAP is provided by the CDF with a minimum degree of variation. Conservatively, the best fit is required for the lower tail of the CDF (below 0 on the vertical axis). The proposed formula has been calibrated so that the standard deviation for the CDF represented by its lower tail is minimized, and the ratio of CAPWAP and capacities calculated using the proposed formula are close to 1. The analysis has resulted in the following expression:

where P = the nominal pile capacity (N), S = the average penetration (mm) of the pile per blow for the last ten blows for steam or Diesel hammers, and E = the energy per blow (N-mm). The developed pile capacity formula is adequate for both concrete and steel piles. The equation does not depend on the type of soil. The proposed method provides results that are in good agreement with CAPWAP. The resistance factor for pile capacity is calculated based on the assumption that the new limit state approach and the traditional allowable stress design result in the same design of a pile. The calculations were carried out for steel H piles, steel pipe piles, and concrete piles, in sand, fine grain soil and rock. Depending on pile type and soil conditions, resistance factors are from 0.70 to 0.85. The lowest values correspond to design cases dominated by dead load. Therefore, the recommended conservative value of resistance factor for all load ratios is 0.7.

435 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Influence of chloride ion diffusion coefficient on the service life of concrete structures subjected to coastal environment

J.I. Park, S.H. Bae & K.G. Yu Department of Civil Engineering, Andong National University, Andong, Korea K.M. Lee Department of Civil and Environmental Engineering, Sungkyunkwan University, Suwon, Korea H.Y. Shin & D.O. Kang Samsung Construction Co., Seoul, Korea

To predict service life of concrete structures exposed to chloride attack, surface chloride content, diffusion coefficient for chloride ion, and critical chloride content for corrosion in concrete, are considered as important factors. Of these, the chloride ion diffusion coefficient in concrete may significantly influence the service life of structures. The qualitative factors affecting the ingress of chloride ion into concrete are water-binder ratio, cement type, age, chloride ion concentration of given environment, wet and dry condition, etc. In this paper, the effects of W/B ratios and cement types on the diffusion characteristics for chloride ion in concrete were investigated through the chloride ion diffusion test. For this purpose, the diffusion characteristics in concrete with three types of cement such as Ordinary Portland Cement (OPC), Binary Blended Cement (BBC), and Ternary Blended Cement (TBC) were measured for the concrete specimens with W/B ratios of 32%, 38%, and 43%, respectively. It was observed from the test result that the resistance against chloride ion penetration increased with decreasing W/B ratio and that of TBC concrete was the greatest of the cement types used in this study. Furthermore, the service life of concrete structures was predicted by using the measured values of the chloride ion diffusion coefficient in concrete.

Figure 1. Service life of concrete structures according to cement types. 436 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Damage to structures due to increasing traffic numbers related to service life predictions

Ane de Boer & Boyke (M.H.) Djorai Ministry of Transport, Public Works and Watermanagement, Utrecht, The Netherlands

ABSTRACT: Increasing traffic numbers will have an extreme effect on the condition of structures in the infrastructure, while the space available in crowded areas leaves little room for expansion of the existing infrastructure. In order to get a view of these effects, an upper boundary approach for universal damage factors for structures has been developed to manage the maintenance or replacement of structures in the future.

1 INTRODUCTION

Many European Union reports and Dutch reports, e.g. the Mobility report of the Dutch Ministry of Transport, predict a growth of 70% in lorry traffic over the period 2000–2020. 30% growth in private vehicles is predicted over the same period. These growth predictions mean an increase in the percentage of lorries when compared to the current mix of road traffic, Lorries will then comprise 15–20% of the total traffic volume. This shift in the percentage of lorry traffic in relation to the total traffic will have consequences for traffic flow on the highways and the physical condition of the structures and the road-pavement. The influence of these changes on the costs related to infrastructure isn’t sufficiently known. In order to get a better view of these items, in particular the changes for the structures and road- pavements’ the main highway routes linking the Netherlands to the rest of Europe, will have to be investigated. In 2006 a project was started to help minimize the consequences of the changes and to respond to these changes with new or better design codes. The first phase of the project was to investigate the negative consequences for traffic flow and the condition of the roads and highway structures. This research was finished in the first quarter of 2007. The Civil Engineering Division was involved in this project in the field related to the condition of structures. In this paper the results will be summarized and the recommendations for the future will be discussed. An overview of the different types of structure, the different construction materials in the struc- tures and the changes in the design codes and traffic loads is given for the last 80 years. Table 1 gives an overview of the vehicle categories on the Dutch highways. Figure 1 illustrates the impact of the changes made since the 1960’s and shows the damage caused in steel and concrete structures due to the vehicle load categories in table 1.

Table 1. Vehiclecategories and the number of passages in Dutch highways.

Category Category Category [kN] Number [kN] Number [kN] Number

200 938,867 750 53,274 1280 117 310 591,093 950 2223 1410 78 490 486,954 1090 312 580 128,973 1220 195 Total 2,205,821

437 Figure 1. Annual damage percentages to the main load bearing system for 10 vehicle configurations.

Table 2. Damage ratio for the growth in the number of vehicles according to the current Dutch vehicles.

Damage ratio caused Damage ratio Present by increasing the number day – anno 2000 [%] of vehicles by 70% [%]

Vehicle categories Steel Concrete Steel Concrete

Category 490 kN 16.0 1.3 29.6 2.4 Category 580 kN 12.1 3.0 44.0 11.0 All categories 100 100 373 370

When the traffic volume increases in line with the EU expectations of 70% by 2020, the amount of damage due to common traffic will increase from 28.1% to 73.6% for steel structures and from 4.35 to 13.4% for concrete structures. The overall increase will be from 100 to 370%, which means that the lifetime of structures will be reduced from 100 years to 27 years.

2 CONCLUSION

The growth of traffic predicted in the DutchTraffic regulations will cause a small amount of damage to existing structures (maximum steel:−14%, concrete:−1.1%). When this growth is applied to all vehicle categories currently present on the Dutch Highways, the maximum upper boundary level damage factor will reduce the predicted lifetime of a structure to 26.8 years. The vehicle weight category 750–950 kN causes the most damage to steel structures, while the heavier vehicle weight category above 1090 kN will cause the most damage to concrete structures.

REFERENCES

Ministry of Transport. 2004. Ministry of Transport, Public Works and Water Management. Report Mobility (in Dutch). The Hague. Vrouwenvelder, A.C.W.M.,Waarts, P.Hen Wit, S. de. 2000. General probabilistic considerations for reliability and modelling of traffic loads for road bridges (in Dutch). 98-CON-R1813. TNO-Built Environment and Geosciences. Delft. CEN-EN 1993-1-9. 1993. Eurocode 3: Design of steel structures – Part 1.9: Fatigue. NEN. Delft. Dijkstra, O.D. & Bentum, C.A. van. 2006. Fatigue bridgedecks equivalent traffic loads (in Dutch), 2005-BCS- R0453. TNO-Built Environment and Geosciences. Delft. Boer, A. de, Djorai, M.H. & Noordzij, R. 2007. Effects of increase of lorry traffic loads on structures till 2020 (in Dutch). CT07-R-7144-001. Civil Engineering Division. Utrecht. 438 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Optimal design of cable-stayed bridges based on minimum life-cycle cost

Sung-Ho Han & A. H-S. Ang Department of Civil and Environmental Engineering, University of California, Irvine, CA, USA

Generally, it seems reasonable to conduct the safety examination using a reliability evaluation method that takes into consideration the effects of uncertainties in the design variables. For this, it is important to consider two types of uncertainty; namely, the aleatory and epistemic types in which the aleatory uncertainty is associated with randomness of nature and leads to a probability of failure or risk, whereas the epistemic uncertainty gives rise to a range of possible values of the calculated probability or risk. Conversely, in performing structural designs, it is reasonable from the standpoint of economic efficiency and safety to find the optimal design based on the minimum life-cycle cost. In this process, the determination of the minimum life-cycle cost must include the uncertainties in the initial cost and damage cost. For this purpose, a procedure for the optimal structural design of cable-stayed bridges is proposed based on minimum expected life-cycle cost; the procedure is illustrated with the optimal design of a cable-stayed bridge in Jindo, Korea subjected to static and seismic loads. Reliability analysis of the bridge was performed taking into account the two types of uncertainty in the capacity and loads. The capacity of the bridge is assumed to be determined by its critical members; this is tantamount to the assumption that the capacities and load effects of the structural members are highly correlated. Various designs of a bridge are considered; namely, a standard design, plus several that are weaker as well as several that are stronger than the standard design. For the different designs, the member sections are decreased or increased relative to those of the standard design. The Life-Cycle Cost (LCC) of a particular design is formulated assuming that the cost components (including the maintenance and social costs) are respectively fractions of the initial cost. Reliability of a design associated with the aleatory uncertainties is assessed for each design, and the corresponding expected LCC and safety index are evaluated. The results of the various designs provide the information, β vs E (LCC), for determining the design with the minimum expected LCC which can be presented graphically. Because of the epistemic type of uncertainty, the LCC as well as the safety index of the optimal design are random variables; the respective distributions (or histograms) are also generated, from which the various percentile values can be obtained. In risk-based engineering, it is important to distinguish the difference between two broad types of uncertainty; the aleatory type which is part of the randomness of natural phenomena whose sig- nificance can be expressed in terms of the probability of occurrence, and the epistemic type which is associated with imperfections in modeling and estimation of reality and leads to uncertainty (lack of complete confidence) in the calculated probability or risk. For practical applications, epistemic uncertainty may be limited to the imperfections in the estimation or predication of the mean (or median) value of a variable or parameter. Because of these epistemic uncertainties the calculated results, such as failure probability, safety index, and expected life-cycle cost, become random variables with respective distributions (or histograms) For decision-making purposes, the distri- butions of the calculated results allow the specifications of high percentile values of the essential design parameters (such as safety index) to insure sufficient risk averseness. For example, the 90% value, or the 75% value, of the safety index may be appropriate, leading to conservative designs (particularly important) for long span bridges.

439 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Design planning decision for deteriorating wearing surfaces based on whole-life design considering life-cycle cost

J.X. Peng & X.D. Shao Institute of Bridge Engineering, Hunan University, Changsha, China M.G. Stewart Centre for Infrastructure Performance and Reliability, University of Newcastle, Australia

ABSTRACT: Improved calculation methods for indirect maintenance cost over the service life of a bridge wearing surfaces are proposed in the present paper. Based on the existing data in a Chinese BMS, bridge condition is assessed and predicted under predefined optimal maintenance strategies, and used as boundary constraints for the proposed optimization model. A lifetime cost optimization methodology for planning the maintenance of structures that deteriorate over time is introduced and illustrated through numerical examples. The cost optimization is based on minimizing the expected total life-cycle cost while maintaining an allowable lifetime condition index. This method incorporates: a) predictive maintenance strategy; b) the effects of ageing, deterioration, and subsequent repair on wearing surface performance; c) the time value of money. The overall cost to be minimized includes the initial cost and costs of patch maintenance and performance-based repair. Selection of the number of bridge lanes (bridge width) is employed to illustrate the potential for cost saving and improving benefits of the proposed approach herein. It is concluded that when the traffic volume is constant, the design scenario that adds a temporary lane based on the 4-lane design one can decrease indirect maintenance cost effectively and keep the bridges in a higher service level.

1 INTRODUCTION

Structural safety, durability and economy are the basic design principles of bridge engineering. Con- ventional bridge design method uses quantitative design approach to purse a minimum acceptable design that meets all the performance requirements specified in the design code. Thus, this paper presents a practical and realistic life-cycle cost (LCC)-based bridge whole-life optimum design based on the previous studies [Peng and Shao 2005; Shao et al 2006]. According to the minimization of present value of expected life-cycle cost, the optimal design scenario is selected under the constraint of bridge performance. A numerical example of selection of bridge width is employed to verify the effectiveness and feasibility of the whole-life optimum design method for a bridge concrete wearing surface.

2 LCC-BASED BRIDGE WHOLE-LIFE OPTIMUM DESIGN

LCC-based bridge whole-life-optimum design needs to evaluate the sum of life-cycle cost from birth of bridge to completion of life in design stage. The basic principle of LCC analysis method is expected value of life-cycle cost is minimum under the constraints of service level (C ≥ Ctarget, Ctarget is target condition index), which can realize the tradeoff between technology and economy. In this study, the selection of number of bridge lane (bridge width) is treated as the design objective to investigate bridge whole-life optimum design problem. The optimal design planning problem deals with cost-effective allocation of design efforts so that a preferred balance among 440 lifetime performance, initial cost and maintenance cost can be obtained. The problem can be concisely stated as follows

where x1 is condition index, and is used to depict performance state of bridge wearing surface; Ctarget is predefined target condition index; E( ) is the present value of expected whole-life cost; x2 is the number of bridge lanes; ui,uj are the upper limit and lower limit of design variable x2, respectively.

3 LIFE-CYCLE PERFORMANCE ASSESSMENT BASED ON CONDITION INDEX

Statistical regression theory is successfully applied to predict bridge condition in (Jiang 2006), and polynomial deterministic condition equation for bridge concrete wearing surfaces. According to statistical data in Qingyuan, Guangdong Province, a reliability index based on least square method can be used to assess performance of bridge wearing surface is assumed as

where C0 is initial condition index, α1, α2, α3 are empirical deterioration rate parameters. All these parameters are random variables. It is assumed that while the condition index decreases to a limit condition index value of C∗(i.e. C∗ = 5.48) at time t∗ the bridge overlay needs to be patched. When the condition index decreases to the target condition index Ctarget (i.e. Ctarget = 0.0), a performance- based essential maintenance is applied. The empirical approach adopted herein should be used with caution. For example, a data fitting based on historical performance data may not adequately capture future performance as bridge inventory is highly non-stationary. According to distribution types of random variables, it is calculated that three time-controlled patch repairs are respectively applied at 8.6 yr, 17.9 yr and 41.1 yr over life of 50 yrs. A performance- based maintenance intervention is applied at 30 yr, and condition index of bridge wearing surface at 50 yr is 2.1.

4 CONCLUSIONS

This paper presents a realistic LCC methodology for whole-life optimum design of deteriorating bridge concrete wearing surfaces considering a bridge deterioration–maintenance model by means of direct and indirect maintenance cost evaluation method. A numerical example is employed to demonstrate the feasibility of the proposed optimum design approach. The following conclusions can be drawn: Firstly, the whole-life optimum design method not only incorporates parameter uncertainties and programs maintenance planning in service time, but also balances the construction cost and maintenance cost. So, this optimum design method plays a role on protecting the environment and optimizing resources, thus enabling sustainability development of bridge engineering. Secondly, according to the concept of whole-life optimum design, a new effort in selecting bridge width in this paper is performed based on the calculated results, when the traffic volume is constant, the design scenario that adds a temporary lane based on the 4-lane design scenario can decrease indirect maintenance cost effectively and keep the bridge in a higher service level. Thirdly, the improved indirect maintenance cost evaluation model is suitable for maintenance of other bridge elements. Finally, whole-life optimum design of bridges has only recently been developed by bridge researchers and is very effective for selection of bridge types and design parameters. Further efforts in selecting suitable decision objectives, deterioration models, optimization models, maintenance strategy and solution method according to specific structures and actual situations are required.

441 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Optimal seismic design of cable-stayed bridges based on LCC concept

D. Hahm & H.-M. Koh Seoul National University, Seoul, Korea W. Park Korea Bridge Design & Engineering Research Center (KBRC), Seoul, Korea K.-S. Park Dongguk University, Seoul, Korea S.-Y. Ok University of Illinois at Urbana & Champaign, USA

Various strategies for the seismic response mitigation have been extensively studied and applied to several engineering structures. Among them, cable-stayed bridges have seen recently increasing applications of control devices such as dampers. However, most of the previous optimized seismic design concepts or schemes for cable-stayed bridges were focusing essentially on the performances of the control devices rather than on their applicability. Need is thus to select a broader approach embracing both performance of the devices and cost-effectiveness in a lifetime perspective. Accord- ingly, this paper presents an optimal seismic design procedure of cable-stayed bridges adopting the Life-Cycle Cost (LCC) concept. The corresponding cost-effectiveness of the optimally-designed aseismic devices is evaluated considering various seismic characteristics such as magnitudes and frequency contents. The second Jindo Bridge located in the southwestern cost of Korea is chosen for the application and comparison is done for various aseismic devices such as elastomeric bearings, MR dampers and seismic isolation devices in order to derive optimal design alternatives. The cost-effectiveness of each control system is defined by the ratio of LCCs of the bridge structure with the control system and the bridge structure with conventional shock transmission units. The LCC function consists of the summation of initial construction cost and expected damage cost due to earthquake event. For the evaluation of the expected damage cost, the failure probability of a cable-stayed bridge system is estimated using simulation methods. Since the damage cost may be very sensitive to the surrounding conditions such as the regional economic status, the scale of damage cost is adopted as a parametric variable in the LCC evaluation of a seismically excited cable-stayed bridge. Finally, based on the LCC concept, the efficiency of the proposed optimal design method is discussed with respect to the various seismic characteristics.

442 Life-cycle structural engineering

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Structural geometry effects on the life-cycle performance of concrete bridge structures in aggressive environments

Fabio Biondini Department of Structural Engineering, Politecnico di Milano, Milan, Italy Dan M. Frangopol Department of Civil and Environmental Engineering, Advanced Research Center for Large Structural Systems (ATLSS), Lehigh University, Bethlehem, PA, USA Pier Giorgio Malerba Department of Structural Engineering, Politecnico di Milano, Milan, Italy

ABSTRACT: For concrete structures, design for durability with respect to chemical-physical damage phenomena is usually based on simple criteria associated with prescribed environmental conditions. Such criteria introduce threshold values for concrete cover, water-cement ratio, amount and type of cement, among others, to limit the effects of local damage due to carbonation of concrete and corrosion of reinforcement. However, due to the interaction between local damage and global structural behavior, the life-cycle performance of concrete structures exposed to aggressive agents may significantly depend on structural aspects associated with the geometry of the cross-sections and the reinforcement layout, as well as on the shape of the structure and the system redundancy. In this paper, structural geometry effects on the life-cycle performance of concrete structures in an aggressive environment are investigated. These effects are investigated with respect to cross- sectional shape factors related to parameters like the free perimeter exposed to aggressive agents, the number of the corners, the number and the dimension of internal cells, the location of the reinforcement bars, and other geometrical parameters. The life-cycle structural analyses required for the numerical investigation are carried out by using a general procedure proposed in previous works for concrete structures in aggressive environments (Biondini et al. 2004, 2006a, 2006b). The diffusion process is modeled by using cellular automata and the mechanical damage coupled to diffusion is evaluated by introducing proper material degra- dation laws. A nonlinear analysis is carried out at different time instants in order to evaluate the time evolution of the structural performance. The effects of the considered shape factors are investigated at the cross-sectional level with reference to the resistant bending moment. The application to a box cross-section, quite typical for bridge piers, shows that the equivalence among cross-sections having the same initial performance in the undamaged state is generally lost in time due to dam- age, and that the loss of performance strongly depends on both geometrical parameters and spatial distribution of the aggressive agent.

REFERENCES

Biondini, F., Bontempi, F., Frangopol, D.M., and Malerba, P.G., (2004). Cellular Automata Approach to Durability Analysis of Concrete Structures in Aggressive Environments. Journal of Structural Engineering, ASCE, 130(11), 1724–1737. Biondini, F.,Bontempi, F.,Frangopol, D.M., and Malerba, P.G.,(2006a). Probabilistic Service LifeAssessment and Maintenance Planning of Concrete Structures, Journal of Structural Engineering, ASCE, 132(5), 810–825. Biondini F., Frangopol D.M., Malerba P.G., (2006b). Time-variant Performance of the Certosa Cable-stayed Bridge. Structural Engineering International, 16(3), 235–244.

445 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

FRP reinforced concrete: Reliability assessment for life-cycle analysis

S.M.C. Diniz Federal University of Minas Gerais, Belo Horizonte, Brazil

In a time of greater demands posed by a deteriorated infrastructure in both developed and devel- oping countries, it is clear that more durable structures must be achieved. This is particularly true for Reinforced Concrete (RC) bridges in aggressive environments. Recently, the use of Fiber Reinforced Polymers (FRP) as internal reinforcement has gained an increased attention for its potential for improved durability and attendant reduction in the structure’s life cycle costs. This fact is easily understood since these materials are noncorrosive; additionally, FRP materials exhibit several properties, such as high tensile strength, that make them suitable for use as structural reinforcement. The use of FRP as internal reinforcement is seen as a great promise in terms of durability; however, currently a major limitation in the use of this material is the higher initial costs. It is clear that these tradeoffs between initial and future costs can be properly dealt with in a life-cycle analysis framework. Nevertheless, while the initial costs can be easily estimated, the estimation of failure costs with respect to ultimate and serviceability limit states are much more involved. In the case of ultimate limit states, a major problem is the evaluation of the corresponding probabilities of failure. In this paper, the evaluation of probabilities of failure associated with flexural ultimate limit states for FRP-RC beams is presented. The beams evaluated have been designed according to the ACI 440 (2006) design guidelines. In the case of FRP-RC structures, the components of interest are those subjected to bending. In order to achieve a ductile failure mode (i.e. steel failure), beams are designed as under-reinforced in traditional RC design. On the other hand, in FRP-RC beam design, the most desirable failure mode is concrete crushing, i.e. a brittle failure is unavoidable. As it can be seen there has been a change in the flexural design paradigm thus granting a careful consideration of its implications. The influence of concrete compressive strength, type of FRP bar, longitudinal reinforcement ratio, and load ratio in the resulting reliability levels are investigated. ACI Committee 440 (2006). Guide for the Design and Construction of Concrete Reinforced with FRP Bars (ACI 440.1R-06), American Concrete Institute.

446 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Seismic performance upgrading of existing bridge structures

G. Furlanetto, L. Ferretti Torricelli & A. Marchiondelli Bridge Design Department, SPEA Ingegneria Europea, Milan, Italy

ABSTRACT: In Italy the most part of the national highway system is committed to Autostrade company, which has to manage about 5500 km of highways, comprising a total of about 3000 bridges. Since the first highways were built in the 60’s, many of them are nowadays inadequate to meet the actual functional and safety standards anymore. In order to answer to the continuously increasing demands due to the high volume of traffic flow, and to comply with the recently issued new seismic regulations, Autostrade and SPEA Ingegneria Europea, its design company, are widely involved in planning the widening and retrofitting of the most part of the highway network. The new Italian seismic code (OPCM 3274/2003) introduced innovative design criteria based on the performance of the structures which deeply differ from the old prescriptive ones. Moreover, the new seismic hazard maps revealed that many regions, that in the past were not classified, are to be considered seismic. In this complex context, Autostrade and SPEA Ingegneria Europea chose to develop an accurate document, called “Guidelines for the evaluation of safety of existing bridges before and after widening interventions”, aimed to represent an important reference for all the designers involved in the design activities. In this document, a unitary approach on the most appropriate analysis and diagnosis methods, and on the most adequate remedial measures to be adopted in order to improve the seismic performance of existing bridges, is presented. In particular, the design of the widening portion is addressed by the concept that this kind of intervention is a part of the strengthening strategy. In fact, by accurately designing the structural members of the widening portion so to lead to uniform distribution of seismic forces, the overall performance of the construction can be improved. According with the Guidelines, seismic analysis can be developed by using a linear static analysis, applicable only for structures which satisfy proper regularity criteria, or a non linear static analysis, called pushover analysis, able to take into account the post-elastic behaviour (Pinto, 2005). On the basis of the results of the analysis, a reliable judgment about the seismic performance of existing bridges and the need of contingent retrofitting interventions can be expressed. In many cases the existing bridges revealed severe structural deficiencies which caused an inade- quate safety level. In order to adequate the structural response to the new safety requirements, proper strengthening interventions are requested (Furlanetto et al., 2007). After a brief overview on the typical retrofitting interventions, a representative case study is presented by describing the results of the structural analysis and the adopted strengthening interventions.

REFERENCES

Furlanetto, G., FerrettiTorricelli, L. & Marchiondelli, A., StructuralAssessment and Rehabilitation of Concrete Bridges, IABSE Symposium, September 19–21, 2007, Weimar, Germany. OPCM 3274/2003: “General criteria for the seismic classification of Italian territory and technical specification for structures”, and following modifications. Pinto P.E. (2005), The Eurocode 8-Part 3: The new European code for the seismic assessment of existing structures. Asian Journal of Civil Engineering, Vol. 6, No. 5. 447 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Estimation of fatigue life for suspension bridge hangers under wind action and train transit

F. Petrini, F. Giuliano & F. Bontempi University of Rome “La Sapienza”, Rome, Italy

The main causes of induced fatigue in the cables and hangers of suspension bridges are wind and traffic induced vibrations. Cause of the high flexibility and reduced weight (in relation of the whole dimension of the structure), during their life-cycle, suspension cable system of these type of bridges can experiment a great number of tension-cycles with significant amplitude. In this paper the fatigue problem regarding hangers of the bridge is treated, these members are subjected to axial fatigue effect that can be faced with simplified damage laws and fatigue curves referring to experimental data. Concerning wind fatigue damage, several authors have done studies to assess fatigue behavior of structures and cables under wind action. Cause of wind stochastic nature, the problem has to be treated in a probabilistic way. The fatigue damage is correlated with the mean wind velocity magnitude, and the last one can be viewed like a stochastic variable having an annual Weilbull probability density function. The fatigue damaged induced by traffic is different in the different members of suspension bridges, depending on the stress rate induced by the passage of the vehicles: in the members which experience low excursions of the stress rate in respect the permanent load, such as the main cables, the expected fatigue damage is low, especially in the case of long span bridges, for which the permanent load is very high. Vice versa, hangers, deck and secondary components have high change in the stress during the passage of vehicles, and so they are prone to be damaged and substituted several times during the service life of the bridge. In the present paper this problem is developed for a case study of long span suspension bridge and the life duration of different structural members is assessed, the rainflow method is assumed as the fatigue counting method, the Palmgreen-Miner law is assumed as the damage accumulation criterion, and the Eurocode 3 fatigue curve is assumed for the members’ damage calculation. The third aspect of discussion, concerning the subject of the paper, is the interaction mechanism between train and wind. First, there is an aerodynamic interaction caused from the local changes in aerodynamic forces due to the train presence on the bridge deck, which changes the local aerodynamic shape of the structure. This effect increase in importance when the geometrical dimension ratios between train and structure increase. The second interaction mechanism concern the structural dynamic behavior, it can be significant in the case of large deflection of the bridge. Third interaction mechanism concerns the fatigue damage interaction: the sum of damages due to wind and train, is not equal to the damage caused from train passage when the structure is subjected to the wind action. In the present paper, interaction mechanisms are not considered. The analyses have been developed in time domain both for wind action and train passage effects investigation. Analyses results show that the train passage influences the fatigue life more than wind action; it is also shown that the critic hanger (more damaged from fatigue point of view) is in different location along the deck for the two actions. Further analyses have to be developed to implement wind-train interaction effects.

448 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Life-cycle bridge management considering member interference

Kyung-Hoon Park & Sang-Yoon Lee Hybrid Structure Research Division, Korea Institute of Construction Technology, Korea Jung-Sik Kong Civil, Environment and Architecture Engineering, Korea University, Seoul, Korea

To predict the future condition of bridges, condition profiles based on experts’ opinions have been used widely because of the lack of quantitative inspection and maintenance records. A weakness of these conventional profiles is that those opinions are subjective. In order to provide reason- able maintenance scenarios for bridges, a method to calculate the optimum maintenance strategy through the connection of bridge performance change and cost during the lifetime was developed. Unlike traditional bridge maintenance methods aiming to find out a single maintenance scenario minimizing the life-cycle cost, this system provide a number of tradeoff maintenance scenarios (Fig. 1) between two conflicting objectives considering not only the life-cycle cost but also the life-cycle performance. The developed program through the application of genetic algorithm including the life-cycle performance and the cost analysis program can generate the optimal tradeoff maintenance scenario for each year not only in the level of bridge member, but also in the level of bridge system. Through applying the developed program to the existing bridge maintenance, it was found that the maintenance scenario with the system level can be generated through the life-cycle analysis and a reasonable annual optimal maintenance scenario can be generated by considering the subordinate relationship in accordance with the replacement between members. The proposed method can be used for provide the optimal maintenance scenario satisfying the various requirements and constraints of the bridge agency.

Figure 1. Solution set for elite group.

449 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Bridge maintenance strategy based on life-cycle cost and rebuild cost stabilization

A. Miyamoto Yamaguchi University, Ube, Japan J. Ishida New Yamaguchi Prefectural Government, Yamaguchi, Japan

ABSTRACT: The needs for efficient bridge management systems have been recognized all over the world. Although there are some commercial products on the global market, the national dif- ferences many times hinder the adoption of foreign bridge management systems. This is one of the main reasons why so many countries are developing their own systems. The local con- ditions (bridge types, macro and micro climatic peculiarities, working traditions, etc.) are best known by the local people. By trusting in local expertise also the customer needs (i.e. needs of the bridge owner) are best taken into account. However, with many new innovations, technolo- gies and methodologies, a lot of benefits can be achieved by international cooperation in this field. In Yamaguchi Prefecture, some studies have been made about planned maintenance jointly by administrative, academic and private sectors. Research and development has been carried out concerning maintenance systems for bridge maintenance planning in a joint research project of Yamaguchi University and the prefectural government. The requirements for future road mainte- nance are enhancing road functions, assuming accountability and reducing costs. A system will therefore be built in the joint research to meet the requirements and bridge maintenance plans will be developed and implemented. Many of the bridges under the management of Yamaguchi Prefecture were built in the 1950s to the 1970s. In 2025, about 63 percent of the bridges under the management of the prefecture will be 50 or more years old. The cost of the repair, strengthening and reconstruction of these aging bridges is expected to increase sharply in the coming years. Under these circumstances, many road administrators including Yamaguchi Prefecture need to manage bridges systematically. For systematic management of the bridges, it is necessary to draw up a management plan by making effective use of a wide range of information such as bridge inspection results, bridge specifications and future deterioration predictions. To do this, there is a pressing need to develop a management planning support system. In this paper, a strategy decision support system for bridge management is described based on both of the Remaining Life Cycle Cost (RLCC) and rebuild cost stabilization, as a joint research project between Yamaguchi University and Yamaguchi Prefecture. The proposed sys- tem that considered not only future maintenance but also rebuild action is applied to existing bridge data on about three and half thousand bridges that stored in J-BMS data base (J-BMS DB) for practical use. Then, it is found that the system helps efficiency bridge engineers related to bridge management understand easily what are the major processes in the deteriorating bridges in the future, how to improve their performance, and what kind of repair/strengthening works is most convenient for the deteriorating bridges. In order to make the Bridge Management Sys- tem (J-BMS) developed by the authors a more practical tool, a Bridge Management Planning Support System (J-BMS’07) that can be used for more than one bridge on the highway net- work has been developed. The system performs three tasks: (1) making it possible to draw up a management plan that takes into consideration not only repair and strengthening based on structural soundness evaluation results but also reconstruction so that the peaks of future

450 reconstruction costs can be shaved; (2) provides a reconstruction cost stabilizing function to cal- culate a realistic service life of each bridge, and (3) drawing up a management plan that satisfies the budgetary requirements for repair and strengthening for each year, taking into consideration budgetary constraints and construction requirements that can be specified by bridge administra- tors, and enabling the user to specify the year in which to perform repair or strengthening for each bridge.

451 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Damage modeling and life-cycle reliability analysis of aging bridges

Fabio Biondini Department of Structural Engineering, Politecnico di Milano, Milan, Italy Dan M. Frangopol Department of Civil and Environmental Engineering, ATLSS Research Center, Lehigh University, Bethlehem, PA, USA Elsa Garavaglia Department of Structural Engineering, Politecnico di Milano, Milan, Italy

In order to assess the life-cycle performance of a deteriorating structural system and to eventually plan a process of essential and/or preventive maintenance, some basic information on the actual development of the deterioration process, as well as on the uncertainty involved in its modeling, is clearly required. Therefore, a life-cycle reliability analysis must be based on a suitable modeling of the damage evolution and on a probabilistic analysis able to incorporate the main features of the time-variant deterioration process. A deterioration process may be formulated as a probabilistic problem where every loss of perfor- mance greater than prescribed threshold values is considered as a failure. Therefore, when a failure is reached, the system passes from the current service state to another service state characterized by a lower level of performance. On the other hand, structural performance can also be improved by maintenance and/or rehabilitation interventions and the system may move from the current service state to another service state characterized by a higher level of performance. However, in both cases the failure process may be defined as a transition process that can be formulated as the semi-Markov processes. These processes can well model the deterioration with the aim to evaluate the life-cycle reliability of the system. Based on these premises, in this paper, a general procedure for life-cycle reliability assessment and maintenance planning of deteriorating structural systems is proposed. Material damage is described by means of damage indices and its effects at the cross-sectional level are considered by using cross-sectional performance indices by taking into account both cross-sectional shape and damaging pattern. Selective maintenance scenarios are considered by applying the repair interventions only to elements heavily deteriorated and/or characterized by a low safety margin. To this aim, a suitable condition index is introduced. The effectiveness of the proposed procedure is finally shown with an example of life-cycle reliability analysis applied to a bridge structure.

REFERENCES

Biondini, F.,Bontempi, F.,Frangopol, D.M., and Malerba, P.G.2006. Probabilistic Service LifeAssessment and Maintenance Planning of Concrete Structures, Journal of Structural Engineering, ASCE, 132(5), 810–825. Biondini F., Frangopol D.M., Garavaglia E. 2006. Probabilistic Lifetime Assessment based on Limited Moni- toring, 3rd Int. Conf. on Bridge Maintenance, Safety and Management (IABMAS’06), Porto, Portugal, July 16–19, 2006, Ed. Paulo J.S. Cruz, Dan M. Frangopol & Luis C. Neves, Taylor & Francis Groups. Leiden, The Netherlands, CD-ROM, Paper_030. Biondini, F., Frangopol, D.M. 2008. Long-term Performance of Structural Systems, Structure and Infrastructure Engineering, 4(2), Special Issue, Fabio Biondini (ed.), Taylor & Francis Publisher (in press).

452 Lifetime-perspective design of Kwangyang suspension bridge with main span 1545 m

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

The planning and design of the long-span suspension bridge connecting Myodo and Gwangyang in Korea

Jae-Hong Kim Bridge Engineering Team, Daelim Industrial Co., Ltd., Seoul, Korea Myeong-Jae Lee, Sang-Hoon Shin & Sung-Bum Chun Structural Division, Yooshin Engineering Corporation, Seoul, Korea

This paper describes the features of the structural design including the investigations performed along the planning and design stages of the long-span suspension bridge connecting Myodo and Gwangyang. This bridge will be constructed in the third section of approach road works of Yeosu National Industrial Complex from 2007 to 2012. The total span length of 2260 m with a central span of 1545 m will boost the bridge at the third place among the longest bridges in the world. The bridge has the floating-type stiffening girder which has no vertical supporting points at pylon. The long central span, the 270 m high pylons and the sag ratio of 1/9 will secure 18,000TEU-class shipping serviceability, reduce risk of ship collision, diminish the size of anchorage, and reduce the construction cost. The choice for a central span longer than 1500 m was inevitable regard to the condition of ship navigation, wind climate, and construction cost, and required to improve the technical level of the structural components applied to date for existing suspension bridge systems. To realize the long-span suspension bridge, close investigation of previously applied design and construction methods and detailed examination of recent advanced design and construction methods were carried out. By introducing twin box shape for the girder, the aerodynamic stability of the whole bridge system could be guaranteed. In addition, the most optimized section type of twin box girder suppressing the vortex shedding vibration and minimizing the drag force coefficient could be found through wind tunnel tests. Additionally, the innovative structural systems are composed of high-strength main cable of 1860 MPa, and stoppers at the pylons and lock-up-type buffers at the ends of stiffening girder.

Figure 1. Aerial view of the bridge connecting Myodo and Gwangyang.

455 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Wind resistance design of Kwangyang Bridge

Soon-Duck Kwon & Seung-Ho Lee KOCED Wind Tunnel Center, Chonbuk National University, Chonju, Chonbuk, Korea Hidesaku Uejima Bridge Aerodynamics Gr., IHI Co., Shin-Nakahara, Isogo, Yokohama, Kanagawa, Japan Myeong-Jae Lee Structural Division, Yooshin Engineering Corp., Yoksam, Kangnam, Seoul, Korea

The Kwangyang Suspension Bridge is located in south seashore of Korean peninsula which is frequently exposed to the typhoon at summer. The main span length of the bridge was finally determined as 1545 m to ensure the safety of heavy ship traffics in Kwangyang Harbor. In order to design the Kwangyang Bridge, a series of wind tunnel tests and analysis were performed to investigate possible aerodynamic problems. Present paper summarizes the details of aerodynamic issues in the design of the Kwangyang Bridge. In the first stage of design, wind climate analysis was performed in conjunction with basic design. For estimating the design wind speed at bridge site, the annual maximum wind speed at local weather stations near the bridge site within 20 km were used. Moreover Monte-Carlo based typhoon simulation was performed to estimate the wind velocity independently. In accordance with wind analysis, preliminary cross sectional shape of the girder was derived based on the experiences of engineers, data base of existing bridges, design codes and computational fluid-structural interaction analysis. Main stage of design was focused on development of girder cross sectional shape from two dimensional wind tunnel tests. The cross sectional shape of main girder was gradually modified from preliminary one by wind tunnel models with three different scales to maximize aerodynamic stability and to minimize drag force. The section was changed from single box shown in Figure 1(a) to streamlined twin box section in Figure 1(b) finally. To reduce the drag force acting on girder, gantry crane rail was repositioned into the corner edge. In the final stage of design, global analysis combined the structural model and aerodynamic data were carried out to provide necessary data for the detail design. Finally the aerodynamic stability was confirmed from the full bridge tests for completion as well as construction stage.

Figure 1. Cross sectional shapes of girder.

456 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Planning, design and construction of the largest concrete pylon in the world

Sang-Hee Lee & Woo-Sun Jang Structural Division, Highcon Engineering Consultants, Seoul, Korea Seung-Bae Oh Geotechnical Division, Highcon Engineering Consultants, Seoul, Korea Kyung-Taek Kim Bridge Engineering Team, Daelim Industrial Co., Ltd., Seoul, Korea

This paper introduces the shape plan, design and construction sequence of the largest concrete pylon in the world which will face the symbol of the long-span suspension bridge (also known as Gwangyang bridge) which connects Myodo and Gwangyang bay area. Considering sag(L/9) of main span 1,545 m and required clear space under deck, it is planned to reach 270 m in height. This pylon is the second highest bridge pylon, shorter than Akashi Kaikyo (Japan, H = 283.8 m steel material), and taller than Great Belt East bridge (Denmark, H = 254 m concrete). But it is the largest concrete pylon in the world. The shape of the pylon has been analyzed referring to examples of worldwide long-span sus- pension bridges and closely examined using prototype model and CG. In the design detail, straight line of the inner part of the pylon is harmonious with the curved line of the outer part of the pylon. This portrays the beauty of the lines which represent imposingness and softness. The shape of the pylon at the base configures the container-ship line, in order to portray the rising of Marine Korea towards the center of the world. Also, the pylons cross beam and stiffening girder’s position is decided according to the traditional stone pagoda’s perfect ratio, in order to portray beauty in sight and stability. During the basic planning period, the foundation’s form was to have open caisson foundation, but there was possible damage hazard on Liquid Pitch Tank-Terminal which was adjacent to the foundation due to blasting of submerged open caisson. After close examination, there was pos- sibility that Liquid Pitch Tank-Terminal’s electric supplying equipments can be damaged, so the foundation’s form has been changed to drilled shaft foundation. Pylon is a important part of the bridge which supports the main-cable, and restrains the stiffening girder’s lateral displacement. The pylon top receives both compressive force and hor- izontal component caused by the main-cable. Also, it receives the effect of load caused by the wind which directly affects the pylon’s legs, the displacement caused by wind load are restrained by the main-cable. The calculation for the section forces is accomplished using non- linear analyzing, due to the consideration of long column effect according to whole system and independent pylon. In order to understand the stress flow on the pylon top, foundation or any other stress concentrated section and structural behavior of the pylon, the FEM analysis was carried out. The concrete pylon’s method of construction considered sufficient concrete supply and qual- ity management, mold application plan for the high placed pylon, and etc. uses constructed form of mold and footing to apply Auto-Climbing Form; which is constructing method that rises automatically. There has been examination regarding the constructability about assembling pre- fabricated reinforcement bars, which is possible to shorten construction term. The establishment of curved slant pylon’s perpendicularity and structural stability has been confirmed. the efficiency

457 of temporary structures has been ensured by minimizing the weight division of temporary shoring with dividing cross beam in the concreting. This bridge is currently under construction, and has a plan to open before the 2012 Yeosu International Exposition. The best efforts will be strived to realize a construction that considers bridge scale, given conditions for the job site, construction term of works and etc.

458 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

The design for anchorage of sea-crossing long-span suspension bridge

Hak-Sung Jang, Young-Il Jang, Young-Suk Choi & Kiung Park Geotech and Tunnel Division, Yooshin Engineering Corporation, Seoul, Korea Kyung-Taek Kim Bridge Engineering Team, Daelim Industrial Co., Ltd., Korea

This paper intends to introduce the anchorages supporting the long-span suspension bridge con- necting Myodo and Gwangyang in Korea. These anchorages are designed to support cable force larger than 400 MN. Myodo and Gwangyang being definitely exhibiting different geological char- acteristics, different types of anchorage have been applied in each side. South anchorage (AN1, Myodo side) is designed as rock anchored type with 35 m anchor length to use the resistance of sound rock at Myodo considering the joints and beddings of rock. The overall stability check of the anchorage considered joints, bedding planes, fault zones, condition of rock structure in situ by photo-lineaments, aerial photograph interpretation and drill-hole logs through the stereographic projection method. This anchorage consists of an access shaft, adit and concrete bearing plate to introduce prestressing force into rock mass. North anchorage (AN2, Gwangyang side) is sat on the bed rock underlaid by soft clay layer of 21 m thickness. Considering these geotechnical characteristics, AN2 designed as gravity anchorage type presenting diameter of 68 m and depth of 34.8 m was proposed to be constructed by installing circular diaphragm wall. The diaphragm wall will have thickness of 1.4 m and will be inserted into soft rock to a depth of 5.0 m to secure fixed condition for the wall. For the stability of trench in diaphragm wall, soil cement wall with diameter of 0.55 m will be installed. The inner area of the wall of AN2 is planned to be filled up with plain concrete by roller compacted concrete method. The factors for resistance of north anchorage for the cable tension of the bridge are self weight and frictional resistance between the anchorage base and the rock.

459 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

IDC for economical and safe design for the suspension bridge connecting Myodo and Gwangyang

Chang-Su Kim, Woo-Jong Kim & Kyung-Sik Cho DM Engineering Co., Seoul, Korea Jae-Hong Kim Daelim Industrial Co., Seoul, Korea YasutsuguYamasaki Maunsell Consultants Asia Ltd., Shatin, Hong Kong

Once completed, the suspension bridge (1545 Bridge) connecting Myodo and Gwangyang will be the longest suspension bridge in Korea and will rank at the third position among the currently longest bridges in the world. Therefore, its planning and design are requiring thorough preparation. The necessity of independent design check (IDC) was brought up together with detailed design because 1545 Bridge is a very long span suspension bridge, complex and abstruse structure, and should secure design quality at the most. Moreover, owing to the short period allotted for the basic design and the start of construction immediately after completion of the detailed design, the IDC has been organized during detailed design stage for the first time in Korea. To save time, the IDC reviewed the basic design for the safety and the optimization from overall design, constructability and economical points of view, pointed out features to be re-examined during detailed design, proposed alternative or optimum details when available, and confirmed the changes made in the detailed design. IDC’s works proceeded as follows; (1) review of basic design, (2) selection of complements and alternatives and, (3) application in the detailed design. The first achievement of IDC was the review and evaluation of basic design’s results including the design criteria, design report, technical calculations and drawings, followed by in-depth check for bridge overall behavior, structural components and construction issues. In the bridge overall behavior, global analysis was accomplished independently along with the review of bridge articulations. The structural components comprise the cable works, suspended deck, tower and side pier. As for construction issues, a set-back system, catwalk system, aerial spinning system and deck erection system were checked. The review of basic design concentrated on the structural appropriateness and safety. But before the detailed design stage, proper alternatives were proposed considering economical and safety aspects and attempt was done to apply them to detailed design. These alternatives were thus established based on the basic design and optimized. From the review of basic design, more refined complements and alternatives were proposed, and then the optimized option was chosen by careful review of each item. During detailed design, IDC reviewed the changes of detailed design by real-time schedule and transmitted the opinions and check results to designer. A large number of IDC’s results were reflected in detailed design with active cooperation between designer and IDC member. The principal check points of IDC involved the global behavior, suspended deck, cable works, tower and foundation. Finally, focus was given to apply IDC’s results in the detailed design in an attempt to achieve the most safe and economical design for the longest suspension bridge in Korea.

460 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

The innovative construction method for the long-span suspension bridge connecting Myodo and Gwangyang in Korea

Sang-Hoon Shin & Pil-Jo Yu Structural Division, Yooshin Engineering Corporation, Seoul, Korea Seung-Wook Jeong Bridge Engineering Team, Daelim Industrial Co., Ltd., Seoul, Korea Yukiaki Takizawa Infrastructure Division, IHI Corporation, Tokyo, Japan

This paper introduces the innovative construction methods adopted for the 2260 m long-span sus- pension bridge to link Myodo Island and Gwangyang City, Korea, in the third section of approach road to the Yeosu National Industrial Complex. A few new construction methods are adopted for this bridge with main span of 1,545 m and pylon heights of 270 m. The bridge will be located at the gate of Gwangyang Port, a main international trading port with heavy shipping traffic. Firstly, for main cable, high-strength wire is selected to minimize the number of wires and to reduce the weight of cables, and a rectangular shape of strands is chosen in order to install two strands simultaneously. The catwalk system with adjustment function is adopted not only to apply uniform tensile force to each strand, but also to improve wind-resistance performance and simplify the structure for saving erection time and dismantlement of the catwalk. Secondly, for erecting stiffening girders, swing method is applied to open navigation channel all the times during construction, and 4-point lifting method is planned for the installation of balanced cantilevers. Finally, for the reinforced concrete pylons, auto-climbing form method, known for its adaptability to the sectional changes, is applied, and GPS is planned to secure vertical accuracy of pylons. Considering the scale of the bridge, site conditions, and construction period, optimal construction methods were developed and adopted to increase the efficiency of erection and secure safety.

461 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

The planning of ship collision protections based on risk analysis

Ho-Chul Kwon, Myeong-Jae Lee & Jin-Ho Park Structural Division, Yooshin Engineering Corporation, Seoul, Korea Henrik Andersen Major Bridges, COWI, Kongens Lyngby, Denmark

In this paper, the planning of ship collision protections based on risk analysis of Gwangyang Bridge (tentative) is introduced. The objective of risk analysis is to assess the risk with respect to ship impacts of the bridge. The present and future ship traffics (year 2030) across the alignment have been estimated based on the data. The characteristics of the waterway such as the bathymetry, currents, etc. have been determined based on the data. Based on the findings of the risk analysis, several alternatives for protective structures have been assessed. For the selected protection alternatives the design criteria have been specified. This has been carried out iteratively with the resistances for pylons and piers. Finally, the protective structures have been optimized based upon impact simulation and risk calculation. The exposure of Gwangyang Bridge with main span of 1545 m and its individual components has been assessed on the basis of the annual collision frequency. The annual collision frequency for the future ship traffic scenario (year 2030) is 0.15 per year which corresponds to a mean time between collisions of 6.7 years. It has been found that the north pylon is the component that is exposed most and accounts for approximately 85% of all ships colliding with the bridge. It is therefore considered to give a high priority to measures reducing the risk for the north pylon. Three design alternatives have been studied. Each of them is acceptable with respect to the AASHTO guideline. The selected design considers a pylon foundation using two 30 m diameter caissons, a block wall protective island around the pylon foundations, the Myodo pier (south pier) foundation using a 16 m diameter caisson, and a sub sea reef protecting the south pier (Table 1).

Table 1. Types of ship collision protections.

462 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Measurement and evaluation of data from wind observation station in Gwangyang

Seong-Lo Lee, Gwan-Mun Han & Young-Sun Gwon Mokpo National University, Mokpo, Korea Yong-Gwi Bae Oceanspace, Seoul, Korea

The new construction for the approach road toYeosuNational Industrial Complex in Jeollanam-do (2007–2012) is a large-scale project for the construction of a 8.5 km bridge (5.2 km sea-crossing), connecting Wolnae-dong of Yeosu, Myodo, and Geumho-dong of Gwangyang, with a budget of KRW88 billion. This project involves the design and construction of sea-crossing long-span bridges that are a suspension bridge and a cable-stayed girder bridge. To secure stable wind-resistance of the long-span bridges and safety of construction, Jeollanam-do and Korea Wind Engineering Research Center of Mokpo National University agreed to install wind observation stations rising at a height of 60 m, and to use their observation data to develop a wind model for Gwangyang-Myodo area. The wind observation station was installed on the idle site in Gwangyang POSCO steelworks nearby the construction site of long-span bridges connecting Yeosu, Myodo and Gwangyang, which will be constructed on the approach road to Yeosu National Industrial Complex in Korea. There were no obstacles around the measurement area. The wind data measurement mast was a 60 m tall tubular tower, which was erected in July 2007. The mast has three wind vane anemometers, which were at 10, 35 and 60 m heights, and one 3-dimensional ultrasonic anemometer at 60 m height. Temperature, relative humidity and atmospheric pressure data were obtained from a thermometer, a hygrometer and a barometer, respectively. A data logger was connected with all sensors on the mast to collect data in a time series. Furthermore, a computer was used to transfer the data from the data logger to be evaluated. Wind data was evaluated to obtain the correlation of wind velocities at the two heights, namely, power law exponent of vertical profile of wind speed. Basic wind speed for the design of all parts of a bridge is generally extrapolated from long-term wind data of nearby weather station. Extrapolation of long-term wind data was performed taking account of topography of Yeosu and Gwangyang region and short-term observations from several wind sensors on 60m mast. Also descriptors of atmospheric turbulence such as turbulence intensities, integral turbulence scales and spectra of turbulent wind speed fluctuations were evaluated for engineering calculations of buffeting effects on bridge. Wind observation tower could miss out data. To solve this problem, neural network will be used to estimate wind velocity. The estimated results are expected to be mostly feasible and will be used to fill in missed out data in the future. The above analysis models will be used to develop a wind environmental model for the Gwangyang-Myodo region. In addition, Computational Fluid Dynamics (CFD) will be used to estimate wind velocity at proposed sites and analyze correlation between wind and temperature/humidity. The field measurements of wind data continue and thus longer periods, in terms of years, can yield quantitative, representative as well as persuasive results.

463 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Vehicle-structure dynamic interaction by displacement constraint equations and stabilized penalty method

Keun-Young Chung & Jae-Min Kim Department of Civil & Environmental Eng., Chonnam National University, Korea Myung-Kwan Song Corporate R&D Institute, Doosan Heavy Industries & Construction Co., Ltd., Korea Jong-Gyun Paik Bridge Engineering Team, Daelim Industrial Co., Ltd., Korea

In this study, to realize the super long-span suspension bridge with twin box-girder deck sys- tem, a close investigation of the dynamic serviceability is carried out. This paper investigates vehicle-structure dynamic interaction effects on dynamic response of the Gwangyang Grand Bridge subjected to running vehicles. In order to model various types of vehicles, a lumped-parameter vehicle element is proposed. The proposed vehicle element is a basic vehicle including three lumped masses interconnected with elastic/inelastic springs and dash-pots, and the element also has a contact spring. By introducing constraint equations between masses of the vehicle elements, the motions of various types of vehicles can be described. In this study, two different types of the displacement constraint equations are defined to connect between car bodies, and/or between bolsters. These constraints are the rigid body connection and the rigid body connection with a pin. To cope with the possible abrupt changes of road surface height due to girder ends deforma- tion, the height of the contact point is filtered considering contact patch between wheel tire and road surface in an average sense. The displacement constraint equation on the contact point of a vehicle element is implemented by the penalty method with stabilization, in which a damping force is artificially introduced. The artificial damping force is in proportional to the negative rate of change of the constraint equation and parallel to elastic force due to artificial stiffness in the penalty method. As a result, the penalty force arisen in the stabilized penalty method is dependent on not only prescribed displacement of a contact point but also artificial damping value and veloc- ity of a contact point. Owing to these techniques, it is possible to stabilize the penalty method in the numerical time integration process without deteriorating accuracy of the solution. For the time integration of dynamic equations of vehicles and structure, the Newmark time integration scheme is adopted. To reduce the error caused by inadequate time step size, adaptive time-stepping technique is also employed. Finally, with the aforementioned analysis schemes, the vehicle-bridge interaction analysis of Gwangyang Grand Bridge is carried out, and its behaviors are briefly discussed.

464 Loads and capacity assessment

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Evaluation of ultimate capacity of deteriorated reinforced concrete bridge columns

M. Tapan Yuzuncu Yil University, Department of Civil Engineering, Van,Turkey R.S. Aboutaha Syracuse University, Department of Civil & Environmental Engineering, Syracuse, NY, USA

Good understanding of the effects of deterioration on the structural performance of concrete columns leads to better evaluation procedures, planning, and cost-effective rehabilitation meth- ods. Corrosion of steel reinforcement is the principal cause of deterioration of concrete bridge elements. Corrosion of steel bars develops as a result of concrete carbonization, introduction of chloride ions, oxygen penetration, insufficient cover, etc. In order to accurately evaluate the remain- ing strength of a deteriorated concrete column, the effect of corrosion on the column’s physical condition, geometric and material properties should be quantified. The ultimate strength of corroded concrete bridge columns can be quantified by developing interaction diagrams based on the actual section and material properties of the deteriorated column. The analysis procedure used in this study accounts for amount of corrosion, exposed corroded bar length, concrete loss, bond failure, and type of stresses in the corroding reinforcement. This paper presents axial load-moment (P-M) interaction diagrams for deteriorated concrete col- umn sections. It focuses on the effect of corrosion of compression bars on the ultimate load carrying capacity of deteriorated reinforced concrete bridge pier columns. The effect of exposed corroded bar length on the load carrying capacity is also investigated using stress-strain diagrams developed for three different exposed compression bar lengths (Lexp = 50 cm, Lexp = 100 cm, Lexp = 150 cm). The results of this study indicate that corrosion of compression bars significantly reduces the load carrying capacity of deteriorated columns. It was also found that relatively small amount of corro- sion would result in significant strength reduction, as the compressive force in corroded steel bars drastically decreases with the increase in the exposed length of the corroded compression bars.

467 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Load test of bonded post-tensioned concrete beams with corroded tendon

S.G. Youn, S.H. Park, C. Lee & E.K. Kim Department of Civil Engineering, Seoul National University of Technology, Korea

ABSTRACT: The paper describes ultimate load tests which were performed to show the effects of prestress and tendon corrosion on the flexural strength of post-tensioned concrete beams. Five test specimens were fabricated in laboratory with the variations of the prestress of tendons and the loss of tendon area. For two specimens, small area of tendon at the center of the beam was exposed by using diameter 25 mm drill and the exposed tendon was corroded intentionally using accelerated corrosion equipment. Test results show that the ultimate flexural strength of test specimens with corroded tendons is smaller than the predicted flexural strength which is calculated considering the loss of tendon area. From the comparison, it is considered that the estimation of flexural strength of prestressed concrete beams with corroded tendons is very difficult just based on the loss of tendon area obtained by one-side visual inspection. In addition, it is very hard to determine the effective tendon area which can be guaranteed to show fully ductile behaviors.

1 INTRODUCTIONS

Many bonded post-tensioned concrete bridges have exhibited tendon corrosion and wire fractures caused by voids in ducts and ingress of water and chloride ions. For safety, strength evaluation of these bridges is required at the time of investigation, but limited information about corroded tendons can be lead to wrong direction or evaluation. This paper describes the results of an experimental test for estimating the ultimate flexural strength of prestressed concrete beams with corroded tendons. The residual tendon area of corroded tendons estimated by the information obtained by one-side visual inspection is used for calculating theoretical flexural strength. During the tests, acoustic signals come from inside the test specimens are monitored by a continuous acoustic monitoring system to detect premature wire fractures. Test results of ultimate flexural strength are compared with those of theoretical values. Based on the research work, strength evaluation method including tendon corrosion is discussed.

2 DETERIORATION OF POST-TENSIONED CONCRETE BRIDGES

Evaluation of deteriorated post-tensioned concrete bridge is needed to determine strength and safety at the time of investigation. If corroded tendon is found, it is required to assess how much section has been lost from the tendon. If corroded tendons are assumed to show fully ductile behaviors, residual area of prestressing steel can be applied to calculate residual flexural strength. However, chloride-induced corrosion is believed to have contributed to pitting corrosion which led to wire fracture. Therefore, if some corroded wires are excluded, Equation 1 can be used for strength evaluation. Due to the lack of technical information, any reliable guideline for the effective tendon area AEP is not provided yet and thus, whole corroded tendons are likely to exclude for the safety.

where, AEP: Effective tendon area.

468 3 LOAD TEST OF POST-TENSIONED CONCRETE BEAMS WITH CORRODED TENDON

The test results of ultimate flexural strength were smaller than the calculated values which obtained by using the residual area of prestressing steel. It is considered that the premature wire fractures are mainly the cause of decreasing the ultimate flexural strength. It means that ultimate flexural strength calculated by using the residual area of prestressing steel, ARP, is not safe for strength evaluation. Therefore, strength evaluation should be estimated by Equation 1 which is obtained by using the effective tendon area, AEP.

4 CONCLUSIONS

Strength evaluation of deteriorated prestressed concrete bridges will be a big issue in near future. Many bonded post-tensioned prestressed concrete bridges in Korea have been constructed without serious consideration of the tendon corrosion and the limitation of inspection method. This paper describes the results of load tests performed to show the effects of prestress and tendon corrosion on the ultimate flexural strength of post-tensioned concrete beams and a way to estimate the effective tendon area for strength evaluation is proposed. This paper will be useful only at the laboratory level and there are so many uncertainties to resolve. Because of the uncertainties of whole tendon condition, strength evaluation of deteriorated post-tensioned concrete bridges is a very complicated task for structural engineers. To overcome these limitations, experimental works for severely deteriorated post-tensioned concrete beams are extensively required to investigate ductility and ultimate flexural strength and reliable method or guideline to estimate the effective tendon area should be established.

ACKNOWLEDGEMENTS

This study is supported by the Infra-StructureAssessment Research Center (ISARC) and the authors would like to thank ISARC.

REFERENCES

Woodward, R.J. & Williams, F.W. 1988. Collapse of Ynys-s-Gwas bridge, West Glamorgan. Proceedings of Institute of Civil Engineers, Part 1, 84: 635–669. Mathy, B., Demars, P., Roisin, F., & Wouters, M. 1996. Investigation and Strengthening -Study of Twenty Damaged Bridges: A Belgium Case History. Proceedings of the 3rd International Conference on Bridge Management, London: 658–666. Woodward, R.J. Hill, M.E. & Cullington D.W.1996. Non-destructive methods for inspection of post-tensioned concrete bridges, FIP Symposium on Post-tensioned Concrete Structures: 295–304. Halsall, A.P. Welch, W.E. & Trepanier, S.M. 1996. Acoustic Monitoring Technology for Post-Grouted Structures, FIP Symposium on Post-tensioned Concrete Structures, London: 295–304. Youn, S.G., Cho, S.K., Kim, E.K., & Lim, J.Y. 2004. Detection of Corrosion-induced Wire Fracture using Acoustic Emission Technique. Proceedings of the First International Conference of Asian Concrete Federation, Vol. 2 Chiang Mai, Thailand: 1005–1012. Cullington, D.W., MacNeil, D., Paulson, P.,& Elliot, J. 1999. Continuous acoustic monitoring of grouted post- tensioned concrete bridges, Structural Fault & Repair-99, Proceedings of the 8th International Conference, London: 47. Fricker, S. & Vogel, T. 2006. Detecting wire breaks in a prestressed concrete road bridge with continuous acoustic monitoring. Proceedings of the Third International Conference on Bridge Maintenance, safety and Management, Porto, Portugal: 847–849. The Concrete Society, 1996. Durable Bonded Post-tensioned Concrete Bridges. Concrete Society Technical Report 47. the Concrete Society, TR047.

469 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

The bridges and the floods

V.Popa Consitrans, Bucharest, Romania

The bridges are engineering structures that ensure the continuity of a road over an obstacle, generally over a river. The length of the bridge over a river is established by hydraulic calculus based on the water flow and hydraulic characteristics of the riverbed. Sometimes the water volume from the river is too much so that the water level overgrows and outruns the river banks generating floods. The phenomenon of over-passing the space ensured for the prescribed water flow in the river bed has a generic name of “flood”. The amplitude of the flood could be smaller or higher, but sometimes it could has catastrophic consequences, causing important loses of materials and even human lifes. These floods cause many damages to the bridges, culminating even with their collapse. The paper presents some cases of bridges affected by floods in Romania. The long term and abundant rains during the spring and summer of the year 2005 caused important damages to many bridges in our country, culminating with collapse of some of them. The paper also analyses the causes that led to bridge problems and the measures which have to be considered in order to avoid such troubles and to save the bridges during the floods. Generally, the mistakes of the bridge length dimensioning, bad placement of the piers or wrong foundation deep are the main causes that induce the troubles during the floods. The paper includes the principles and methodology for the hydraulic calculus of the bridges and ways of designing a bridge in order to be safety during the floods. Finally, the paper presents measures and recommendations for the rehabilitation of the bridges affected by floods. In conclusion, the author shows that the bridges could be designed and constructed without the risk of being affected at the floods, if the project and execution take into consideration the correct principles and recommendations shown in the paper as well as the real data regarding water flow and hydraulic characteristics.

470 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

3 DOFs collision model for the analysis of bridge super-structures and deck house collision

Gye Hee Lee & Seong Lo Lee Mokpo National Maritime University, Chonnam, Korea

3 DOFs model for the collision analysis of a bridge super-structure and a deck house of navigating vessels were proposed and analyzed. The collision event between the deck house or the high-rised structure of vessel and the super-structure of bridge are differ from the normal collision event that occurred at sub-structure of bridge. Because of its long moment arm, the stability force of vessel could affect the collision behaviors. To consider this effect, 3 DOFs model including two translation DOFs and one rotational DOF were introduced (Fig. 1). The restoration forces of the collision system were considered as nonlinear springs. The equations of motion were derived in form of differential equations and solved by 4th order Runge-Kutta method. The accuracy and the feasibility of this model were verified by the numerical example with parameter of moment arm length.

Figure 1. 3 DOF collision model.

471 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Modelling granular soil to predict pressures on integral bridge abutments

J. Banks, T. Knight & J. Young Mott MacDonald, Croydon, UK A. Bloodworth University of Southampton, Southampton, UK

Integral bridges eliminate many of the problems associated with expansion joints and bearings used on conventional bridges. However, such structures are not free from problems, and of particular concern are the lateral soil pressures which occur on the abutments. The cyclic nature of the abutment displacements, caused by thermal loading of the deck, results in increased lateral soil pressures from the granular backfill. Previous experimental work at the University of Southampton investigated the fundamental behaviour of a granular soil element under integral bridge loading. A literature review showed that no existing constitutive soil model could replicate the observed behaviour. A constitutive soil model has therefore been developed based upon the experimental results. This constitutive model was designed to take into account the changes in radial strain and density resulting from the rolling/sliding behaviour of soil particles close to the active state that the previous work found to be so important in the granular soil’s behaviour. The model was implemented into FLAC, a finite difference method program, and validated by modelling of the experimental triaxial test. The experimental work upon which the constitutive model was based had only investigated the soil at the element level. However, in order to use the model to investigate the lateral soil pressure behind a whole abutment, a FLAC model was required with a grid of elements to represent the abutment/soil system. This paper describes the initial series of such analyses that were carried out in FLAC. The aim of these analyses was to investigate whether the soil model worked as expected at the whole-system level, and to test the influence of Young’s modulus profile on the resulting soil pressures. Published data from centrifuge model tests was used for validation for this phase of modelling. The values of lateral soil pressure on the abutment predicted by the constitutive model were compared to the experimentally measured values, showing that the constitutive model performed as expected. The lateral stress distributions that developed for three different profiles of Young’s modulus were compared. It was found that the value of Young’ modulus at the surface, and the gradient of the profile, could both have a significant impact on the resulting lateral soil pressures. This was evidence that further work to carry out a full parametric study of the Young’s modulus profile would be needed to assess its impact when using the developed granular soil constitutive model. This research has shown that the type of constitutive model developed has the potential to provide good estimates of soil pressure. This could be used to further investigate integral bridge abutment pressures with the aim of improving and refining the design of such structures.

472 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Assessment of bridge capacity through proof load testing

J.D. Gómez & J.R. Casas Universitat Politècnica de Catalunya, Barcelona, Spain

ABSTRACT: Aging is a matter of increasing concern for most bridges that are part of the road and railway systems of the European Union. Many of these bridges are very old bridges without documentation. It is common to find that old bridges carrying normal traffic satisfactorily fail to pass the assessment calculation. In other cases, bridges that were designed with old codes or standards fail to pass the assessment when new codes prescribe higher levels of live load. These bridges are ranked as deficient, so they need to be posted, strengthened or in the worst case closed and replaced. A possible way to assess their capacity is by means of a proof load test. Two main questions should be solved when facing the execution of such a test: a) Which is the load level that the bridge should resist during the proof test to guarantee a predetermined service load with a required and appropriate safety level?, and b) When the increase of loading should stop in order to not damage the structure? The answer to these questions is part of the task of the EC 6th Framework Programme European Project ARCHES (Assessment and Rehabilitation of Central European Highway Structures). The paper shows the basis of the method as well as the application to an existing bridge located in Barcza (Poland). The assessment via proof load test is analyzed here in a reliability-based perspective. In figure 1 is shown the influence of the type of distribution, variability and percentage of characteristic live load as defined in the Eurocode, in the proof load effect (bending moment) to be introduced in the most loaded beam of the bridge. As can be seen in figure 1, the type of distribution is almost negligible for low values of the coefficient of variation of the traffic load. However, the coefficient of variation seems to have a higher high influence. The influence of the percentage of total traffic in the proof load effect is almost linear independently of the type of distribution and coefficient of variation.

Figure 1. Proof load effect (bending moment in the most loaded beam) for β = 2,3.

473 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Micro-simulation modelling of traffic loading on medium- and long-span road bridges

E.J. OBrien University College Dublin, Ireland A. Hayrapetova Roughan ODonovan, Consulting Engineers, Ireland C. Walsh Trinity College Dublin, Ireland

ABSTRACT: This paper presents a new approach to the modelling of congested traffic loading events on long span bridges. Conventional traffic load models do not account for the mixing between lanes that takes place as traffic becomes congested. It is shown here that cars move out from between trucks as traffic slows down which results in a higher frequency of long platoons of trucks. These longer platoons increase characteristic load effects under the slow lane by a modest but significant amount – about 10% for the bridge considered. Micro-simulation, the process of modelling individual vehicle/driver behaviour that is widely used in traffic modelling, is presented here as a means of predicting imposed traffic loading on long-span bridges more accurately. The traffic flow on a congested bridge is modelled using a random mixing process for trucks and cars in each lane, where each vehicle/driver unit is modelled individually with driver behaviour parameters assigned randomly in a Monte Carlo process. Over a number of simulated kilometres, the vehicles move between lanes in simulated lane-changing manoeuvres. The algorithm was calibrated against video recordings of actual driving on a 100 m span bridge in the Netherlands – Figure 1. The driver behaviour properties are adjusted to achieve the observed distributions for numbers of lane changes, etc.. After calibration, a good correspondence is achieved between the observed and simulated histograms of truck platoon length. A collapsing of the gaps between trucks results in fewer long platoons. The bridge used to validate the microsimulation approach is heavily trafficked with a significant number of trucks and is subject to congestion twice daily with fully stopped and stop-and-start driving cases. Strain sensors on this bridge were used to collect maximum-per-hour strain data. Acknowledging that the data may not have converged to an Extreme Value distribution, it is fitted to a Normal distribution raised to the nth power where n, the number of independent events, is found empirically. Observed and microsimulated data are used to determine characteristic mid-span bending moments for a number of return periods. There is a good match between characteristic observed and microsimulated results. The stay-in-lane case is also considered and is found to result in characteristic bending moments that are 9% to 11% less than those found from the measured data.

Figure 1. Congested traffic on the Moerdijk Bridge, actual (left) and microsimulated (right). 474 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Nonlinear analysis of PSC structures with internal tendon by strengthened using external tendon

Jae-Guen Park, Ju-Hyoun Cheon, Monn-Young Kim & Hyun Mock Shin Sungkyunkwan University, Suwon, Korea Byeong-Ju Lee Korea Highway Corporation, Hwasung, Korea Jung-Ho Choi Hankyong National University, Ansung, Korea

In recent years, external prestressing using PS tendon has become a primary method for strengthen- ing existing concrete structures and has been increasingly used in the construction of newly erected ones, particularly segmental bridges. This paper presents an analytical prediction of Nonlinear characteristics of prestressed concrete bridges by strengthened of externally tendon considering the work sequence, Using beam-column element and bonded and unbounded tendon element. The beam-column element was developed with reinforced concrete material nonlinearities which are based on the smeared crack concept. The fiber hysteresis rule of beam-column element is derived from the uniaxial constitutive relations of concrete and reinforcing steel fibers. The tendon element represent the bonded and unbonded tendon behaviors. Beam-column element and tendon element was be subroutine A computer program, named RCAHEST (Reinforced Concrete Analysis in Higher Evaluation System Technology). The proposed numerical method for prestressed concrete structures with internal tendon by strengthened of externally tendon is verified by comparison with reliable experimental results.

Figure 1. Experiment and analysis sequence of internal tendon cutting and strengthen of external tendon.

475 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Enhancement of bridge serviceability due to a strong wind

A. Krecak, P. Sesar & M. Masala-Buhin Civil Engineering Institute of Croatia, Zagreb, Croatia

This paper presents some facts about wind protection structures – from the design issues to assessment of their efficiency during strong wind occurrences, namely wind speed reduction ratio. Built wind protection structures in Croatia are situated on the motorway section from Tunnel Sv. Rok to Maslenica Interchange which passes through rough terrain. Location of the motorway section is on the rocky slopes of the Velebit Mountains where inland meets seacoast: Geographical complexity severely influences generation of strong winds in this region. The strongest one is bora, which mostly blows from north and northeastern direction. It reaches extreme velocities which directly influence traffic regimen on the motorway. Therefore it was decided to invest in wind barrier structures along the motorway section that would help increasing serviceability as well as traffic safety on the mentioned motorway section. Several types of structures were proposed regarding their efficiency and aesthetic issues. At they were tested in wind tunnel to prove their efficiency on physical models in scale 1:5. Special attention was given to designing wind barriers on the several viaducts along the motorway in order to avoid unfavorable effects on vehicles. Wind speed near and on the motorway can be locally decreased by application of adequate protection systems. Permeable wind protection systems enable an acceptable level of wind speed reduction on certain locations on the route and on the viaducts. On viaducts, permeable protection systems should not significantly increase the load on viaduct structure. This project is part of the research which aim is also to establish criteria for conduction approximately continuous driving conditions along the motorway route. One of the built types of wind protection structures is porous wind barrier on the viaduct Barice- vic. It is quite sophisticated structure similar to several wind barriers recently built on viaducts in the world (Millau Viaduct in France, Crni Kal Viaduct in Slovenia). After it was built, measurements of wind speed on the viaduct with protecting structure were done. Both statical and dynamical mea- surements were done. Statical measurements comprise measurements on the meteorological station while dynamic measurements are done by means of measurements set installed on the roof of the car which drives along the motorway and measure influence of wind on the driving vehicle. Driving speed is excluded from the analysis and only dynamic influence is observed (i.e. turbulences). Results are showing that reduction of the wind speed satisfy demanded level necessary for safe traffic ongoing. This proves efficiency of porous wind protection structure on the viaduct. Similar structure can be built from steel deflectors as well. The main reason for choosing plexiglas deflectors is achieving better visibility and avoiding “tunnel effect” on the viaduct. Research in this field is still under way on the scientific project on protecting motorway from extreme wind, which is run by Civil Engineering Institute of Croatia. Results presented in this paper concern only one type of wind protection structure implemented on viaduct.

476 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Comparison of theoretical and measured temperature distributions for concrete slab bridges

E.-S. Hwang Kyunghee University, Yongin, Korea J.J. Lee Korea Bridge Design & Engineering Research Center at Seoul National University, Seoul, Korea

For short span bridges, reinforced or prestressed concrete slab bridges are widely used in the country. About half of all bridges in Korea is actually slab bridges. Of many loads affecting the structural behavior of the bridge, thermal effects are usually considered as the variation of the average temperature and temperature gradient. In the design code of Korea (MOCT, 2005), the variation of average temperature is specified for concrete bridges but temperature gradient is not specified for slab bridges and is not considered in the design. This paper deals with comparison of the experimental and theoretical study on temperature distribution for the slab bridges. For the experimental study, long-term measurements of temperature distributions along the height of slab have been performed on a rahmen type slab bridge. 20 temperature sensors were installed during the construction of the bridge. Hourly temperature data are collected using a data acquisition system for about four years to find the daily and seasonal variation of temperature distribution of the slab. Effects of temperature distribution of the section can be divided into three parts; average (uniform) temperature, linear vertical temperature difference, and remaining self-equilibrating temperature. From the measured data, temperature parameters, such as average temperature and linear vertical temperature difference of the section, are calculated. Environmental parameters, including ambient temperature, solar radiation, wind speed, and etc., have significant effects on design temperature load for bridges. Since these parameters are random variables, the analysis should be based on probability and statistics. In this study, the estimation method for design temperature load using long-term thermal analysis and extreme analysis is proposed. The thermal analysis utilizes the two-dimensional finite element model and is verified through measured data in a constructed concrete slab bridge. Values of ambient temperatures and solar radiation intensity are collected through meteorological stations near the bridge for experimental study and are used for input data. Appropriate values of concrete thermal properties are selected. Analyses are performed on various locations in the country. From the measured and theoretical data, temperature parameters (average temperature and linear vertical temperature difference of the section) are calculated and compared. Following conclusions can be made from the study. – From the comparison, appropriate values for solar radiation absorptivity coefficient, convection hear transfer coefficient, and emissivity are selected. It is found that temperature distributions from experimental and theoretical studies are similar. – From analysis of various slab depth, it is found that the temperature distributions from surface to about 0.45 m are almost identical. Temperatures are almost constant for lower part of the section. – The load model for the vertical temperature gradient is proposed as the second-order polynomial function from the top surface to 0.45 m below the top surface and constant temperature at lower portion.

477 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Numerical analysis of old masonry bridges supported by field tests

J. Bie´n& T. Kami´nski Wrocław University of Technology, Wrocław, Poland Ch. Trela BAM Federal Institute for Material Research and Testing, Berlin, Germany

ABSTRACT: The paper presents a complex approach to analysis of masonry bridges based on the FE modelling and supported by field tests. Some NDT and MDT methods and 2D and 3D FE numerical models are presented. Results of calculations are compared against the field tests. Finally sensitivity of the ultimate load of the structure to selected parameters is studied.

Old masonry bridges are very often devoid of any documentation providing essential technical data on the material properties or even basic geometry of the bridge. A reasonable solution of the problem should be based on results of the experimental tests but cannot require techniques disturbing the traffic or threatening stability of the bridge. Thus, preferably it should involve Non-Destructive Testing (NDT) or Minor-Destructive Testing (MDT) methods providing sufficient amount of information for further assessment of the structure. Description of an approach to the problem is explained with an example of an old masonry bridge in Ole´snica in Poland. The structure was extensively investigated within the scope of international cooperation organized by the Integrated Project “Sustainable Bridges – Assessment for Future Traffic Demands and Longer Lives”. The applied methods of testing were: georadar measurement, electrical conductivity measurement, endoscopy tests, dynamic penetration tests and laboratory tests on masonry and soil samples. On the basis of the testing results advanced 2D and 3D FE models of the structure are proposed. 3D model of the structure is presented in Figure 1a. The models are applied in analyses indepen- dently under service and ultimate loads. Exemplary failure of the arch barrel under the ultimate load received by means of the 3D model is shown in Figure 1b. The results of analysis under service load are verified by means of field load tests involving various modern and independent methods of deformation measurements: LVDT, laser and microradar techniques. Finally sensitivity of the ultimate load to selected bridge parameters investigated by means of the testing methods is analysed and thoroughly discussed. Conclusions on those parameters which are crucial for mechanical behaviour of the structure are given.

Figure 1. FE analyses: a) 3D model with removed part of the spandrel wall, b) failure mode of the arch.

478 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Input ground motion for seismic design considering near fault effects in stable continental regions

Jung Han Kim & Jae Kwan Kim Seoul National University, Seoul, Korea

This paper proposes a method for modeling design earthquake considering near fault effects in stable continental regions. The near fault ground motion is characterized by a single long period velocity pulse of large amplitude. In order to model the velocity pulse, its period and amplitude need be determined in terms of earthquake magnitude and distance from the causative fault. Because there have been observed very few near fault ground motions, it is difficult to derive the model directly from the recorded data in stable continental regions. Instead an indirect approach using fault rupture mechanism is adopted in this work. The two parameters, the velocity pulse period and its amplitude are derived by comparing related data between active tectonic regions and stable continental regions (Fig. 1). A time history of the near fault ground motion is synthesized by superposing method. The velocity pulse time history is represented by Gabor Wavelet formula. The pulse period and the maximum amplitude expected in moderate size earthquake magnitude in stable continental regions are determined. Far field ground motion is generated by the stochastic method to match the ordinary design response spectrum. Near fault ground motion in stable continental regions is generated by superposing pulse type long period term on stochastic high frequency term. And pseudo acceleration response spectrum of a single degree of freedom system excited by near fault ground motion in stable continental region is studied.

Figure 1. Pulse period and Vmax in stable continental regions.

479 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Calculation of dynamic interaction of train and an arch bridge

J. Györgyi Department of Structural Mechanics, Budapest University of Technology and Economics, Budapest, Hungary G. Szabó Pont-TERV Ltd., Budapest, Hungary

ABSTRACT: If we want to calculate the dynamic effect of a moving vehicle during the test of a railway bridge we have some possibilities to build up the model of the structure and the vehicle too. We considered an arch bridge and besides a truss bridge for comparison the dynamic sensitivity of them due to vehicle loading. We used three different FEM models to the arch bridge. There were the two-dimensional beam model, a three-dimensional beam and a fully shell model. We modelled the vehicle in three ways. At first the train was considered as only moving forces, after calculated by moving mass points too and at third we built up a dynamic model to the vehicle. If we model the train not only with moving forces, the matrices will be time dependent. Using the eigenvectors of the undamped system we can apply the quasi-modal analysis in this case. With this technique we can detect the critical speeds where the dynamic effects are large. On Figure 1 we show the maximum vertical displacements versus train speed curve in case of a monitored point of the arch bridge. Comparing the behaviour of the two bridge types we found great differences.

Figure 1. Maximum displacement versus train velocity curve of the arch bridge at a certain point of the structure.

REFERENCES

Bathe, K.J. & Wilson, E.L. 1976. Numerical methods in finite element analysis. New Jersey: Prentice-Hall. Frˇyba, L. 1972. Vibration of solids and structures under moving loads. Prague: Academia. Györgyi, J. 1996. Application of direct integration in the case of external and internal damping. Proceedings of the Estonian Academy of Sciences Engineering, 2: 184–195.

480 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Structural behavior of corroded reinforced concrete structures

K. Zandi Hanjari & K. Lundgren Department of Civil and Environmental Engineering, Chalmers University of Technology, Gothenburg, Sweden P. Kettil Department of Applied Mechanics, Chalmers University of Technology, Gothenburg, Sweden M. Plos Department of Civil and Environmental Engineering, Chalmers University of Technology, Gothenburg, Sweden

ABSTRACT: There is a growing need for reliable methods to predict the load-carrying capacity and remaining service life of deteriorated reinforced concrete structures as a decision basis for optimized maintenance and repair strategies. In an ongoing research project, load-carrying capacity of deteriorated reinforced concrete structures is studied on the component level. The part of the project presented here is focused on deterioration (damage) due to the corrosion of reinforcement. The corrosion process transforms steel into rust, leading to (a) area reduction of the reinforcement bars and (b) volume expansion that generates splitting stresses in the concrete, which may crack and spall the concrete cover and affect the bond-slip between reinforcement and concrete. This has been studied by many researchers; for a state-of-the-art report see fib (2000). While previous research has been mainly concerned with corrosion around a single reinforcement bar, see e.g. Lundgren (2003), relatively little research has been concentrated to the practically important problem of assessing the residual load-carrying capacity of corroded concrete structures. The aim of this research is to develop a methodology to quantify the load-carrying capacity of corroded reinforced concrete structures. This paper presents a methodology to analyze the mechanical behavior and remaining load- carrying capacity of corroded reinforced concrete structures. The effect of corrosion is modeled as a change in geometry and properties, i.e. reduction of steel area, removal of spalled concrete and modification of bond-slip properties. The modification of bond-slip properties is based on previous research where a detailed 3D solid finite element model has been developed and used to determine the 1D bond-slip response for corroded reinforcement, Lundgren et al. (in progress). The methodology was used in combination with non-linear fracture mechanics in finite element analysis using the program DIANA. The results of the analyses were compared with available experimental results to verify the capability of the proposed method to predict the load-carrying capacity of the corroded structural members. Based on the results of the analyses, the following conclusions can be drawn:

1. The experimental setups for the Coronelli beams had too low corrosion penetrations (≤150 µm) to give significant effect on the load-carrying capacity of the member. 2. For the Coronelli beams, the agreement between the analysis results and the experiments for the presented cases is reasonable in terms of load-carrying capacity, i.e. maximum load, but differences occurred in the post-peak behavior. 3. For the Rodriguez beams, the agreement between the analysis results and the experiments is reasonable.

481 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Field evaluation of dead and live load hanger rod stresses in a continuous steel girder bridge

Stephen Pessiki Department of Civil and Environmental Engineering, Lehigh University, Bethlehem, PA, USA Ian Hodgson Center for Advanced Technology for Large Structural Systems, Lehigh University, Bethlehem, PA, USA

Field testing was performed to evaluate the dead and live load stresses in hanger rods used to support a continuous curved steel two-girder bridge structure that is suspended from the steel multi-girder bridge above from six hangers (three hangers on each girder). Each hanger comprises four 2-1/4 inch diameter ASTM A322 Grade 5140 steel rods. One set of rods has been damaged as a result of a vehicular impact. Vibration testing was used to measure the tension force in each rod. For each test, two uni- axial accelerometers were temporarily clamped to a hanger rod (oriented to measure transverse accelerations). Each rod was struck multiple times at two separate locations with a rubber mallet. The impact from the hammer excited multiple vibration modes (the first three are of interest). High-speed time-history data were recorded during each test, from which the natural frequencies of vibration were extracted. Additional testing with strain gages and a test truck of known weight and geometry was used to evaluate live load stresses. A mock-up of a single rod of identical diameter and length as that of a selected rod on the bridge was constructed in the laboratory. Under constant axial load, the rod was vibrated using a mallet while data were recorded. Tests were conducted at a number of different axial load levels. The results of the laboratory test were used to experimentally establish the load versus natural frequency relationship since the load was measured directly. Finally, finite element analyses of each hanger rod were performed. For each model, a modal analysis was performed with the rod under discreet axial loads (in 20 kip increments). Nonlinear geometrical effects were considered. A load versus natural frequency relationship was developed for each rod and was used to correlate the frequency measurements obtained from field measurements to hanger tension force. This paper describes the tests and analyses that were performed, and compares their results.

482 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Dynamic behaviour of soil-steel road bridge made from corrugated plates

D. Beben Civil Engineering Department, Opole University of Technology, Opole, Poland Z. Manko Institute of Civil Engineering, Wroclaw University of Technology, Wroclaw, Poland

ABSTRACT: The paper presents the results and conclusions of dynamic load tests that were conducted on a road bridge over the Mokrzyca river in Wroclaw (Poland) made of galvanized Cor- rugated Steel Plates (CSP).The critical speed magnitudes, dynamic coefficients, velocity vibration, vibration frequency were determined in the paper. Conclusions drawn from the tests can be most helpful in the assessment of behaviour of this type of corrugated plate bridge with soil. In consid- eration of application of this type of structure in the case of small-to-medium span bridges, the conclusions from the research will not be yet generalized to all types of such solutions. The detailed reference to all type of such bridge structures would be requiring additional analysis (field tests and calculations) on the other types of soil-steel bridges.

The main aim of the this paper is to present the results of the research on the new bridge in the dynamic load test domain of studies as the basis upon which the quality of realiza- tion, durability, critical speed magnitude, dynamic coefficients, velocity vibration, vibration frequency were determine. Due to the important location of the bridge in the road network of Wroclaw and considering its prototypical character as well as the comprehensive and thor- ough research on the structure together with the detailed analysis of the obtained results (from analysis of displacements, strains and dynamic effects), the conclusions derived from this complex study can be very helpful in engineering practice, especially in the control and acceptance field tests which carried out during construction of steel bridges made from corru- gated or flat plates – particularly in bridges of similar geometric structure and similar material characteristics. Based on unit strains at characteristic points and cross sections of the span, as well as deflec- tions obtained at three different points of shell structure the dynamic coefficient values φ were determined for all the load variances I–VI. On their basis, among other things, the critical speed has been determined on the level vcr. = 30 km/h. It was also noticed that the magnitude of the dynamic coefficients were in the limits of 2–18% lower in comparison to values calculated in accordance with the Polish Load Standard (PN-85/S-10030), with exception to one load scheme VI (points: 1, 5, 9, 13, 17, 19, 21, 23, 29, 31 in transversal direction of the bridge). It should be clearly mentioned that the calculated dynamic coefficients were adopted just as the traditional road steel bridges (which in the experimental researches the values were obtained mainly much lower), whereas normal regulations in this range (in Poland as well as Sweden) do not yet relate to the such solutions in a design of the new structures. The above summary and final conclusions refer to a structure of tested bridge of given geometrical characteristics and rigidity of various elements as well as specified effective span. In order to be able to directly apply the results obtained in the research of dynamic load effects on bridges (among

483 others, dynamic coefficient φ, critical speed vcr.) to other types of bridges, additional tests must be conducted on other bridges, consisting mainly of spans with different geometrical longitudinal cross-section, different backfill thickness, different kinds of steels of different span structures as well as different proportions of rigidity of various elements e.g. different types of dimensions, corrugations, and strengths.

484 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Development of live load model using Weigh-In-Motion data

E.-S. Hwang Kyunghee University, Yongin, Korea I.R. Paik Kyungwon University, Seongnam, Korea J.-J. Lee Seoul National University, Seoul, Korea

The live load model used in the bridge design represents the effects of traveling vehicle on the bridge. The current live load model in Korea Bridge Design Code needs to be updated to consider the fast growing truck traffic and weight. Also the reliability-based code requires statistical data of the load model. This paper deals with the development of live load model using data from BWIM (Bridge Weigh-In-Motion) system and portable WIM (Weigh-In-Motion) system. Truck weight data are collected on various sites in the area. They include highways, national roads and provincial roads. Procedures for determining the maximum load effects are determined. From the collected data, maximum weights are estimated for each truck type and sites using the extreme analysis. Only upper 10% of data are used in the estimation and assumed as having extreme type distribution (Fig. 1).

Figure 1. Truck weight data on Gumbel probability paper.

Figure 2. Proposed live load model and multi-lane loading factors.

485 Various factors such as truck types, total weights, headway distances, and correlation of truck types and weights are considered for each lane. The maximum load effects are evaluated for dif- ferent span lengths of the bridge. Multiple presence of trucks in one lane (series of trucks) and two or more lanes(side-by-side trucks) are considered. The probability of multiple presences of trucks is determined from the video recording and other studies. New design live load models for bridges with multi-lane loading factors are proposed for the new LRFD Korea Bridge Deign Code as shown in Figure 2.

486 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Bridge safety analysis considering heavy truck loading

Jiang Du & Da-Jian Han Department of Civil Engineering, South China University of Technology, Guangzhou, Guangdong Province, China

ABSTRACT: Heavy trucks are increasing much rapidly along with the development of economy in china, and the load capacity of trucks is also on the increase, thus nowadays bridges are confronted with challenge to their capacity from the sporadically extremely heavy trucks. Furthermore, the quite serious overloading makes the bridges especially those in the highway deteriorating rapidly and at risk. In order to investigate the safety level of bridges under heavy truck loads, a batch of heavy truck samples passing the exit of the North-Ring Highway in Guangzhou are collected. And the collected data are analyzed using the point process model in extremal statistics. Based on the samples, the generalized pareto distributions are fitted to the gross weight and axle weight data in their tails, and the return levels are estimated and compared with the design standard. The results show that the current operating loads on the analyzed bridges are far above the design standard, and they are still rising. The normal traffic state of vehicles is simplified to a one-truck-passing state for middle and short span bridges, thus the load effects can be easily determined, but obviously the results are somewhat less than the real value. The load effects are analyzed using similar methods to that used in the analysis of truck weight. The generalized pareto distribution is fitted and the return level is estimated and compared with the designed live load effects. The analysis result shows that, the load effects on bridges in the normal traffic state are at a quite high level. The maximum likelihood estimations of return levels of a year of the load effects are below the bridge capacity, but if the confidence interval of the return level at a 95% confidence level is considered, the security is doubtable. The nightly traffic jam state is an extreme case of the dense traffic state and is of special meaning for northing-ring bridges. The generalized extreme value distributions are fitted using the point process model to the load effects samples of truck queues obtained by the Mont-Carlo simulation, and the distributions of maximum load effects in the time horizon are then estimated to assess the failure probability of bridges. The result shows that, the reliability index are far below the allowable value when the traffic jam happens at night, and it would decrease rapidly if the jam lasts for a long time, thus bridges are at huge failure risk. It should be noted that, the analysis results of the vehicle loads, the load effects and the security of bridges are all conservative and reliable due to the following reasons: (1) almost all the fitting result of distribution are a little less than the actual sample data for the sake of estimation steadiness; (2) only one truck is considered and concurrent trucks are neglected in the normal traffic state, and vehicle queues are only laid on two of the three lanes in the nightly traffic jam state.Thus in fact, the security of the bridges would be even worse than the result obtained in the paper.

487 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Behaviors of bracing members in U-type trapezoidal steel box girders

K. Kim & J.H. Park GS Engineering & Construction, Seoul, Korea

This paper investigates mechanics of U-type steel box (tub) girders with top lateral and internal transverse bracing systems. Presented are analytical equations for the estimation of brace forces in lateral bracings and internal transverse bracings. A lateral bracing system is installed at the top flange level in the open top box girder to form a quasi-closed box, thereby increasing the torsional stiffness. Single diagonal and crossed diagonal bracing systems are typically considered for lateral bracing systems. The lateral bracing system in composite tub girders is only required for the construction load. Once the concrete roadway deck is completely hardened, the composite concrete deck takes over the structural function provided by the lateral bracing system. Because of this temporary nature of the bracing system, there is a tendency to take a short-cut on the part of the designers and contractors, thereby experiencing occasional embarrassing experience. Internal transverse bracing (also known as internal cross-frames) are provided in the box and spaced every one or two lateral bracing panels to retain the original box shape, thereby controlling the magnitude of the distortional warping stress level within a prescribed range, usually 5 to 10 percent of the vertical bending stress. Forces are induced in bracing members due to bending as well as torsion of the tub girder. A single diagonal lateral bracing system, currently preferred to by contractors, exhibits significantly different responses from those of a crossed diagonal lateral bracing system. The applied torsional moment is primarily resisted by St. Venant torsion in a girder with a closed cross section. Therefore, stresses (both normal stress and shearing stress) induced due to warping torsion are negligibly small and are not separately considered in the design of the box girder. Since the stiffness of a box section against distortion is relatively weak compared with the bending stiffness or torsional stiffness, the distortional stresses may become fairly large if adequate distortional stiffnesses are not provided. If cross-frames are provided at a proper spacing, the distortional effects can be controlled in a horizontally curved girder. Torsional moments can be decomposed into a pair of couples inducing a net shear flow and distortion of the cross section. Distortion of box girders is caused by the distortional components. The magnitude and distri- bution of the distortional components on box girders are a function of the applied torsional loads as well as the cross-sectional geometry of the boxes. Equations to decompose the force couples causing pure torsional shear flows and distortion of the cross section are given. Equations to pre- dict member forces of cross-frames in trapezoidal box girders have been derived herein. Member forces computed by the proposed predictor equations have been compared with those obtained from three-dimensional finite element analyses and reasonable correlations have been observed.

488 Monitoring and assessment of bridges using novel techniques

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Wireless sensor networks for model based bridge monitoring

S. Deix, M. Ralbovsky & R. Stütz Arsenal research, Vienna, Austria

ABSTRACT: The research area “intelligent sensor networks” is an ongoing field of develop- ment to promote structural health monitoring applications. The idea of tiny smart sensors working together on a common objective to enhance population’s safety and reliability of buildings is an inspiring one. The proposed fields to integrate this technology are as various as scientific topics. But especially in guaranteeing a sustainable and reliable traffic infrastructure the smart-wireless communicating sensors appear to be outmost applicable. In this paper the different application ranges will be well investigated with real time measurements and implementations. The data qual- ity and smart algorithm capability of these systems needs to be evaluated to develop new trustworthy applications and open new implementation strategies in a larger scale. For this purpose wireless sensor test equipment is used in different laboratory scenarios to compare their signals with data from accepted wired sensors. Different communication strategies based on the 2400 MHz ZigBee standard are evaluated to test their usefulness for various tasks. But even the real scenarios on infrastructure sites themselves (bridges, pavements) are tested to analyze the standard behavior of wireless sensor networks.

1 EXECUTIVE SUMMARY

Our bridges are a key factor for the traffic infrastructure and are important for a vital economic development and prosperity of the respective area or nation. Inspection and maintenance is required to ensure safety and reliability of bridges in our traffic networks. The observation of the bridge condition over time is an important aspect in civil engineering. Usually bridges become visually inspected in regular time intervals to detect any defects and to decide whether additional rehabil- itation might become necessary to maintain structural condition. Due to the fact that large parts of European bridges have been built 30 to 40 years ago the costs for rehabilitation are nowadays steeply increasing. Wireless sensor networks have the potential to change the perception of struc- tural health monitoring due to a lot of advantages. Decreasing hardware costs enable disperse deployment and ubiquitous intelligent sensors can accomplish various tasks. But the enabling tech- nology needs to meet the requirement of structural monitoring. Hence, a specific level on sensor sensitivity and available sampling rate is needed to deduce significant objective parameters. Model based approaches are needed to reduce the data transmission and energy consumption on the node. According to the fast development and emerging technologies of the (sensor) hardware market the compliance with the requirements is only a question of time.

491 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Structural health monitoring and passive vibration control of an Austrian road bridge

M. Reiterer Department of Structural Dynamics, BERNARD Engineers, Vienna, Austria L. Praxmarer Department of Civil Engineering, BERNARD Engineers, Tyrol, Austria

ABSTRACT: The Murbridge in Austria was erected in 1964 and scheduled to be demolished and rebuilt because of its poor condition. The construction consists of a prestressed concrete box-girder with a main span of 60m and additional cantilevers of 14,50m on either side. Both ends of the superstructure are attached to the basis of the pillars by prestressed tie bars. They are not visible because they are covered with soil and therefore increase the risk of failure. Shortly after completion a large deflection in midspan occurred. Additionally over the years the deformations increased and in 1977 cracks on the bottom side of the midspan appeared. In 1995 leveling showed a maximum deflection difference from the nominal value of about 17 cm. Also, a slight inclination of 4‰ and 6‰, respectively of the pillars was measured. Additionally, as the asphalt was damaged periodically over the years, a depression emerged on both sides of the bridge. On this account the permitted maximum live load was reduced to 16 tons. The engineering company BERNARD Engineers was contracted by the State of Styria to look for the most economical method for bridge maintenance. The service life of the bridge should be extended for an additional 25–30 years. If possible, the reduction of the permitted live load should be reversed. A detailed static and dynamic analysis was performed and pointed out that a horizontal movement of the foundation is likely responsible for the deformations. The shear forces at the foundation level had not initially been taken into account in the structural analysis. Hence, an inclination of the pillars occurred and caused the large vertical deflection in the center of the midspan. In addition to the performed static and dynamic FE-analysis, a permanent structural health monitoring system was installed to monitor the bridge. At first the results of the measurements were used to calibrate the numerical model of the bridge and afterwards, a static and dynamic proof load test was performed. The results of the load test indicated, that the stiffness of the bridge was underestimated within the performed numerical simulations. In view of the permanent operation of the structural health monitoring system a digital remote maintenance system was installed and a special emergency plan was developed. The plan identifies the responsibilities and actions in case of critical bridge states. In addition the dynamic proof load test indicated that the Murbridge is very vibration prone. The damping coefficient of the Murbridge was found to be about 0.8%. Hence, it was decided to attach an innovative passive damping device, the so calledTuned Liquid Column Dampers (TLCD). Numerical simulations show, that due to the installation of the TLCD a increasing of the bridge damping up to 1.6% is achieved. It is concluded that a basic reconstruction of the Murbridge in combination with a perma- nent installed structural health monitoring system and damping devices is more economical than demolishing the bridge and built it new.

492 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Recent Austrian activities in bridge monitoring

R. Geier Schimetta Consult ZT GmbH, Vienna, Austria

EXTENDED ABSTRACT: To offer reliable bridge monitoring it is required to combine know- how in investigation, assessment and rehabilitation of structures. Besides the conventional structural inspection, usually static and/or dynamic measurements and advanced identification algorithm are employed. Monitoring definitely offers an enormous potential for observation and assessment of existing structures. The key word in this context for future monitoring concepts will be integration of different methods to obtain a maximum of information concerning structural condition. As each specific monitoring technique offers advantages and at the same time has some disadvantages, it is important to combine existing techniques in such way, that the advantages supplement each other. Several projects have shown that these technologies are powerful to solve our future problems in maintenance and assessment of bridges. In this context the paper should give an overview concerning some selected Austrian activities in the field of bridge monitoring. In the given examples it was tried to combine different approaches and technologies in a way to obtain the information required for decision making. The elaboration will focus to three major examples of use successfully realized during the last two years. The first aspect deals with experiences gained for the development of a remote system for cable force measurements combined with a semi-active damping device for bridge stay cables. A second section will be devoted to the practical experiences gained for assessment of railway bridges. These tasks are devoted to permit higher train speeds on bridges considering resonance effects. An innovative calculation approach in combination with field measurements will be presented. A short section is devoted to a classical monitoring system which is currently realized for an integral abutment bridge in Upper Austria. Finally it is very important to mention the key sentence once again: in civil engineering we always have to deal with very complex structures. Therefore it is required to be open-minded and to combine different monitoring technologies to obtain reliable results.

493 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Computational model updating for bridge maintenance planning

S. Deix, M. Ralbovsky & H. Friedl Arsenal research, Vienna, Austria

ABSTRACT: An ongoing topic of research and active field of development is the usage of computational model updating for damage localization methods. Applying system identification algorithms and modal analysis on vibration measurement data reveals physical and dynamic behav- ior of structures. Natural frequencies and mode shape functions are directly affected by changes in stiffness or mass of solid structures.An accepted method to identify structural changes is fine-tuning Finite Element models (FE) by varying certain variables. This process is actually an optimization problem. Therefore modern model updating methods use optimization algorithms to change mass or stiffness matrices of finite element models. With these procedures an area of reduced stiffness can then be located and identified as damage. At the moment the tasks for computational model updating routines can not be considered to function by automatic means. The choice of correction parameters is important for the success of the updating procedure. Other influencing factors are the quality of the measured signals and the propagated uncertainty in this process. Due to temperature dependency and other undesired effects the special temperature models can aid to filter the distur- bance. The accomplished work will be reported and will mainly deal with different computational updating routines as well as various case studies conducted in Austria. For instance one algorithm is based on an iterative sensitivity matrix based function. In this the objective function consists of weighted vector consisting of natural frequencies and mode shape information. Another algorithm is based on the modal force residuals. The task will be to identify and localize artificial applied damage.

1 EXECUTIVE SUMMARY

Damage localization by means of vibration measurements is the scientific ambition of several national projects in Austria. In one case a steel structure with a length of 8 m has been tested (steel construction test) and another test was conducted with a 5m concrete beam. Both measurements and corresponding data will be used to show weakness and strengths of proposed damage localization methods. In particular the modal parameters eigenfrequencies, eigenmodes were used for model updating. The results of the FE-models are adapted to the actual modal data by an optimization process (Model Updating). The results from the concrete beam test confirm the applicability of vibration-based damage detection on reinforced concrete beams. The excellent agreement between the observed cracks and identified stiffness reductions was possible due to high quality of acquired measurement data. The results from the steel construction show that different FE-models with varying boundary conditions can produce varying detection results. The damage identifications with further damage states (bigger changes in modal parameters) show a better correlation. Both updating methods (method 1 and method 2) produce similar results, due to the same input param- eters. Although the differences in residual formulation results in more or less stable and accurate damage detection.

494 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

AIFIT – user orientated identification for infrastructure, theory

R. Wendner, S. Hoffmann, A. Strauss & K. Bergmeister University of Natural Resources and Applied Life Sciences, Vienna, Austria

ABSTRACT: Reinforced concrete structures such as bridges do not only have to be built, but must also be maintained to preserve their operational reliability. It is most essential for cost minimized maintenance to identify damage as soon as possible. Traditionally used visual inspection methods are very time-consuming and costly. Alternatively global parameters including their changes can be used to assess a structure’s condition. Aim of the presented work done within project AIFIT is the development of robust identification methods suitable for practical use. This paper shows different approaches considered, ranging from quasi-static techniques such as bending lines and influence lines to several methods based on modal data. Following the theory and requirements all techniques are applied to laboratory tests. For this nine 2-span bars were concreted, damaged stepwise and several measurements performed to gain the necessary input data. Finally the results of all methods are compared and verified against each other. For most developed countries infrastructure challenges will shift from building new infrastructure to the maintenance of existing structures. This will demand detailed knowledge of the structure’s state not only to guarantee sufficient safety of the structures, but to have a most economical maintenance planning and assignment of funds. Global identification methods applied to structures can offer an attractive method for the estimation of a structure’s health status, including rough quantification and localization of damages, expressed by local stiffness changes. The main goal of project AIFIT is the development of user orientated identification systems for engineering structures which should be highly automated, robust, largely unbiased and necessitate as little as possible engineering efforts. A total of six different approaches for the determination of stiffness distributions based on global data were developed and/or used for verification purposes in course of this project. Apart from DELFI an algorithm based on static deflection lines at different load levels and ILIAS working with influence lines for bearing reactions of hyperstatic systems the other four approaches use modal characteristics – eigenfrequencies and eigenformen as input data. MOBEL uses mode shape data with high spatial resolution to determine inertia forces, then calculates a pseudo static bending line and fits it to the original mode shape. STRIDE is a sensitivity factor based model update in this case using eigenfrequencies and amplitude information in 2 points only although the tool actually is able to work with all types of monitoring data as long as they can be modeled. Laboratory tests conducted at our institute confirm that all tools are capable of identifying and to a certain extent quantifying damage. Although there are some system-immanent differences the results of the different approaches correspond pretty well to each other and the observed crack pattern and most of them comply with the demands of user orientation and robustness. The application to a real structure as shown in the contribution “AIFIT – user orientated identification for infrastructure, application” of these proceedings will finally determine the benefits, demands and practical feasibility of all tools.

495 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

AIFIT – user orientated identification for infrastructure, application

S. Hoffmann, R. Wendner & K. Bergmeister University of Natural Resources and Applied Life Sciences, Vienna, Austria M. Mautner ÖBB-Infrastructure Construction Corporation, Vienna, Austria W. Steinhauser ZT Steinhauser consulting engineers, Vienna, Austria

EXTENDED ABSTRACT: For most developed countries infrastructure challenges will shift more and more from building new infrastructure to the maintenance of existing structures. This will demand detailed knowledge of the structure’s state not only to guarantee a sufficient safety of the structures, but to have a most economical maintenance planning and assignment of funds. Inspection methods should allow the identification at a very early damage state to enable the best preventive actions. Especially for concrete structures damage identification and localization is mostly limited to damages visible at the surfaces of the structure. Several attractive method for estimations of the health status, including rough quantification and localization of damages, expressed by local stiffness changes are under development (compare the contribution “AIFIT – user orientated identification for infrastructure, theory”). The research project “AIFIT” aims for the enhancement of well-known or new approaches to make them more user-orientated. Laboratory tests generated data for new methods based on directly measured influence lines and additional quasi-static data, as well as modal data. The stiffness distributions identified by different methods showed their sensitivity for the localization of damages resulting in bending cracks. A field tests demonstrated less, but feasible sensitivity by stiffness loss of app. 13% resulting from a well visible crack pattern on the bottom side of the middle span. Not accessible areas like above the piers, where cracks from bending will occur on the upper side hidden by the surfacing and asphalt, were assessed. One major claim of the project AIFIT was fulfilled by transferring the experience gained in the laboratory test to the field application and conducting a full stiffness analysis of a real bridge structure with feasible effort. Further analysis and implementation of reliability assess- ments using the identified characteristics will be future tasks of this project. Furthermore detailed tests will be conducted on the instrumented bearings on the abutment of the considered bridge. In addition a significant influence of cracks on structural stiffness was shown, which challenges existing codes especially regarding the accounted deformation.

REFERENCES

Hoffmann, S., Wendner, R. & Strauss, A. 2007b. Comparison of Stiffness Identification Methods for Reinforced Concrete Structures, 6th International Workshop on Structural Health Monitoring, September 11–13, Stanford, USA, pp. 354. Ralbovský, M. 2007. Damage Detection in Concrete Structures Using Modal Force Residual Method, Phd thesis at the Slovak University of Technology in Bratislava.

496 Strauss, A., Bergmeister, K., Lehky, D. & Novak, D. 2006. Inverse statistical nonlinear FEM analysis of concrete structures, in Meschke, G.; de Borst, R.; Mang, H.; Bicanic, N.:: Euro-c 2006. Computational Modelling of Concrete Structures, March 27–30, Mayrhofen, Tirol, Austria, p. 897. Wendner, R., Strauss, A., Hoffmann, S. & Bergmeister, K. 2007. Novel identification methods for the assessment of engineering structures, in Luc Taerwe, Dirk Proske (Eds.), 5th International Probabilistic Workshop, Universiteit Gent, BOKU, ESRA, 5th International Probabilistic Workshop, November 28–29, Ghent, Belgium, pp. 135.

497 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Stochastic nonlinear finite element analysis of bridges

R. Pukl Cervenkaˇ Consulting, Prague, Czech Republic M. Voˇrechovský & D. Novák Brno University of Technology, Czech Republic

ABSTRACT: Stochastic finite element method is a methodology, which accounts for uncertain- ties and spatial variability in structural properties. Spatial variability of material properties is an important phenomenon, which should be accounted for in realistic computer simulation of structural response behavior and damage. Representation of spatial variability of structural properties is made by random fields, which is a high level of its modeling with a scientific background. Methodology for introducing spatial variability of material properties into a nonlinear finite element program system is presented together with a brief description of selected applications.

SUMMARY

The nonlinear finite element analysis employs advanced constitutive models for concrete based on damage mechanics, nonlinear fracture mechanics and plasticity theories with smeared crack approach. It is a proven tool for computer simulation of reinforced structures including failure mechanism and post-peak behavior. It was extended by encapsulating of the existing material models to incorporate spatial variability of material properties. For the time-intensive calculations like nonlinear fracture mechanics of concrete, the small- sample simulation techniques based on stratified sampling of Monte Carlo type represent a rational compromise between feasibility and accuracy.Therefore, Latin Hypercube Sampling method (LHS) was selected as a key fundamental technique. The representation of spatial variability of structural properties by random fields is a high level of their modeling with a scientific background. Random fields describe the spatial distribution of a structural (material) property over the region representing the structure based on the prescribed correlation length and the autocorrelation function. Application of the proposed methodology using random fields is illustrated on two examples: development of cracks in four-point bending beam, and reliability analysis of a real highway bridge.

CONCLUSIONS

The presented technology enables to model inhomogeneities in the nonlinear finite element solu- tion. It allows accounting for spatial variability of material properties in the nonlinear finite element framework. Random occurrence of structural damage (cracks) can be simulated even in a homogeneous stress state. These features are in particular important in realistic analysis of resistance of concrete bridges. The numerical results can be directly statistically evaluated and used for safety and reliability assessment.

498 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Overview of 40 bridge monitoring projects using fiber optic sensors

Daniele Inaudi & Branko Gliši´c SMARTEC SA, Switzerland

ABSTRACT: In the last 15 years, fiber optic sensing has become a useful and increasingly widely used tool for structural health monitoring of bridges and other civil structures. We have been fortu- nate to participate to their development and introduction to real field applications. This paper is an overview of 40 bridge monitoring projects carried out over the last 15 years in 13 different countries. In particular we concentrate on the analysis of the different types of bridges that were monitored, their situation (new construction, existing structure, refurbishment…) and the main purpose of the installed monitoring system. Two main categories emerge form this analysis: new bridges with innovative aspects or particular relevance and existing bridges with known deficiencies. We will than analyze the most typical sensor architectures used to monitor those bridges, giving statistics on the number and type of installed sensors, their survival rate and the duration of the monitoring project. For some projects, the monitoring concentrated on a specific event in the life of the bridge, for example its construction or refurbishment. Other projects aimed to long-term monitoring and are still running, in some cases after almost 10 years. In the early applications of fiber optic technology the cost of the measurement instruments and their fragility were discouraging permanent monitoring and many projects called for periodic manual measurements. In the last 5 years, significant progresses in the instrument reliability and lower cost have enabled a growing number of permanent instrumentations. Another interesting aspect concerns the entities involved in each monitoring project. We will analyze weather the project originated form a research entity (university, research institute) or was driven more by the bridge owners. Finally we will summarize the main findings of each project and show concrete examples of actionable information that could be gained thanks to monitoring. This overview on the monitoring of 40 bridges with optical fiber sensors shows how diverse and multi-faced this domain can be. The projects include everything form a simple short-term test with a couple of sensors to verify a design hypotheses to a large-scale instrumentation project with hundreds of sensors to extend the lifetime of a bridge with known problems. After an initial phase where many projects were driven by the curiosity of both universities and owners towards a new technology, we have now moved to applications where the customer wants to address a specific question or increase safety in the case of known deficiencies or degradations. Those projects show that a well planned and executed monitoring projects can provide actionable information to the owner and the bridge engineer.

499 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Reliability assessment of an existing bridge using long-term monitoring

Alfred Strauss, Dan M. Frangopol & Sunyong Kim Lehigh University, Bethlehem, PA, USA

Numerous monitoring and maintenance concepts for engineering structures have been developed during the last decades (Bergmeister and Santa 2000; Frangopol and Messervey 2007). Most of these concepts originate from the demand to capture uncertainties in environmental and mechanical conditions, material properties, and loading history. The overall aim is to support cost-optimized intervention planning of deteriorating civil infrastructure systems. Optimization strategies, numer- ical reliability assessment methods, degradation models and cost-optimized intervention strategies are considered for the maintenance and operation of highway networks. Despite these advancements there is still a need for the efficient incorporation of monitored data in structural assessment and prediction models. For instance, the updating of prediction models by monitored data, at present underutilized, can have a serious influence on cost based intervention planning. Monitoring strategies for engineering structures have been subjected to a rapid development during the last decades. In general, these strategies should provide important information for intervention planning (e.g., maintenance, repair, rehabilitation, replacement) of new and existing structures. However, the reliability assessment and prediction of structural performance, based on monitoring strategies, still require a sophisticated treatment of structural monitoring data to allow a cost optimized maintenance. Statistical quality assurance and acceptance sampling provide the necessary basis for such monitoring data treatments. The objectives of this paper are to discuss (a) the incorporation of monitored data in the structural reliability assessment process, and (b) the use of monitoring data for improvement of structural performance prediction models. The proposed approaches are demonstrated on an existing highway bridge. The monitoring program of this bridge is provided in Connor and Santosuosso (2002), and Connor and McCarthy (2006). The paper also provides a very brief summary of this program, including the bridge description. Both yield strength performance and fatigue performance are assessed based on long-term monitoring data.

REFERENCES

Bergmeister, K. & Santa, U. 2000. “Global monitoring concepts for bridges.” Nondestructive evaluation of highways, utilities, and pipelines IV,Society of Photo-Optical Instrumentation Engineers, 3995, 14–25. Connor, R.J. & McCarthy, J.R. 2006. “Report on Field Measurements and Uncontrolled Load Testing of the Lehigh River Bridge (SR-33).” Lehigh University’s Center for Advanced Technology for Large Structural Systems (ATLSS), ATLSS Phase II Final Report. Connor, R.J. & Santosuosso, B.J. 2002. “Field Measurements and Controlled Load Testing on the Lehigh River Bridge (SR-33).” Lehigh University’s Center for Advanced Technology for Large Structural Systems (ATLSS), ATLSS Report 02-07. Frangopol, D.M., & Messervey, T. 2007. “Integrated life-cycle health monitoring, maintenance, management and cost of civil infrastructure,” Proceedings of the International Symposium on Integrated Life-Cycle Design and Management of Infrastructures, Lichu, F., Limin, S., and Zhi, S., eds., Tongji University Press, Shanghai, 216–218; full 12 page paper on CD-ROM.(keynote paper).

500 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Evaluation of the redundancy of bridge superstructures and substructures

Michel Ghosn The City College of New York,/CUNY, New York, USA Dan M. Frangopol Department of Civil & Environmental Engineering, Center for Advanced Technology for Large Structural Systems (ATLSS), Lehigh University, Bethlehem, USA

In an effort to reduce the risk to civil infrastructure from man-made and natural hazards, researchers have recently renewed their interest in developing methods for evaluating the redundancy of struc- tural systems and recommending the design of more robust structures. For highway bridges, these efforts concentrated on both superstructures and substructures (Frangopol and Nakib; 1991, Ghosn and Moses; 1992, Ghosn & Moses; 1998, Liu et al.; 2000). This paper provides a very brief back- ground on the contributions made by the authors and their co-workers on these topics with emphasis on bridge substructures. In fact, for most bridges, the substructure represents the most vulnerable system and is associated with the highest level of risk. A bridge substructure is a highly complex system that includes structural elements such as bents and columns, piles or footings, as well as the supporting soil. A common method to account for the uncertainties in estimating the safety of a structural system or member is to use the reliability index, β, which is inversely related to the probability of failure. Since redundancy is defined as the capability of a system to continue to carry load after the failure of its most critical member, the redundancy of a system can be related to the relative or incremental reliability index, βu, which is defined as the difference between the reliability indices of the system, βult, and that of the most critical member, βmember. Thus, a substructure system will provide adequate levels of system redundancy if the relative reliability index is adequate. For the practical implementation of the results of the reliability analysis, this paper illustrates how to calibrate a set of system factors that can be included in the LRFD design equations of substructure systems to improve the safety of non-redundant bridge substructures and allow for reduced safety levels for redundant configurations.

REFERENCES

Frangopol, D.M., and Nakib, R. 1991. Redundancy in highway bridges. Engineering Journal. American Institute of Steel Construction. 28(1), 45–50. Ghosn, M. and Frangopol, D.M. 2006. Redundancy of structures: A retrospective, Proceedings of the 13th WG 7.5 Working Conference on Reliability and Optimization of Structural Systems, Kobe, Japan. October 13–15, 2006; in Reliability and Optimization of Structural Systems: Assessment, design, and life-cycle performance, Frangopol, D.M., Kawatani, M., and C-W. Kim, eds., Taylor & Francis Group plc, London, 2007, 91–100. Ghosn, M. & Moses, F. 1998. Redundancy in Highway Bridge Superstructures, NCHRP Report 406, Transportation Research Board, Washington DC. Liu, D., Ghosn, M. & Moses, F. 2000. Redundancy in Highway Bridge Substructures, NCHRP Report 448, Transportation Research Board, Washington DC.

501 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Damage detection by pattern recognition at bridge components

H. Wenzel & R. Veit-Egerer VCE Holding GmbH, Vienna, Austria

ABSTRACT: Bridge components, structural and non-structural parts, earn more attention with damaged observed and failures occurred recently. Methodologies to check and assess the structural integrity of components have become essential. This contribution deals with experience made at the following elements of bridges: (a) Torsional bracings of a steel bridge, (b) Overhead traffic signs, (c) Lighting poles, (d) Handtrails, (d) Inspection equipment. For all of these components failures have been found. Most of them are related to traffic accidents in case that they are exposed to traffic on the surface of a bridge. Nevertheless cases of fatigue failures also occurred on elements of less than 5 years of age. As components of bridges are elements of less value huge campaigns are monetarily not justified. Quick and easy methodologies are required. This contribution describes the results of a non-linear vibration analysis that clearly indicates damage. The method is based on the well known non-linear distribution of frequencies depending on the impact. As bridges are loaded unequally during the monitoring procedure these non-linear effects become visible.

1 INTRODUCTION

Health monitoring for civil engineering structures is a challenge. Our structures are a prototype each and show small safety margins and a great exposure to the public. Bridges for example were the backbone of powerful empires from China to Rome and the Incas in America. Currently the transportation infrastructure is directly related to economic success of a nation. All these facts make structural health monitoring of civil engineering structures more difficult than any monitoring of a well defined mechanical structure. The major number of uncertainties in geometry, material properties and the influence of the environment might have a higher impact on monitoring results than any minor damage. Therefore only complex approaches under consideration and compensation of the already known phenomena will be successful. In civil engineering the procedure and tools are best developed for bridges. Some kind of structural health monitoring always existed in this sector. The following figure shows how these procedures have developed from simple inspection routines to highly sophisticated monitoring campaigns.

REFERENCES

Forstner E., and H. Wenzel, 2004. “IMAC – Integrated Monitoring and Assessment of Cables, Final Technical Report”, IMAC Project. Wenzel, H., and D. Pichler. 2005. “Ambient Vibration Monitoring,” J. Wiley and Sons Ltd., Chichester – England, ISBN 0470024305. Veit R., H. Wenzel and J. Fink. 2005. “Measurement data based lifetime-estimation of the Europabrücke due to traffic loading – a three level approach”, In International Conference of the International Institute of Welding. Prague. 502 Peeters, B.; De Roeck, G.: “One year monitoring of the z 24-bridge: environmental influences versus damage events.” In Proceedings of IMAC 18, The International Modal Analysis Conference, pp. 1570–1576, San Antonio, Texas, USA, February 2000. De Roeck, G.; Peeters, B.; Maeck, J.: “Dynamic monitoring of civil engineering structures.” Computational Methods for Shell and Spatial Structures IASS-IACM 2000, Greece, 2000.

503 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Dynamic damage identification of Colle Isarco viaduct

D. Lehký, D. Novák & P. Frantík Faculty of Civil Engineering, Brno University of Technology, Brno, Czech Republic A. Strauss & K. Bergmeister Department for Civil Engineering and Natural Hazards, University of Natural Resources and Applied Life Sciences, Vienna, Austria

The paper is focused on damage identification of dynamically loaded structures using two alterna- tive methods: artificial neural network and sensitivity based methods. The damage and variability of material properties (stiffness) along the structure are studied. The modal properties of a struc- ture (eigenfrequencies and modeshapes) are used as input parameters for identification. Parametric study of their applicability for damage identification was carried out. The neural network iden- tification method is based on coupling of artificial neural network (multilayer perceptron) and stochastic analysis of structure for preparation of appropriate training set. The sensitivity based method is based on an iterative change of input parameters depending on their sensitivity with respect to structural response to get numerically as close as possible to experimental measurements. Proposed methodologies are verified using Colle Isarco viaduct. Continuous health-monitoring of structures (bridges) is essential part of its maintenance. There- fore damage localization and its level is the subject of research of both academic and industrial research groups during last decade. Non-destructive testing – vibration measurements to get modal data (eigenfrequencies, modeshapes) is the most promising technique as it can be performed using a structure in usage. The task is based on the fact that damaged structure has smaller stiffness in some parts – and this difference will affect vibration (modal data). The comparison of vibration of virgin (undamaged) structure and damaged structure can be used for the detection of damaged parts (localization of damage). The aim of the paper is to show efficiency and possibilities of damage identification using artificial neural network based approach (Lehký et al., 2007) and sensitivity based approach (Strauss et al., 2007). Integral part of research is the study of modal properties and analysis of its applicability for damage identification. Both techniques are briefly described. But main aim of the paper is to show theirs efficiency for damage identification using case study: Colle Isarco viaduct.

REFERENCES

Lehký, D., Novák, D., Frantík, P., Strauss, A., Bergmeister, K. 2007. Dynamic damage identification based on artificial neural networks, SARA – part IV. The 3rd International Conference on Structural Health Monitoring of Intelligent Infrastructure, Vancouver, British Columbia, Canada, 183. Strauss, A., Bergmeister, K., Wendner, R., Novák, D., Lehký, D. 2007. Sensitivity factor based dynamic damage identification, SARA – part III. The 3rd International Conference on Structural Health Monitoring of Intelligent Infrastructure, Vancouver, British Columbia, Canada, 186.

504 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Virtual testing of bridges for life cycle reliability assessment

R. Pukl & V. Cervenkaˇ Cervenkaˇ Consulting, Prague, Czech Republic B. Teplý & D. Novák Brno University of Technology, Czech Republic K. Bergmeister University of Natural Resources and Applied Life Sciences, Vienna, Austria

ABSTRACT: Tools for life cycle analysis should capture degradation and retrofitting processes in order to support maintenance of bridges. The presented system employs advanced computer modeling based non-linear FEM – so called “virtual testing” of structures.

1 VIRTUAL TESTING OF BRIDGES

Virtual testing of structures provides better insight into the structural properties and resistance and enables realistic assessment of the structure. For realistic life cycle reliability analysis it incorporates a suitable stochastic approach and efficient tools for modeling of degradation.

2 RELIABILITY ASSESSMENT BASED ON NONLINEAR SIMULATION

SARA is a complex software system, which combines efficient techniques for nonlinear numer- ical analysis of engineering structures, probabilistic methods and degradation assessment. The structural resistance is modeled by nonlinear finite element program ATENA. Using appropriate input parameters, it can analyze the structure in intact or deteriorated conditions. Stochastic engine FReET utilizes an advanced statistical technique in order to achieve sufficiently accurate results keeping the number of simulations acceptable small. Input parameters for nonlinear analyses are generated randomly; structural safety (reliability index) is evaluated from the calculated resistance. Reduced material parameters or their statistic for deteriorated structure are obtained from FReET-D, a complex software tool for degradation assessment.

3 MODELING OF DEGRADATION

FReET-Dprovides stochastic modeling of degradation phenomena in concrete structures, statistical and sensitivity analyses, assessment of service life and assessment of reliability measures. The degradation models for carbonation, chloride ingress and reinforcement corrosion are incorporated in form of a library of functions.

4 CONCLUSIONS

Nonlinearities in material, damage, degradation and durability aspects are employed in virtual testing of the particular bridge structure, including uncertainties and randomness involved in all 505 these phenomena. Presented approach is applicable in sense of performance-based design for the assessment of service life (for newly designed bridges) or residual service life (for existing bridges) and relevant safety and reliability levels. The consideration of input parameters as random variables is essential for realistic results enabling a reasonable decision-making.

506 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Degradation modelling of bridge components based on cellular automata

D. Novák, B. Teplý, J. Podroužek & M. Chromá Faculty of Civil Engineering, Brno University of Technology, Brno, Czech Republic A. Strauss IKI BOKU, Vienna, Austria

The paper presents cellular automata approach to the problem of lifetime assessment of concrete structures, particularly bridges, under diffusive attack from external aggressive agents. The dif- fusion process is modeled by cellular automata approach. The effectiveness of the methodology is supposed in the combination with nonlinear analysis of concrete structures. Cellular automata can simulate degradation of material as a 2D task. A consequent step after diffusion modeling is the utilization of results from cellular automaton approach for a particular stochastic degradation model to capture the corrosion of reinforcement. The whole methodology is described with respect to application within the framework of SARA software tools, namely nonlinear finite element calculation ATENA, reliability modelling FREET and degradation tool FREET-D. Application for a particular bridge is provided. Figure 1 shows the heuristically estimated initial distribution of concentrations (half of the bridge), maximal surface chloride concentration is 0.2% Cl- per cement content. Predictions of chloride concentrations for 80 years along the bridge are shown in Figure 1. The light blue colour represents the undamaged state of a cell, while the darker colour represents the degraded state. Black regions represent parts where chloride concentrations were greater than critical chloride concentrations (here 0.4% Cl- per cement content was considered).

Figure 1. Predictions of chloride concentrations for 80 years.

507 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Stochastic aging model for infrastructure buildings

Markus Petschacher Petschacher Consulting, Feldkirchen/Feldkirchen, Carinthia, Austria

Information about the durability of constructions under the impact of traffic and other conditions is highly important regarding the planning of maintenance work and budget. This study investigated bridges on Austrian roads regarding their future behaviour. Data source for this project was the building database Austria, which holds time variant and invariant information on all constructions like noise barriers, tunnels or bridges. Life-Cycle-CostAnalysis is the evaluation of investment costs for initial construction, operation, maintenance, renewal and sometimes disposal of a construction over a chosen period. The LCC- model consists of 3 levels in which the costs are divided by life-cycle, cost object and components. The typically 8 components of a bridge do not exist in every case, so using a rigid component scheme is not always possible. This leads to the definition of 3 meta-components, to which the components are assigned to and which permits a multidimensional state description of a bridge. In order to obtain the best state description, an approach called “Best Scalarization” was used, in order to convert the multidimensional description into a scalar value. For this bridges were divided in two groups, which were transformed into a scalar value with a linear transformation. As mentioned before, the data source was the building database Austria, where all relevant data – especially “year of construction” and “year of last maintenance” – were sorted and stored. The data analysis began with the formation of bridge sets with similar behaviour patters. To optimize the reliability of a network and to reduce maintenance costs it is necessary to create an aging model, which allows the user to predict future situations within a system. Core of the analysis was to build up a warning level, normally bridges in condition class 4 are taken into such a level. So the model must be able to predict how long it will take, until such a bridge reaches condition class 5 and fails. The knowledge of the tendentious aggravation of the asset and the renewal rate form the basis for the generation of a higher network quality. In principle the aging model or condition decline of bridges bases on the assumption of probabilities. When data are interpreted and used as basis for system forecasts, calculations of probabilities are used; this study uses the Cohort Survival Model. The CSM can easily be applied on infrastructure, because the life-cycle process for humans and bridges can be seen as similarly processing. For the simulation the bridge model – 3 meta-components – was combined with the CSM. In order to model the aging behaviour of the bridges, the Herz-distribution was used. The simulation model offers different kinds of strategies, starting from a “None”-Strategy and ending at different combined budgetary scenarios. The simulation was carried out for a period of 50 years and looked at both the global stock and used different budgetary and non-budgetary scenarios. As result of the study it can be said, that the tendency of the condition development was portrayed in a good way and can hence be used as decision basis. The question of an optimal strategy could only be discussed in outlines. This solution is strongly linked to the formulation of an utility function, which is the basis for global optimization. The utility function will depend on the point of view of its creator, i.e. utility is different for administration, end user and community. With the model it could be shown that underfinancing will not cause important deteriorations in the near future. But the longer the simulation lasts, the worse the condition will become. In such a sit- uation the budget is often not the only edge condition – also the number of constructions sites, which lead to handicaps for the community will be highly important and can thus be seen as edge condition. 508 New developments in large-scale model studies of bridge components and systems subjected to earthquakes

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Hybrid distributed simulation of a bridge-foundation-soil interacting system

A.S. Elnashai, B.F. Spencer, S.J. Kim, C.J. Holub & O.S. Kwon University of Illinois at Urbana-Champaign, Urbana, Illinois, USA

ABSTRACT: In this paper, a hybrid-distributed simulation that involves large scale reinforced concrete physical pier specimens, computer models of the bridge deck and soil analysis software is described. This hybrid-distributed simulation was one of the earliest to utilize the George E. Brown, Jr. Network for Earthquake Engineering Simulation (NEES) including collaborating part- ners at the University of Illinois at Urbana-Champaign (UIUC), Lehigh University (Lehigh), and Rensselaer Polytechnic Institute (RPI). The bridge structure under investigation was partitioned into five separate modules which were either experimentally or analytically loaded at one of the three participating NEES equipment sites (Fig. 1). The simulation successfully produced pier shear failures similar to those observed in the prototype structure, Collector-Distributor 36 of the Santa Monica (I10) Freeway which was damaged during the 1994 Northridge Earthquake. Moreover, the redistribution of forces between the two sites with the bridge piers as either of the two suffered partial failure shows that full interaction was taking place between the geographically distant sites. Thus, the simulation provides insight into the behavior and design of shear critical bridge piers and showcases the ability of NEES facilities to investigate complex failures in large and complicated soil-structure-foundation systems.

Figure 1. Substructuring of the bridge structure and distribution to participating sites.

511 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Research and application of precast segmental bridge columns for seismic regions

K.-C. Chang & M.-S. Tsai Department of Civil Engineering, National Taiwan University, Taipei, Taiwan Y.-C. Ou & G.C. Lee Department of Civil, Structural and Environmental Engineering, University at Buffalo, Buffalo, NY, USA J.-C. Wang Runhorn Pretech Engineering Co., LTD., Taipei, Taiwan P.-H.Wang National Center for Research on Earthquake Engineering, Taipei, Taiwan

In this paper, two research projects on segmental bridge columns for seismic regions are reviewed. Both were carried out at the National Center for Research on Earthquake Engineering (NCREE) in Taiwan. In the first research project, new types of segmental bridge columns with enhanced hysteretic energy dissipation capacity and lateral strength were developed. This was accomplished by adding mild steel reinforcing bars (also referred to as Energy Dissipation (ED) bars) across the segment joints, strengthening the joint at the base of the column and increasing the height of the hinge segment. Four large-scale column specimens were designed and tested. Test results of specimens with the proposed design concepts showed good ductile behavior, and increased energy dissipation capacity and lateral strength. The second research project started in 2005 and is in progress. The research is a joint venture between NCREE and the Multidisciplinary Center for Earthquake Engineering Research (MCEER) in the U.S. The proposed segmental columns in this research can be categorized into three types. The first type has no mild steel bars continuous across the segment joints. ED bars are provided in the other two types of column to increase the columns’ capacity in terms of hysteretic energy dissipation and lateral strength. In the second type, the ED bar ratio is proportioned to the amount of the total axial force to achieve satisfactory energy dissipation capacity while keeping the residual displacement small. For the third type, the ED bar ratio is further increased so that the resulting hysteretic energy dissipation capacity emulates that of a conventional column. Large-scale column specimens with the above characteristics were designed and tested with lateral cyclic loading and pseudo-dynamic loading. Under cyclic loading, the specimens exhibited satisfactory ductile behavior. Three types of hysteretic behavior were observed. Results of pseudo-dynamic testing showed that the specimens were capable of sustaining two successive design earthquakes with little strength degradation and surviving one maximum probable earthquake. Large-scale testing on the first type of column with seismic isolation systems will be carried out in 2008. Towards the end of this paper, an application of precast segmental columns in a highway to be constructed in a region of high seismicity in central Taiwan is discussed. Seismic isolation systems will be used for reducing the seismic demand.

512 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Seismic performance of a two-span bridge subjected to fault-rupture

H. Choi & M. Saiid Saiidi Civil & Environmental Engineering University of Nevada, Reno, USA P. Somerville URS Corporation, Pasadena, California, USA S. El-Azazy California Department of Transportation, Sacramento, USA

Near-fault ground motions, which have caused much of the damage in recent major earthquakes, typically contain a strong and long-period velocity pulse that is the result of the forward directivity effects, as well as static permanent ground displacements caused by the relative movement of the two sides of the fault. Nearly all current bridge design criteria are based on far-field ground motions, and the effects of these unique characteristics of near-fault ground motions are not generally accounted for. More than 73% of bridges in California are close to active faults, thus necessitating a thorough study of near-fault earthquake effects and fault-rupture effects on bridges. The purpose of this study is to investigate near-fault ground motion effects on bridge columns and to determine the effects of fault-rupture on bridges that cross active faults. The focus of this paper is on testing and analysis of the performance of a two-span bridge model crossing a fault line with fling effects simulated by differential movement of shake tables. A quarter scale reinforced concrete bridge model with a total length of 20.5 m was subjected to a series of incoherent ground motions that included the fault-rupture effects. The bridge super- structure was continuous post-tensioned slab representing typical California bridge superstructure properties. The piers were all two-column bents with height varying from one bent to another making the bridge unsymmetric. The test was conducted in late May 2007 and analysis of the data was recently completed. The measured data showed that the seismic response under fault-rupture resulted in significant demands on bridge columns. In particular, damage modes and the location of the most damaged column were drastically different from a similar bridge model tested subjected to coherent earthquake records.

513 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Nonlinear modeling of a two-span reinforced concrete bridge model from pre-yield through failure utilizing contemporary analytical methods

N. Johnson, M. Saiidi & D. Sanders University of Nevada, Reno, USA

With the growing need for performance based design, accurate analytical modeling of bridge structures during seismic events is exceedingly important. The focus of this paper will be on nonlinear dynamic modeling of a reinforced concrete bridge system that was tested to failure at the University of Nevada, Reno research laboratory. A large scale two-span reinforced concrete bridge tested on three shake tables was subjected to a suite of earthquake motions in the strong direction. The bridge was composed of three two-column bents of varied height each rigidly connected at the base to one of three shake tables. The motions, derived from the 1994 Northridge earthquake, were applied to the bridge with increasing amplitude from a pre-yield state until failure of the bridge substructure. Upon completion of testing, in depth analytical modeling was conducted to determine the effectiveness of structural analysis software in duplicating the bridge model response. Both SAP2000 v.9 and Drain-3DX were utilized for this purpose. SAP 2000 represented a state of the art commercial structural analysis program, while Drain-3DX is a popular FORTAN based program written specifically to model nonlinear response of reinforced concrete structures. Although both models simulated reasonable results, the Drain-3DX model was determined to be most effective for both efficiency and accuracy to duplicate the nonlinear bridge model response. This computer model in conjunction with the experimental results was then used to further study system response of the shake table model as well as the response of other bridge systems. This paper will discuss the analytical modeling including comparisons of the model predictions with the measured shake table response. Computational modeling issues such as bond slip, shear stiffness, and damping will also be discussed.

514 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Development of an innovative seismic damper for large-scale bridges and sub-structured hybrid earthquake loading tests

H. Iemura, A. Igarashi & A. Toyooka Graduate School of Civil Engineering, Kyoto University, Kyoto, Japan

For enhancement of seismic performance of large scale bridges, installation of additional energy absorbing damper is expected to be very effective to reduce earthquake response of the structures. To achieve this objective, energy absorbing dampers shall have long stroke and large damping force with reasonable cost. It is quite difficult to satisfy these physical and cost requirements with the conventional dampers, such as viscous-type, inelastic-type and friction-type devices In this study, a HDR (High Damping Rubber) damper with long stroke and large damping force is newly developed as a seismic response control device for bridges. The special feature of the HDR damper is that it absorbs large amount of energy without axial force due to the development of a new rubber material. In this paper, fundamental performance test and hybrid earthquake loading test of the HDR damper is conducted. In the fundamental test, equivalent stiffness and damping ratio of the damper, and their strain dependence are investigated. Although, enough damping capability is confirmed, equivalent stiffness and equivalent damping ratio are found decreasing with the increase of shear strain. Furthermore, the dynamic response of a long-span steel cable stayed bridge with the proposed HDR damper is investigated by conducting the sub-structured hybrid earthquake loading test. These results show the proposed HDR damper has very high efficiency to reduce both the acceleration and displacement response of the bridge down to 50% with relatively small amount of the absorbed energy.

515

Practical application of BMS and BMS-DB

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Decision making processes and deterioration models of bridge management systems in Korea

B.-G. Kim, J.-N. Park & S.-H. Lee Yonsei University, Seoul, Korea M.-S. Park Korea Expressway Corporation, Gyeonggi-do, Korea

ABSTRACT: This paper reviews the decision making processes and deterioration models applied to network level BMS practically used in Korea. Two major systems called Korean Bridge Management System (KOBMS) and Highway Bridge Management System (HBMS) are inves- tigated. Researches for development of network level BMS were started from 1987 in Korea. KOBMS and HBMS were developed in 1995 and 1999, respectively. The major objectives of KOBMS and HBMS are to store and manipulate structural characteristics, inspection and mainte- nance records, and to support decision making on Maintenance, Repair and Rehabilitation (MR&R) activities together with priority ranking. Increase of computer capability has led development and use of more rigorous analysis for structures. It is true that KOBMS and HBMS have contributed to accumulation of inspection data and support decision making on MR&R. However, the decision support processes and deterioration models are highly depends on condition state of bridge components. It is expected that additional raw data would effectively assist in procuring more objective validity of the decision making processes and deterioration models. However, limitation of data reusability in digital data processing will need much manual work in quantitative analysis of relationships among diverse inspection data. One of the best approaches for management and manipulation of the raw data is to use model based document. In the recent studies, the extensible markup language (XML) is widely used as interface among heterogeneous systems or devices. The model-based document can be effectively used in automatic data transferring from measurement devices to database. In addition, raw data in the model-based document can be explicitly reported in human readable format as well as in computer readable format.

519 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

J-BMS database system 2007 for management of existing bridges in Yamaguchi prefecture

Kei Kawamura & Ayaho Miyamoto Yamaguchi University, Ube, Yamaguchi, Japan Jun-ichi Ishida Yamaguchi Prefectural Government, Yamaguchi, Japan

The Japanese local governments are deeply concerned about the management of existing bridges. This is because the number of structurally deficient or functionally obsolete bridges is increasing. It is likely to accelerate in coming years due to increasing volume of traffic, increasing weight of road vehicles, continued bridge aging and deterioration, and limited funds for Maintenance, Repair, and Rehabilitation (MR&R). In addition to the situation, the budget limitations of local governments for bridge maintenance actions and new construction are much more severe than those of the central government. The Yamaguchi prefectural government is one of the leading prefectures interested in develop- ing a practical management method and a computerized system for existing bridges due to the existence of over 3500 bridges having spans of 2 meters or more. The distribution of the number of bridges constructed has a high peak around 1960. The many bridges reach at least 50 years of age around 2010. By 2020, most of reinforced concrete bridges will be older than 50 years. Therefore, the replacement and maintenance of existing bridges would become a serious social concern. Consequently, the need for rational bridge management is undeniable. The authors contributed to the development of a comprehensive Bridge Management System (BMS) with co-workers at Yamaguchi University and the Yamaguchi prefectural government. The BMS for the Yamaguchi prefecture is referred to as the Japanese Bridge Management System (J-BMS). The system consists of three main subsystems: performance evaluation system, maintenance planning system, and bridge management database system. This paper presents the web-based database system for management of existing bridges in the Yamaguchi prefecture. The database system is referred to as the J-BMS database system 2007 (J-BMS DB 2007). The system has three main databases: bridge specifications database, inspec- tion database, and MR&R history database. The data items in the databases follow the guidelines of making bridge inventory, established byYamaguchiprefectural government in 2006. The inventory was developed for having a database for rational bridge management including the bridge identi- fication information, specifications, inspection results, etc. After briefly reviewing the outline of each subsystem of J-BMS DB 2007, this paper presents the scheduled inspection database system mainly. Several screen shots of the computer program are also presented in order to demonstrate how this system performs.

520 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Practical application of J-BMS to existing bridges in Yamaguchi Prefecture

A. Miyamoto & K. Kawamura Yamaguchi University, Ube, Japan J. Ishida Yamaguchi Prefectural Government, Yamaguchi, Japan

ABSTRACT: The authors have been developing a Practical Bridge Management System (J-BMS) for bridge network in Yamaguchi Prefecture, which integrated with the Concrete Bridge Rating Expert System (BREX) that can be used to evaluate the serviceability of existing concrete bridges. The J-BMS will be able to predict the deterioration process of existing bridge members, construct a maintenance plan for repair and/or strengthening based on minimizing maintenance costs and maximizing quality, and estimate the maintenance costs in single bridge. In this system, the Genetic Algorithm (GA) technique was used to search for an approximation of the optimal maintenance plan. In this study, a comprehensive decision support system is developed for maintenance strategies with/without annual budget limitations based on life cycle cost analysis of an entire bridge inventory, which form part of a highway network. Especially, an attempt is made to evaluate the applicability of the “maintenance plan optimization system”, a subsystem of the J-BMS, based on the inspection data for 169 concrete bridges in Yamaguchi prefecture. The main results are obtained in this study as follows: (1) An integrated bridge management system for Yamaguchi Prefectural Government has been developed based on a basic policy for systematic bridge management, practical database system, etc with IT-based advanced monitoring system. (2) The system was applied to an actual bridge network in Ube City area to evaluate its effectiveness. As the results, it will be able to make the priority of repair/strengthening works of existing bridges based on the various information from the database system, then the system helps bridge administrators to establish the rational maintenance strategies.

521 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Rational approach for the management of a medium size bridge stock

E. Brühwiler Ecole Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland

EXTENDED ABSTRACT: This paper presents first results of an ongoing project to establish a rational approach for the management of a medium size bridge stock. The bridge stock comprising 667 bridges is subdivided into four major groups depending on the construction type and construc- tion material, i.e. masonry bridges, masonry–concrete bridges, reinforced concrete bridges and steel bridges. In the first part of the paper, existing data is edited to deduce the most relevant characteristics in terms of the nature and condition of the bridges. The basic methodology to make cost estimations for the maintenance of the bridge stock based on present bridge condition evaluation is outlined in the second part of the paper. The available basic statistical data shows that 36% of all bridges are exposed to severe environ- mental conditions and only 23% of all bridge decks are equipped with a waterproofing membrane. All bridges are systematically inspected; the condition rating data shows that there are 20% of all bridges in deteriorated or bad condition. This indicates that it is no longer possible to allocate only minimal resources to maintain the bridge stock. Consequently, a period of time with additional resources is needed to catch-up. The basic idea of the suggested methodology (fig. 1) consists in constructing for each of the four construction types a dependency between the bridge condition and the intervention cost. The bridge condition is described by the “equivalent age”. In this way, bridge condition degra- dation can be derived from the assigned age equivalents and intervention costs are directly linked to condition. The intervention cost includes only the direct construction cost of the intervention, not considering indirect or user costs. The likely costs of intervention for the different construc- tion types as a function of their “equivalent age” are estimated based on information about past intervention costs as well as expert opinion. The sequence of interventions and the corresponding maintenance cost for the bridge stock is determined by minimizing all costs during the considered period while assuming that each bridge has only one planned intervention and budget constraints are respected.

Figure 1. Dependency between equivalent age (bridge condition) and intervention cost. 522 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Proposal for BMS deterioration curves based on the analysis of Hanshin Expressway inspection data

H. Nakajima & T.Yamagami Hanshin Expressway Engineering Company Limited, Osaka, Japan T. Kagayama Hanshin Expressway Company Limited, Osaka, Japan M. Hayashida Hanshin Expressway Engineering Company Limited, Osaka, Japan

ABSTRACT: The Hanshin Expressway forms an urban highway network in Hanshin area in Japan, and 90% of its total 233.8 km length under service consists of bridge structures. In 1964, the first 2.3 km section of the Hanshin Expressway network was opened to traffic and now, 40% of whole length is more than 30 years old. In future, with the increasing tendency of structural aging, it can be expected that the budget necessary for maintenance will increase. Hanshin Expressway, almost 20 years ago, database system which can store the large amount of asset, inspection and repair data, was developed and made available on line at each office. Although actual field works not always go as planned, this system can provide actual structure information. The system has been improved over the years, and now as it can be applied as asset management system, it contributes to support the rational and efficient management works of the Hanshin Expressway. Among the inspection results, analysis on “coating conditions” and “rust & corrosion” was carried out, and a deterioration model was established focusing on regression analysis and deterio- ration speed. The present paper reports the investigation steps and proposes to prepare deterioration curves based on the accumulated data.

523 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Methodology for determination of financial needs of gradually deteriorating bridges

B.T. Adey & R. Hajdin Infrastructure Management Consultants, Zürich, Switzerland

Bridges cost a substantial amount of money to build. In the short term, after they are built they require a relatively small amount of money, if any, to maintain. In the medium to long term, assuming there is no change in use and there are no extreme events, such as an earthquake, the amount of money required to maintain bridges varies considerably depending on the bridge type, the material of construction used, and the intervention strategy being followed. For example, one generic intervention strategy is to do nothing to a bridge until there is an unacceptable risk of failure of the bridge and then replace it, and another intervention strategy is to protect a bridge regularly against the processes of deterioration that affect it, e.g. painting a steel bridge at regular intervals, so there is only a negligible probability that the bridge will experience any significant corrosion. The former strategy will require no money in the medium term to maintain. The latter strategy will require money in the medium term to replace the protective coating of the structure. Both of these intervention strategies are valid and neither necessarily results in a negative impact to the infrastructure users. The choice between these intervention strategies, and all others, depends on the type and speed of the deterioration processes affecting the infrastructure, and the cost and effectiveness of the interventions. Taking these parameters into consideration, it is the task of bridge managers to determine and execute the optimal intervention strategies, i.e. the strategies that ensure the use of the bridges for the lowest long term costs. It is, however, in some cases desirable to follow strategies that do not result in the lowest long term costs, e.g. when it is known that the bridges will be decommissioned or during times of exceptionally tight budgets. When the bridges are to be kept in service in the long term, however, any deviation from the optimal intervention strategies results in increased overall expenditures. Although many bridge managers know in general which intervention strategies are optimal in the long term for their bridges, the infinite number of exact possibilities renders them unsure that the optimal strategies have indeed been determined and are being executed. This uncertainty coupled with external pressure to reduce short term maintenance costs is making it necessary to demonstrate both numerically and systematically the intervention strategies that are indeed least expensive, their costs and the expected condition of the structures if they are executed, as well as to quantify the consequences if they are not. In this paper a general methodology that can be used to meet these needs is outlined, and an example is given of how these predictions are possible using normally existing data. The method- ology is similar to those used in state-of-the-art, but far more data intensive, management systems, such as KUBA (Hajdin, 2008), and PONTIS (Cambridge Systematics, 2001).

REFERENCES

Cambridge Systematics, 2001. Pontis release 4.0: Technical Manual, American Association of State Highway and Transportation Officials (AASHTO), Washington D.C., U.S.A. Hajdin R. (2008). KUBA 4.0 Technical Manual, Federal Roads Authority of Switzerland, Bern, Switzerland.

524 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

The measure towards advanced of Bridge Management System for Expressway Bridges in Japan

YoshinoriWada, Shuhei Sakai, Takeshi Ohshiro, Atsushi Homma & Norio Ogata Nippon Expressway Research Institute Co., Ltd., Tokyo, Japan

ABSTRACT: In The Nippon Expressway Company Limited (NEXCO) which is the old Japan Highway Public Corporation (JH), Bridge Management System (BMS) which supports the well- planned maintenance by determining bridge condition, predicting future deterioration, and selecting optimal timing and method for repair and/or reinforcement was developed in 2003 and has been applied to bridge management in NEXCO after that. As a result, integrated maintenance database such as inventory data, inspection records and repair/improvement history evaluation data was built and we could grasp transition of total deterio- rated bridges which need repair and/or reinforcement by using several kinds of prediction formula based on Engineering knowledge. However, further advanced use of BMS for each bridge such as decision of medium-to-long term planned maintenance and risk management, the elaboration in function of deterioration prediction and rule for structional soundness evaluation is required. In this paper, study themes for the elaboration about deterioration prediction for Chloride attack, and structional soundness evaluation, e.g. quantitative evaluation for concrete bridges, and how to keep them in the good working order are discussed by analyzing inspection data and taking the local characteristic into consideration.

525 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Application of China bridge management system in Qinyuan city

B.F.Yan & X.D. Shao Institute of Bridge Engineering, College of Civil Engineering, Hunan University, China

ABSTRACT: After an enormous investment of highway networks and urban transportation sys- tem undertaken in the past 20 years in China, the highway networks among most key cities are close to completion. The bridge engineering will gradually transform from the stage of constructing new bridges and maintaining the existing bridges to the inevitable stage of bridge replacement and reha- bilitation. In the recent 3 years, some 10,000 (3∼4%) of the total bridges (>400,000) are considered deficient and are eligible candidates for replacement and rehabilitation. In order to preserve the bridge performance in a safe and serviceable condition for success of the infrastructure system, the comprehensive China Bridge Management System (CBMS) has been established to determine optimal strategies within budgetary constraints for Maintenance, Rehabilitation and Replacement (MR&R). The CBMS consists of a series of activities involving information gathering, interpreta- tion, condition assessment, deterioration prediction, cost accounting, decision-making, budgeting, and planning. The CBMS incorporates the following 6 principle modules: data management sys- tem, statistical and inquiry system, performance evaluation and decision making system, cost evaluation model, deterioration forecast and maintenance planning system, and GIS. The paper places emphases on outlining the visual inspection, bridge condition assessment, cost model and the maintenance planning in China Bridge Management System (CBMS). Particularly, the evaluation model for the visual inspection based condition assessment, evaluation of load carrying capacity, and bridge network sufficiency in network level is presented in detail. The practical applications of CBMS in the highway network of Qinyuan, Guandong, China and its deficiencies are finally presented. Due to the lack of the historical inspection data and the weakness of the weighted average bridge condition rating in determining the weight values, much more subjectivity and the qualitative analysis is introduced in rating, sorting, and planning. Moreover, the essential deterioration curves, statistic assessment and optimization techniques were not included in the present version, which requires improvement in the future version.

526 Practical applications of SHM techniques for railway systems

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Structural monitoring of a maglev guideway with wavelength division multiplexed FBG sensors

Wonseok Chung, Donghoon Kang, Inho Yeo & Jun S. Lee Korea Railroad Research Institute, Korea

Among several types of fiber optic sensors, there has been growing recognition of the potential use of Fiber Bragg Grating (FBG) sensors. FBG sensors have significant more advantages than other fiber optic sensors, as to include immunity to EMI, multiplexing capability, absolute measurement, high temperature endurance, and the added convenience of embedment. In the structural monitoring of Maglev guideways, electromagnetic interference (EMI) can be a significant problem as the Maglev train is powered by high-voltage electric feeding systems. Recently, researchers have successfully applied fiber optic sensors to modern railway structures mainly due to the EMI-immunity. This study presents a methodology on integrated monitoring system for a maglev guideway using Wavelength Division Multiplexed (WDM)-based Fiber Bragg Grating (FBG) sensors. The physical quantities such as stains, curvatures, and vertical deflections are measured in the field test. The strains are directly measured from multiplexed FBG sensors at various locations of the test bridge followed by curvature calculations based on the plane section assumption. Vertical deflections are then estimated using the Bernoulli beam theory and regression analysis. Frequency contents obtained from the proposed method are compared with those from a conventional accelerometer. Verification tests were conducted on the newly-developed Korean Maglev test track. Five FBG sensors are multiplexed in a single optical fiber and installed in parallel pairs along the entire length of the bridge by surface attachment, with one set at the top portion and the other at the bottom portion of the maglev guideway. In addition to the FBG sensors, a conventional displacement transducer and accelerometer are installed at the mid-span of the bridge for comparisons. The maglev train is passed over the guideway at different speeds ranging 10 km/h to 40 km/h to monitor its dynamic response. It has been shown that good agreement between the measured deflection and the estimated deflection is achieved. The difference between the two peak displacements was only 3.5% in maximum and the correlations between data from two sensing systems are overall very good. It has been also observed that the results give sufficient dynamic resolution in frequency domain. This confirms that the proposed technique is capable of tracing the dynamic behavior of the maglev guideway with an acceptable accuracy. Furthermore, it is expected that the proposed scheme provides an effective tool for monitoring the behavior of the maglev guideway structures without electro magnetic interference.

529 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Evaluation of modal parameters of a full scaled prestressed concrete beams for railway bridges

S.I. Kim Korea Railroad Research Institute, Korea N.S. Kim Pusan National University, Korea

Equidistant and repetitive moving forces induced from the passing train can cause unpleasant behavior of railway bridges. Resonance of the structure can be broken out when natural frequency of the bridge coincides with exciting frequency of moving forces. Therefore, exact application of dynamic properties of the structure will conduct exact understanding of dynamic behavior of the structure under moving train loads. As an alternative of conventional prestressed concrete (PSC) girders, various types of PSC girders are being developed and applied in bridge structures. Incrementally prestressed concrete girders and concrete girder with encased steel I-shaped beam are one of these newly developed girders. According to design concept, these new type of PSC girders are of some advantages to reduce their self-weight and make spans longer. However, dynamic interaction between bridge superstructures and passing trains would be sometimes one of critical issues in these more flexible railway bridges. Therefore, it is very important to evaluate modal parameters of PSC girders newly designed before doing dynamic analyses. In the present paper, a full scale incrementally prestressed concrete girder of 25 meters long as a test specimen was fabricated and modal testing on it at every prestressing stage was carried out to evaluate modal parameters including natural frequencies and modal damping ratios. Young’s modulus can also be obtained from static loading tests. In the modal testing, a vibration exciter controlled digitally as well as an impact hammer is applied to obtain frequency response functions more exactly and the modal parameters are evaluated varying with construction stages.

530 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Practical acceleration reducing method in high-speed railway bridges

W.J. Chin, J.W. Kwark, J.R. Cho, E.S. Choi & B.S. Kim Korea Institute of Construction Technology, Goyang, Korea

Field measurement tests conducted on PSC box girder bridges of the high-speed railway crossed by the KTX (Korean Express Train) revealed that acceleration responses exceeded the vertical acceleration limit of 0.35 g in winter. Even if no problem regard to the running safety of the high- speed train has been reported to date, the fastening force of the ballast will gradually degrade due to the excessive vibration of the bridge and lead to risk of failure of the track on the bridge. This study performed field application of various local vibration reducing devices to mitigate the excessive accelerations of the high-speed railway bridges occurring locally following the crossing of train traveling at high speeds and, proposed a simple and effective method consisting of the addition of mass in existing bridge. This study intends to investigate the applicability of vibration reducing methods using local vibration reducing devices and added masses as solutions to mitigate such excessive accelerations. Both analytic and experimental methods were conducted to examine the applicability of the local vibration reducing devices and verify their reduction effects. The vibration reducing effects were evaluated by means of vibration tests on a prototype bridge in order to reduce the local vibrational acceleration responses in real bridges. In addition, field test on an actual high-speed railway bridge was carried out considering additional masses of which results allow to derive vibration reduction alternatives and to propose directions for the improvement of their performance.

Figure 1. View of installed vibration reduction device. Figure 2. View of installed additional masses.

531 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Real-time damage detection of railroad bridges using acceleration-based ANN algorithms

J.T. Kim, J.H. Park & D.S. Hong Pukyong National University, Busan, Korea J.H. Yi Korea Ocean Research and Development Institute, Ansan, Gyeonggi, Korea

Recently,ANN algorithms have been studied for vibration-based damage detection due to the advan- tage in dealing with various types of input and output and the efficient pattern-recognition capability with various training patterns. Many researchers have made efforts to develop ANN techniques for identifying the location and the extent of damage, to develop a sub-structural identification method for complex structures using multilayer perceptron, to implement the ANN techniques using modal data to health monitoring of bridges. In this study, a real-time damage detection method using output-only acceleration signals and artificial neural networks (ANNs) is developed to monitor the occurrence of damage and its location in structures. In order to achieve the objective, the following approaches are used. Firstly, theoretical backgrounds are described. The problem addressed in this paper is defined as the stochastic process. An ANN-algorithm that uses output-only acceleration responses to detect changes in structural parameters in real-time is newly designed as shown in Figure 1. As the feature representing the structural condition, we select the cross-covariance function of two acceleration-signals measured at two different locations. By means of the acceleration features, a set of neural networks are trained for a series of potential loading patterns and damage scenarios of the target structure for which its actual loading histories and structural conditions are unknown. The feasibility of the proposed method is evaluated from numerical tests on a simply supported beam model under the effect of model uncertainty due to the variability of impulse excitation patterns used for training neural networks.

Figure 1. Schematic of acceleration-based neural networks for damage detection.

532 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Active piezoelectric sensor nodes and sensor self-diagnosis for structural health monitoring

S. Park & C.B. Yun Department of Civil and Environmental Engineering, KAIST, Daejeon, Korea G. Park The Engineering Institute, Los Alamos National Laboratory, Los Alamos, NM, USA D.J. Inman Center for Intelligent Material Systems and Structures, Virginia Tech, Blacksburg, VA, USA

This paper presents two emerging issues in the piezoelectric sensors-based Structural Health Mon- itoring (SHM) for critical members in civil infrastructures. They are active piezoelectric sensor nodes and sensor self-diagnosis. Firstly, tailoring the impedance-based method to wireless sens- ing technology is considered. An active sensing node incorporating on-board microcontroller and Radio Frequency (RF) telemetry is introduced. All the process including structural interrogation, data acquisition, signal processing, and damage diagnostic is being performed at the sensor loca- tion by the microcontroller. A Principal Component Analysis (PCA)-data compression algorithm is embedded into the on-board chip of the active sensing node. The feasibility of the active sen- sor node and PCA algorithm is validated through an experimental study inspecting loose bolts in a bolt-jointed aluminum structure. Secondly, a piezoelectric sensor self-diagnosis, that per- forms in-situ monitoring of the operational status of piezoelectric sensors in SHM applications, is investigated. A new electro-mechanical impedance model is proposed to incorporate the effects of sensor degradation and de-bonding for both functions of structural damage detection and sensor self-diagnosis. New parameters for sensor quality assessment of a PZT and coupling degradation effects between a PZT and bonding layer were incorporated into the traditional electro-mechanical impedance model for better estimation of the electro-mechanical impedance signatures and sen- sor diagnostics. The feasibility of the modified impedance model for sensor self-diagnosis using the admittance measurements was demonstrated by a series of parametric studies using a simple example of PZT-driven single degree of freedom spring-mass-damper system. Finally, discussions are made for further studies and future applications.

533 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Modal parameter extraction of high-speed railway bridge using TDD technique

B.H. Kim Department of Civil Engineering, Kyungnam University, Masan, Kyungnam, Korea J.-W. Lee & T.-Y.Yoon Research Institute of Industrial Science & Technology, Steel Structure Research Laboratory, Hwaseong, Gyeonggi, Korea

The TDD technique, that is an output-only modal parameter extraction method based on time domain, has been applied to the extraction of modal parameters of a high-speed railway bridge. Since the train passes by the bridge very quickly, the available time response samples are of small quantity. Using the traditional frequency-based modal analysis method, the accurate estimation of the modal parameters is difficult for such a high-speed railway system. For such problem, the TDD technique draws special attentions because the method requires relatively small number of time samples to identify the accurate modal parameters. In order to identify the temporal modal parameters such as natural frequency and damping ratio, a new algorithm that utilizes the system identification technique has been proposed. The proposed algorithm utilizes the modal cross correlation of acceleration responses that results from the mode shape extraction algorithm of the TDD technique. The proposed algorithm has been numerically examined, and applied to a high speed railway bridge. The results have been compared to those obtained by the existing methods. Based on the results, the following three conclusions could be made. First, the free vibration responses can be obtained from the ambient responses using the cross correlation in conjunction with the TDD method. Second, the proposed SVD process significantly removes the orthogonal noises in the extracted modal cross correlation. This is one of the original contributions on this paper. Third, the temporal modal parameters such as natural frequency and damping ratio can be accurately extracted by the well-known system identification approach.

534 Reliability and risk management

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Challenges for structural maintenance in coastal and offshore zones – Floating structures

E. Watanabe Kyoto University, Kyoto, Japan Regional Planning Institute, Toyonaka City, Osaka, Japan

It has been found that the most severe corrosions of offshore structures occur at the splash zone or at the zone immediately below the Mean Low Water Level. This paper summarizes the state- of-the-art of the long-term maintenance of floating structures in the coastal area. Figure 1 shows an example of a floating bridge. Described in the paper are various factors such as diffusion of oxygen, dissolved oxygen, different distribution of corrosion characteristics along the depth of water and ground with respect to surfaces such as splash zone, ebb and flow zone, mean high water level (M.H.W.L.)and mean low water level (M.L.W.L.),collisions with ships and floating obstacles, littoral transports, different corrosion process micro cells or macro cells, marine attached organisms, oxidization and deoxidization. Lastly, Particular anti-corrosion measures and investigations to predict the residual life are described: petrolatum lining, cathodic protection, electro-chemical corrosion forming process, heavy-duty coating such as stainless steel lining, titanium-clad steel, fluorine resin coating rather than polyurethane resin coating. The maintenance of infrastructure has in fact, become one of the world-wide urgent concerns. Even for bridges alone, it is considered to be neither easy economically and socially nor technically to maintain all of them in a healthy condition because of their large number and volume. One cannot optimistically expect sufficient budget for the maintenance if the current unfavorable economic condition of infrastructure holders and the people’s inherent indifference on the needs of such maintenance are taken into account. The strategy for maintenance of offshore structures including floating structures should follow the similar principle as that for bridge and land-based structures.

Figure 1. An example of a floating structure: a floating bridge on two pontoons.

537 As compared with the land-based structures, offshore structures are considered to be much more vulnerable to corrosion. However, in the area of floating structures, the maintenance policy has not been well established because of its relatively new history and only a few floating structures are in existence. Thus, it is proposed that the necessary concepts of AMS of land-based structures be adopted also for floating structures.

538 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Reliability analysis of Steel – Concrete Hybrid Cable-Stayed Bridge during construction

J.H. Yun & C. Moon ENVICO Consultants Co. Ltd., Seoul, Korea J.W. Sun & H.N. Cho Department of Civil Engineering, Hanyang University, Ansan, Korea

In case the side span ratio (side span length/center span length) due to site condition is extremely small and the construction condition of side span is favorable hybrid cable-stayed bridge may be adopted which reduce uplift reaction at the abutment and (or) intermediate pier using heavy-weighted-girder system at side span. The steel-concrete hybrid cable-stayed bridge which longitudinally connect steel and concrete girder is a difficult bridge type to design and construct because of nonlinear behavior and elaborate construction method as well as hybrid behavior of a different kind of materials. The bridges are usually unsafe during construction rather than after opening to traffic because there are many uncertainties to assess load and resistance effects. More- over the cable-stayed bridge is more risky than the general bridge because construction stage is complicated, boundary condition is varied and also erection is gradual. In the study, the models and methods for the safety assessment of Steel–Concrete Hybrid Cable-Stayed Bridge, which consists of steel composite girder and concrete girder erected by the FCM (Free Cantilever Method) and FSM (Full Staging Method) are proposed for the assurance of structural safety and the prevention against bridge collapse during construction. By the structural reliability approach the resistance and the load distribution characteristics of the bridge are defined and the strength limit state equa- tions of permanent structures and temporary structures during construction are suggested. An AFOSM algorithm and MCS technique are used for the reliability analysis of cables, pylons, gird- ers, steel–concrete conjunction part and temporary bents. Also, component reliability analyses are performed at the construction stages based on the structural system model. To demonstrate their rationality and practicality, the proposed models and approaches are applied to a real bridge. The sensitivity analyses of main parameters are performed in order to identify the critical factors that control the safety of similar bridges. As a result, it may be stated that the proposed models could be implemented as a rational and practical approach for the safety assessment of the bridges erected by FCM and FSM during construction. Based on the observations and the results of the applications, it may be concluded as follows : (1) Due to the fact that the reliability index β was higher than the general target reliability index βo it may be stated that the example bridge is safe enough during construction. It was known that the lower part of temporary bent is the most critical under FSM construction stage and the steel composite girder of adjacent pylon is the most critical under FCM. (2) From the results of reliability analyses it is formed that the reliability index of the temporary structure was lower than the those of the permanent structure. And thus it can be realized that the temporary structure is more dangerous than the permanent one and judged that the safety of temporary structure is more needed to notice during construction. (3) Based on the results of sensitivity analyses it is observed that the safety of both temporary and permanent structure are most sensitive to the variation of resistance strength and therefore, it may be noted that the reduction of variation of material, such as quality control, should be more ensured.

539 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Reliability analysis of a high-speed railway bridge system based on an improved response surface method

Andrzej S. Nowak Department Civil Engineering, University of Nebraska, USA Taejun Cho KR Technical Research Institute, Korea Rail Network Authority, Daejeon, Korea Do Hyung Lee Environmental and Railroad Engr. Paichai University, Daejeon, Korea Myung-Kwan Song Doosan Heavy Industries & Construction Co., Ltd., Korea

Reliability analysis for the target bridge system considering the uncertainties in the stiffness, moment of inertia, and the damping ratio of primary suspension in load terms, the uncertainties for the geometries of girder, slabs, and the elasticity of steel girders as resistance terms is performed for a high-speed railway bridge system. The limit states of driving safety of trains and the driving comfort of passengers are determined based on UIC code and Korean design specifications. It is hard to calculate the probability of failure of complicate structures using Monte-Carlo Simulations or using First Order Second Moment method, both of which are hard to calculate the derivative terms in implicit limit state functions. For the implicit limit state function, an improved Response Surface Method (RSM) is developed to evaluate the reliability of the implicit limit states of complex structures. For maximizing the adaptation of RSM, a weight matrix is used as a penalty function which accelerates the convergence of reliability. The results of improved RSM, compared with basic and adaptive method are verified with the improved convergence to the exact solution. To identify the response of the bridge, a new finite element model for three-dimensional FE analysis of high-speed train–bridge interactions is considered, The track structures are idealized using beam finite elements with the offset of beam nodes and assumed to be beams on a two- parameter elastic foundation. The vehicle model devised for a 300 km/hr train is employed. Evaluated reliabilities for the performance of the target bridge and the comfort of high speed train are compared with the conventional safety indices. The results of this study indicate “The important design parameter and the quantity”, to be improved to obtain the improved quality control of the high-speed train service. Keywords: Reliability analysis; High-speed railway bridge; Vehicle–bridge interaction analysis; Variable-node finite element; Improved response surface method

540 Table 1. Comparison of the converged reliability between types of response surface functions when using · = 3 + 2 + 3 − limit state function of g( ) x1 x1x2 x2 18 (Mean of x1 and x2 are 10, with C.O.V. of 0.5 and 0.505, respectively).

Reliability Probability Error Analysis method index of failure (%) Comments

MCS 2.533 0.006 0.0 Number of Simulation = 1,000,000 Linear RSM 1. Basic 0.923 0.178 3047.97 2. Adaptive 2.790 0.003 −53.39 Rackwitz-Fiessler 3. Adaptive Weighted 2.722 0.003 −42.62 Rackwitz-Fiessler with Kaymaz-McMahon Quadratic RSM 4. Basic 1.086 0.139 2353.59 5. Adaptive 1.878 0.030 433.92 Rackwitz-Fiessler 6. Adaptive Weighted 1.988 0.023 313.93 Rackwitz-Fiessler with Kaymaz-McMahon

Figure 1. Comparison of the converged reliability between types of response surface functions (g1: Basic Linear RSM, g2: Adaptive Linear RSM, g3: Adaptive Weighted Linear RSM, g4: Basic Quadratic RSM, g5: Adaptive Quadratic RSM, g6: Adaptive Weighted Quadratic RSM).

541 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

ANN-based reliability analysis of a fiber reinforced polymer deck

Jintao Cui & Dookie Kim Department of Civil Engineering, Kunsan National University, Kunsan, Korea Dong Hyawn Kim Department of Ocean System Engineering, Kunsan National University, Kunsan, Korea

In order to predict the failure probability of a complicated structure, the structure responses usually need to be estimated by a numerical procedure, such as finite element method. In this paper, an artificial neural network (ANN)-based reliability analysis is proposed. In this method, the relationship between the random variables (input) and structural responses is estimated using ANN models. ANN model is then connected to a reliability method, such as First Order Second Moment (FOSM), or Monte Carlo simulation method (MCS), to predict the failure probability.The proposed method is demonstrated through application on a Fiber Reinforced Polymer (FRP) deck. To reduce the computational effort required for reliability analysis, response surface method could be used. The obtained results show that theANN-based reliability analysis method has comparative accuracy and efficiency.

Figure 1. Structure of back-propagation neural network.

Figure 2. Comparison of the safety indexes using AFOSM.

542 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Strength of the chain for suspended scaffolds

Yasumichi Hino National Institute of Occupational Safety and Health, Japan

ABSTRACT: Suspended scaffolds consist of steel chains, chain clumps, and tube pipes, which have been used at the bridge construction site. The purpose of this study is to investigate the strength of the chain for suspended scaffolds. Suspended scaffolds have mainly been used for the painting works, assembling work of form panel, and so on. However, the Hyogoken Nanbu earthquake in 1996 triggered carrying out many aseismic reinforcement works of existing building or bridges, and consequently, specific model of the suspended scaffolds which can uphold heavy load materials at large-scale construction projects have also been needed. But this usage of suspended scaffolds is over the conventional way, and also, one of the con- struction accidents that heavy weight construction materials roughly 10 KN fell from the suspended scaffolds to the cars driving at the expressway have occurred. Many reinforcement works will con- tinue in the future, and accordingly, the safety of suspended scaffolds, where the heavy weight construction materials are used, must be investigated. Hence, the experimental studies on the chain, which is the main component of suspended scaffolds, were carried out, and the fundamental characteristics were investigated in this study. The major findings obtained in this study can be summarized as follows: (1) The maximum strength of the suspended scaffolds is roughly depending on the number of chains. However, the maximum strength of suspended scaffolds constructed with small length can not be estimated by only the quantity. (2) Suspended Scaffolds generally has the difference in their chain’s length, and thus their chains respectively have different loads under the elastic limit. The fact is very important because this difference makes the stress concentration for local chains. The design load of suspended scaffolds should be smaller than the elastic limit. Therefore, to equalize the length of each chain is necessary under the suspended scaffolds construction.

543 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Multimode analysis of extraneously induced excitation due to turbulence on cable-stayed bridges, including temporary stabilizing measures

Jae-Young Cho, Young-Rae Cho & Hak-Eun Lee Korea University, Seoul, Korea

In this study, the multimode frequency-domain approach is applied to the aerodynamic and aeroelastic analysis of cable-stayed bridges considering the construction sequence and temporary stabilizing measures. The accuracy of the analytical and experimental methods used for predicting prototype bridge performance is fundamentally important in assessing bridge safety. Numerical models (Figure 1) for accurate buffeting analysis of cable-stayed bridges, especially free cantilever superstructures are deduced for the vibrations due to extraneously induced excitation caused by tur- bulence. The three-dimensional analytical approaches in frequency domain are applied. The power spectral density and elemental internal forces of the cable-stayed bridge structure are computed using the developed computational code based on finite element method and random vibration the- ory. For the refined FE model, cable element and eccentric connections were first used. The static force coefficients for deck as well as pylon, based on wind-tunnel tests, were applied in the next step. Finally, most buffeting theories except aerodynamic admittances were implemented into the code. Full model test results were compared with those of numerical analysis to verify a simulation modeling strategy which is efficient and accurate for a cable-stayed bridge with various stabilizing measures in turbulent wind. And the effect of alternative temporary stabilizing measures was inves- tigated through various configurations on cable-stayed bridges. This, therefore, suggests the most effective placement of wind cables. Moreover, the effects of wind characteristics and aerodynamic parameters on buffeting response were investigated. It is concluded that aeroelastic full bridge model test results were successfully reproduced by multimode buffeting analysis combined with section model test results. The buffeting RMS responses, e.g. vertical and lateral deflections, rota- tion of deck and pylon foot moment, were acceptably accurate. The results from the parameter study, the level of turbulent vertical component, Iw and Lw can severely impact on the buffeting response.

Figure 1. Temporary stabilizing measures in 93% erection stage of model: (a) A-E2-NW; (b) A-E2-W1; (c) A-E2-W2; (d) A-E2-W3.

Figure 2. The buffeting RMS responses (93% erection stage): (a) Vertical Dir.; (b) Lateral Dir.; (c) Torsional Dir.; (d) Pylon foot moment.

544 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Reliability assessment of seismic expansion joints in bridges

J.E. Padgett Rice University, Houston, TX, USA R. DesRoches Georgia Institute of Technology, Atlanta, GA, USA

Movements at expansion joints and abutments in multi-span bridges have been a source of extensive damage is past earthquakes due to pounding and unseating at joints. Traditional expansion joints do not have the capacity to accommodate large hinge openings to avoid unseating and do not have properties that would facilitate reducing pounding. Seismic expansion joints provide a means to accommodate multi-directional structural movements, can rotate freely in all directions, can accommodate large displacements and can dissipate energy at the joint. This paper studies the effect of seismic expansion joints on improving the performance of a typical multi-span continuous bridge. The study uses a probabilistic framework to illustrate the effect of using a seismic expansion joint by evaluating the probability of meeting or exceeding specific damage states for a common bridge class, with and without seismic expansion joints. In addition, the study evaluates how the demands on various components of the bridge changes as a function of type of expansion joint used. A detailed model of a typical multi-span continuous bridge is developed with explicit considera- tion of the seismic joint behavior using the finite element analysis platform, OpenSees. Nonlinear time history analysis is run using a suite of ground motions representative of hazards in the central and southeastern US. The results of a deterministic analysis show that the bridge with seismic joints has roughly 20% lower expansion bearing deformations and 45% lower passive abutment deformations (or deformations in compression due to pounding). Similarly, the column demands for the bridge with seismic joints are approximately 20% lower than the bridge with non-seismic joints. The results from the fragility analysis show that the use of the seismic expansion joints not only improves the potential joint performance and reduces the likelihood of joint damage in an earthquake, but it also improves the performance of the overall bridge system. The probability of complete damage of the bridge using non-seismic joints is approximately 51% at a PGA of 0.4 g. However the damage probability is reduced to 40% when seismic joints are used for the same level of earthquake. This study illustrates that seismic expansion joints are a viable approach for improving the performance of a bridge system. The improved benefit achieved by the use of these joints may warrant the additional cost for seismic joints.

545 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

System-level reliability evaluation of bridge structures and networks by matrix-based system reliability method

J. Song & W.-H. Kang University of Illinois, Urbana-Champaign, IL, USA

ABSTRACT: Bridge structural systems and networks are often described as complex “systems” whose performance is modeled as logical functions of “component” events such as the occurrences of structural failure modes and the failures of structural members or bridges. For optimal decision- makings on bridge designs, maintenance and hazard mitigation strategies, it is essential to assess such system risks in an efficient and accurate manner. However, computing the probability of such a system event is often costly or infeasible due to complexity of the system event definition, statistical dependence between the component events, and the lack of complete information. For example, direct application of existing system reliability methods such as theoretical bounding formulas (Ditlevsen 1979) and first order reliability method approximations (Hohenbichler & Rackwitz 1983) is feasible only for series and parallel systems with little flexibility in incorporating various types and levels of available information. Song & Der Kiureghian (2003) proposed a method for computing bounds on the failure probabil- ity of any general systems with high flexibility in incorporating available information. The method divides the sample space into mutually exclusive events and describes the system failure probability and the available information by use of vectors and matrices. By solving a linear programming (LP), one can obtain the narrowest possible bounds on the system failure probability for given information. Recently, Song & Kang (2007) generalized this LP bounds method to a Matrix-based System Reliability (MSR) method in order to make use of the matrix-based formulation of system events and probabilities for the cases in which complete information is available as well. Unlike existing system reliability methods, the MSR method is uniformly applicable to general systems regardless of their complexity because both the system event and the joint probabilities of com- ponents are always described by vectors or matrices regardless of complexity of the system event definition. Since the matrix representation of a system can be obtained by algebraic manipulations of matrices representing component events or other system events, the MSR method provides a convenient framework of identifying/handling the system events as well. This paper presents the MSR method and demonstrates its merits in system-level reliability analysis of bridge structural systems and transportation networks by two numerical examples. First, the MSR method is used for estimating the likelihood of seismic damage of a bridge structure system. Using the analytical fragilities of various bridge components and the statistical dependence between the seismic demands (Nielson 2005), we analytically estimate the probability that at least one component fails. Various other system risks are conveniently assessed using the MSR method. This example demonstrates how the MSR method can deal with the statistical dependence between component events even when the source of dependence is not identified explicitly. Second, the risk of disconnections in a bridge transportation network is assessed (Kang et al. 2007) based on the seismic vulnerability of its constituent bridges (Gardoni et al. 2003). Also estimated are the probabilities of various system events such as the probability that each county will be disconnected from the hospital and the conditional probabilities of bridge failures given a disconnection as measures of the relative importance of bridges in the network.

546 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Rehabilitation and monitoring of a marine bridge in Ireland

A. Farrell & L. Duffy National Roads Authority, St. Martins House, Dublin, Ireland A. O’Connor Department of Civil Engineering, Museum Building, Trinity College Dublin, Dublin, Ireland J. Kelly Roughan O’Donovan, Consulting Engineers, Arena House, Dublin

ABSTRACT: This paper describes the strategy of rehabilitation and monitoring adopted for a reinforced concrete bridge located in a marine environment on the South Eastern Irish coast. The bridge, which is under the control of the Irish National Roads Authority (NRA), was constructed in 1980 using prestressed concrete beams with an in-situ reinforced concrete deck. The results of a structural assessment of the bridge required that repairs should be carried out to the main deck and crossheads of the structure to comply with the controlling limit states. It was decided that each of the 7- crossheads on the structure should be repaired using a different repair strategy and that these repairs should be instrumented so that their relative efficiency could be studied. This paper presents information on the project to repair and instrument Ferrycarrig Bridge. Each of the chosen repair strategies is outlined along with the reasons for its use, the instrumentation of the crossheads is presented demonstrating the monitoring equipment employed and finally the internet based data management system is presented to illustrate how the repair efficiencies will be tracked and analysed over time.

1 INTRODUCTION

Ferrycarrig Bridge which carries the N11 single carriageway over the River Slaney is located in Wexford on the south east coast of Ireland. Built in 1980, the 125.6 m long structure consists of 8 spans of precast, prestressed beams with a reinforced in-situ concrete infill, supported on intermediate piled pier foundations with reinforced concrete abutments at both ends. In August 2002 a Principal Inspection of the structure was undertaken in accordance with the Eirspan Bridge Management System. Due to the presence of extensive cracking in the crossheads a Special Inspection of the substructure was called for in the Principal Inspection Report. Follow- ing the completion of the inspection, testing and structural assessment of Ferrycarrig Bridge, a number of deficiencies from standards that affected the long term serviceability of the structure were identified that required repair and strengthening works to be carried out. These issues are summarised as follows (i) observed cracking in the pier crosshead beams was due to a lack of rein- forcement to resist the SLS stresses (shrinkage, thermal, creep). In addition, there was insufficient reinforcement to resist the applied ULS torsional moments. As the cracks were anticipated to con- tinue to develop and would thereby compromise the integrity of the piers it was concluded that pier strengthening works should be carried out, (ii) chloride levels in the concrete were found to indicate a distinct concentration gradient decreasing rapidly through the concrete. The levels of chloride ion concentration were found to be relatively high in the cover zone, (iii) the existing bridge deck waterproofing system had failed and needed to be replaced throughout, (iv) the existing bridge deck expansion joint over the central pier had failed and needed to be replaced, (v) the existing mechanical bearings at Pier 4 were no longer functioning due to extensive corrosion and needed 547 to be replaced. On the basis of these findings 3 alternative management strategies were presented to the NRA: (i) immediately repair and strengthen the structure, (ii) do nothing for 10 years before carrying out repair works or (iii) do nothing.

2 BRIDGE MANAGEMENT CONSIDERATIONS

The decision was made to repair and strengthen the structure immediately. The project afforded an opportunity to utilise the NRA’sresearch and development strategy in order to improve the decision- making process regarding the choice of concrete repair options for other structures in the medium to long-term. It also provided the NRA with an opportunity to implement a structural health monitoring system on the repaired elements; this would be the first time that such a system was used on a national road bridge in Ireland, and the NRA could witness the benefits and experience the issues which arise from installing and operating it. It was decided that each of the seven crossheads on the structure should be repaired using a different repair strategy and that repairs should be instrumented so that their relative efficiency could be studied with time. This provides the combined benefits of repairing and strengthening the bridge, enhancing durability, and gathering information which will help NRA bridge managers to make more well-considered decisions regarding repair options for concrete bridges which will deteriorate in the future. The results of monitoring the efficiency of the repairs will enable the NRA to gain better value for money in future rehabilitation schemes given engineers will be better informed in the choice of concrete repair strategies which offer a lower whole-life cost.

3 REPAIR AND INSTRUMENTATION OF FERRYCARRIG BRIDGE

In July 2007 refurbishment works on the Ferrycarrig Bridge commenced. In the following dis- cussion only the repairs of the crossheads are discussed. In this regard the following repair methodologies were employed: (i) Crosshead 1 – Repair Option: OPC mix + standard formwork, (ii) Crosshead 2 – Repair Option 1 + increased cover, (iii) Crosshead 3 – Repair Option 1 + surface treatment, (iv) Crosshead 4 – GGBS Mix, (v) Crosshead 5 – Repair Option 1 + mixed in corrosion inhibitors, (vi) Crosshead 6 – Repair Option 4 and finally (vii) Crosshead 7 – Repair Option 1 Six of the repaired crossheads (No’s 2–7) were instrumented with a combination of (i) chloride ion penetration probes, (ii) corrosion potential sensors and (iii) corrosion rate sensors. The probes & sensors were fitted on either side of the crosshead and on the end face on the sea-ward side of the structure to facilitate investigation of the (i) the relative performance of the employed repair mea- sures and (ii) the spatial variability of the performance results individual crossheads. The probes & sensors are monitored remotely.

4 CONCLUSIONS

The paper presents information on the survey, assessment and repair of a marine bridge located in South East Ireland. Following the decision to repair the structure it was decided that an opportunity existed to study the relative efficiencies of different concrete repair techniques. In this regard the paper outlines the varying repair methodologies chosen and the reason for their choice. The repaired crossheads were also extensively instrumented to allow for remote monitoring. The paper presents details of the assessment and rehabilitation of the structure and of the instrumentation scheme used for the study.

ACKNOWLEDGEMENTS

The Irish National Roads Authority are gratefully acknowledged for the financial support provided in facilitating this project.

548 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Repair of damaged footbridge after strike of excavator

A.G. Mordak Faculty of Civil Engineering, Opole University of Technology, Opole, Poland Z. Manko Civil Engineering Institute, Wroclaw University of Technology, Wroclaw, Poland

ABSTRACT: The paper is presented the case of damaged steel footbridge located over the Pocztowa Street in Legnica (Poland) as a result of strike excavator despite keep of road limit- ing clearance. The considerable deformations and destructions of main elements of the structure occurred as a result of this strike due to irresponsibility of road users caused of serious damages of that object. The elements tests as well as an analysis of results footbridge calculation were shown that as a result of considerable reduced of main girder (bottom flanges and web) and crossbeams strength until to their repairs, the object was closed for a normal service. The aim of this paper is to present analysis of safety threat problem of that object connected with its damages as well as its usefulness estimation for the future service. The important problem in such case is to undertake a right decision about the choice of repairs ways of damages and destruction as well as about admittance of the object to the future safe service. A lot of bridges located over roads or streets i.e. footbridges, flyovers, viaducts undergoing to mechanical damages despite keep of road limiting clearance. These are results of irresponsibility of road users. Deformations and damages elements of structure (mainly carrying) occurring as a result of strike can be cause of serious breakdowns of these objects. This is especially dangerous in the case of aged structures, very corroded with a reduced of load capacity. In steel structures the bottom flanges or chords (truss bridges) of main girders, the wind braces and very often elements of the deck grid are greatly subjected to damage. The right decision of bridge admittance to a normal service or even limited and choose of the proper method of occurred repair damages and destructions of possible become a very important problem. The main aim of this paper is to present an analysis of safety menace problem of steel footbridge connected with its damages by excavator and an assessment of its usefulness for a future safety service (Manko & Mordak 2004, Suchy & Angerman 2004). Own experience and studies concerning damages of similar road and railway objects were used to estimation of a technical state of the footbridge main girders and deck plate, and an analysis of conditions of the footbridge future service. The footbridge was closed for a current service until its repair due to very poor technical condition of the main girders according to the research of carrying elements and analysis of footbridge computation results. The range of repair works was designed in this way to the footbridge after its repair has got the required load capacity according to the Polish Loads Standards (PN-85/S-10030). This means exactly, that after the realization of repair it was possible to use object without no traffic limitations. For a protection before similar damages in this object in a future occur, the setting on access roads to footbridge of clearance durable limiters was recommended. After the analysis of static-strength calculations results, the opinion of technical and physical state of the span structure and supports as well as steps after repair executed it was affirmed that the object state was good, and that it was fully suitable to the future normal service.

549 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Effect of corrosion on the reliability of a bridge based on Response Surface Method

S.I. Jo Seoul Metropolitan Government, Seoul, Korea T. Onoufriou & A.D. Crocombe University of Surrey, Guildford, UK

The resistance response surface method is known to be a powerful tool for the evaluation of failure probability of complex structures. Advancement in FE analysis programmes makes it to investigate the effects of variation in basic parameters such as the width of the web on the reliability of a bridge system. In this study, the effects of corrosion of steel girders on the reliability of a beam-and-slab composite bridge were addressed. For the damaged state, the thickness loss parameter was used as an RV in addition to the strength of structural steel and the thickness of pavement for the evaluation of resistance RSFs. Figure 1 shows the change of reliability index for a single-lane loading and a two-lane loading. It needs to be noted that the probability of occurrence of the two-lane loading case (0.1) was not incorporated into the reliability shown in Figure 2. Note that reliability would be very similar to those in Figure 1, because the effect of the single-lane loading was negligible. It was found that the present bridge model is still safe in spite of severe corrosion on steel girders. It can be said that this type of bridge is ‘damage tolerant’ with regard to corrosion of steel girders. It was found that the 20% damaged bridge is in a ‘good’ state in terms of the reliability states proposed by Frangopol et al. On the other hand, a 40% damaged bridge is in a ‘fair’ state, which indicates that the reliability of the current bridge model is reduced significantly. However, one important thing to be noted is that the bridge model is not in the ‘unacceptable’ state in spite of the fact that 40% corrosion has occurred on steel girders. Figure 1 shows how reliability index varies according to damage on steel girders. Less conservative results obtained from the system reliability-based bridge assessment method- ology developed in this research will lead to rational use of a limited resources. As a result, it is expected that the use of this assessment tool will provide great benefits with bridge management authorities.

Figure 1. Change of reliability index (total). Figure 2. Change of reliability index (single-lane).

550 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Vibration control of cable-stayed bridge and derrick crane system during construction

Hyung-Jin Pae, Dong-Seok Kim Seoul National University, Seoul, Korea Wonsuk Park Korea Bridge Design & Engineering Research Center, Seoul, Korea K.-S. Park Dongguk University, Seoul, Korea H.-M. Koh Seoul National University, Seoul, Korea

As the span length becomes longer and the system becomes more flexible, the need for vibration control of long-span bridges has been increasing steadily. Especially for cable-stayed bridges under construction, the whole system which includes pylon, deck, cable and derrick-crane is very vulner- able to wind-induced vibration. During construction, workability is frequently declined due to the vibrations, which could result in the delay of whole construction process. Sometimes aerodynamic instability could be occurred. In this paper, therefore, we propose control systems for a cable-stayed bridge under construction by cantilever method and perform wind-induced vibration analysis. An optimal control method is presented in order to reduce vibration of the hoisting derrick crane as well as the bridge deck. For this purpose, we develop a reasonably simplified cable-stayed bridge model with a derrick crane mounted, which can simulate mass hoisting and jib angle change. We also present the effectiveness of the proposed method by investigating the acceleration and displacement responses of the system with respect to safety and workability criteria. Keywords: cable-stayed bridge, derrick crane, cantilever method, wind excitation, vibration control.

551

Repair and strengthening

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Ductility of CFRP strengthened concrete bridge girders

S. Kim Restl Designers Inc., Gaithersburg, MD, USA R.S. Aboutaha Syracuse University, Department of Civil & Environmental Engineering, Syracuse, NY, USA

Concrete bridge girders may require increase in the flexural strength to allow for heavier traffic loads. An increase in the flexural strength could be achieved by the addition of structural elements, e.g. Carbon Fiber Reinforced Polymer (CFRP) composite sheets. The response of CFRP strength- ened concrete girders is not as ductile as un-strengthened girders. In order to improve the ductility of CFRP strengthened girders, end anchors are utilized. These anchors could be made of end CFRP diagonal sheets, or a set of steel plates and adhesive bolts. This paper presents an experimental of a series of reinforced concrete girders strengthened with CFRP composites to enhance the flex- ural capacity and ductility. Some of the girders were also provided with an end CFRP diagonal anchorage system to increase flexural ductility. The main variables were the amount of CFRP composites, the amount of the longitudinal and shear reinforcement, and the effect of end CFRP diagonal anchorage system. Sixteen full-scale girders were tested to investigate the effectiveness of the end CFRP diagonal anchors in increasing the flexural ductility of CFRP strengthened concrete girders. Test results suggest that CFRP diagonal anchors are very effective in increasing the flexural ductility of CFRP strengthened concrete girders. The test results also suggest that the effective- ness of end CFRP diagonal anchors depends on the detailing of the anchors and orientation of the carbon fibers.

555 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Modal analysis and step-by-step repair operation of a two span concrete skew bridge to replacement of its elastomeric bearings

Reza Akbari & Shahrokh Maalek University College of Engineering, Department of Civil Engineering, University of Tehran, Tehran, Iran Houshang Ashayeri Isfahan Department of Road & Transportation, Isfahan, Iran

Condition assessment and extraction of dynamic properties of a two spans simply supported con- crete post-tensioned skew bridge in the middle of Isfahan-Tehran highway during repair operation, step-by-step, is discussed in this paper. The repair operation was carried out in three main steps. First, the asphaltic pavement was removed, then the expansion joints was removed and repaired, and finally the old neoprene elastomeric bearing was replaced with the new ones. For extraction of modal properties of the bridge, the Operational Modal Analysis technique was used. Moreover, the effect of step-by-step repair operation and number of the bearings is considered using several modal tests of the bridge during repair operation. Finally, an FE model of the bridge was constructed and the effect of shear modulus of elastomeric bridge bearing on dynamic characteristics of the bridge was studied in detail. This bridge is tested in five stages: First stage: modal test of bridge deck with traffic, before any repair operation. Second stage: modal test of bridge deck without traffic, before any repair operation. Third stage: modal test of bridge pier before any repair operation. Fourth stage: removing the expansion joints and modal test of bridge deck without traffic. Fifth stage: replacement of elastomeric bearings and modal test of bridge deck without traffic with new elastomer pads. Also, several modal tests are performed on plain elastomeric bearings of the bridge. Results of these tests are used to determine the shear modulus of elastomeric bearings, as a key parameter in FE modeling of the bridge. The results show that: • Worn out elastomers can reduce the natural frequencies of the bridge and its modal damping. • Replacement of elastomeric bearings causes an increase in natural frequencies up to 15 percent in test results and in FE results. • Asphaltic overlay do not affect the dynamic properties of the bridge significantly. • FE results show that this bridge vibrate in-plane that it can be critical during strong ground motions, like earthquake. For this reason, additional repair operations are need for construction of two shear key in each support as transverse restraints.

556 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Bridge widening – technical, economical and aesthetical aspects

G. Boro´nczyk-Plaska& W. Radomski Warsaw University of Technology, Institute of Roads and Bridges, Poland

Bridge widening is one of the fundamental operations connected with road modernization. Depending on the individual situation, the bridge widening can include widening of the super- structure only or both the superstructure and substructure. In some cases, a new bridge structure can be constructed in parallel and close to the existing one. In any case the bridge widening provides more or less change in the appearance of the existing bridge structure. In some cases replacement of the existing bridge by the new and wider one can be also analyzed and applied. Therefore, bridge widening should be considered taking into account technical, economical and aesthetical aspects. The fundamental cases of bridge widening are classified into 9 groups as well as characterized and commented in particular. Important role of the theoretical analysis and bridge investigation performed prior to the bridge widening are emphasized. The safety margins in the old existing bridge structures resulting from the simplified calculation models assumed during their design are also emphasized. The situations when the bridge widening requires to be accompanied by the structural strengthening are characterized. Moreover, the role of redistribution of the internal forces in the widened bridges is also presented as well as an increase of the load life level in the widened bridges is discussed. A special attention is paid to the widening of old bridges with some historical values, where conservation of their original appearance can be the decisive factor. General rules of the economical analysis concerning the bridge widening and its economic efficiency are presented, including the relevant formulae for such analysis. The necessity of the analysis of maintenance costs during the expected whole service life of the widened bridge is explained. The above mentioned economical rules can be also applied when the removal of the existing bridge and its replacement by the new and wider structure is considered as alternative solution. Some of the bridge widening cases are exemplified and briefly characterized, taking also into account the aesthetical aspects. The general rules concerning the widening of the historical bridges are also discussed and exemplified. Finally, some conclusion remarks are formulated concerning the bridge widening as an important element of bridge engineering in many countries.

557 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Behavior under compressive loads of steel structural members repaired by heating and pressing

M. Hirohata Osaka University, Osaka, Japan Y.-C. Kim Joining and Welding Research Institute, Osaka University, Osaka, Japan

When large infrastructures like as bridges are damaged by a fire, traffic accident or earthquake, it is required that they are quickly repaired so as to ensure the traffic of emergency and transportation of aid goods. Sometimes, local buckling damages of steel members, whose damages are mainly a little, are rapidly repaired on site by correction with heating and pressing. In repairing damaged steel members, correction by heating and pressing is a useful and effective method. However, it is still not elucidated clearly the effect of correction by heating and pressing on mechanical behavior of steel members. It is necessary to confirm safety and reliability of members corrected by heating and pressing. In order to investigate the effect of correction by heating and pressing on mechanical behavior of steel members, a series of experiments was carried out. Test specimens were shown in Figure 1. Compressive experiments for virgin specimens were conducted and local buckling was generated. Buckling deformation was corrected by heating and pressing below A1 transformation temperature. After that, compressive experiments were carried out again. From the results of each compressive experiment, it was confirmed that the ultimate strength of the repaired specimen was almost same as that of the virgin specimen. This result indicated that the mechanical properties of the steel never deteriorated with correction by heating and pressing below A1 transformation temperature. However, the buckling mode of the repaired specimen was considerably changed comparing with that of the virgin specimen. So as to verify the dominant factors governing the behavior under compressive loads of the specimen after the correction by heating and pressing, the experiment was simulated by the elastic- plastic large deformation analysis. From the results of the analysis, those factors were; the residual imperfection, which was the deformation inevitably left in the specimen by incomplete correction, and the increase of yield stress resulting from work hardening by local buckling and its correction.

Figure 1. Test specimen.

558 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Effectiveness of prestressed Carbon Fibre Reinforced Polymer (CFRP) sheets for rehabilitation of prestressed concrete girders

Y.J. Kim North Dakota State University, Fargo, ND, USA M.F. Green Queen’s University, Kingston, ON, Canada C. Shi Behlen Industries, Brandon, MB, Canada J. Ford Halsall Associates Ltd, Toronto, ON, Canada L. Bizindavyi SNC Lavalin, Montreal, QC, Canada R.G. Wight Royal Military College of Canada, Kingston, ON, Canada

ABSTRACT: This paper discusses the effectiveness of prestressed Carbon Fibre Reinforced Poly- mer (CFRP) sheets for strengthening prestressed concrete girders. Eight medium-scale prestressed concrete girders (L = 3,600 mm) have been tested under static and fatigue loading configurations. Three of the girders have experienced 80 cycles of freezing-and-thawing effects between tem- peratures of −20◦Cto+20◦C. A 3-dimensional computational model is developed to predict the behaviour of the test girders. The girder strengthened with prestressed CFRP sheets shows an increase of 74% in the load-carrying capacity. The prestressed CFRP sheets significantly con- tribute to stress redistributions in the internal prestressing tendons, cracking behaviour, and fatigue resistance of the strengthened girders. Although the CFRP sheet shows brittle failure characteris- tics, ductility of prestressed concrete girders is improved after strengthening with prestressed CFRP sheets. The effect of the freezing-and-thawing cycles is an important consideration when prestressed concrete girders are externally strengthened using CFRP sheets. Finally, some recommendations are given for such a strengthening method.

559 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

The use of polymer concrete materials for construction, maintenance, rehabilitation and preservation of concrete and steel orthotropic bridge decks

Arthur M. Dinitz Transpo Industries Inc. Sidney Park Woo Bang Ceratech Corp.

ABSTRACT: Over the past 20 years Thin Polysulfide Epoxy Slurry (Filled) Overlays, which are water, skid and UV resistant have been used as effective wearing surfaces for the rehabilitation and preservation of concrete and steel orthotropic bridge decks. The unique slurry application results in an overlay that is 9 mm to 12 mm in total thickness. The single slurry application reduces the time required to complete a project and eliminates the potential for problems due to rain or contamination which has been a major problem when using standard multiple coat broom and seed polymer overlays. The thin overlay adds minimal dead load to a structure which is very important; in addition it does not require joints to be raised which can be costly. Other polymer concrete material systems such as Methyl Methacrylate (MMA) have been devel- oped to obtain high strength in one hour in ambient temperatures from −415 deg C to 40 deg C. These materials have been successfully used for concrete patching, bearing pads, joint headers and closure pours. Other development has lead to High Molecular Weight Methacrylate (HMWM) polymer sealers that can heal/seal small cracks in existing concrete surfaces and are applied using only the force of gravity and do not require pressure injection. Precast Polymer Concrete Panels “stay in place forms” have been used for bridge rails, tunnel panels and safety barriers.Their high strength, resistance to corrosion and ability to be manufactured in virtually any shape make them ideal for fast track projects. This paper will discuss the physical properties of these polymer material systems as well as Precast Polymer Concrete Panels and the proper selection of a specific material to meet project requirements. It will also give details on the use of these systems in Korea, United States and Canada.

560 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

The correlation between crack and residual stress generated by repair welding in service

Y.C. Kim & S.H. Lee Joining & Welding Research Inst., Osaka Univ., Osaka, Japan Y. Agano Katayama Stratech Corp., Osaka, Japan

1 INTRODUCTION

The confirmation of safety of the structure, which is repaired or reinforced, is so important. But, the basic characteristics of residual stress generated by welding for repair/reinforcement of steel bridge in service are not yet enough understood. So, in order to extend the range of repair/reinforcement works with welding in service, the weld- ing experiments for repair/reinforcement are carried out under load. And, in order to investigate the basic characteristics of residual stress generated by welding for repair/reinforcement, the residual stresses are measured by stress relaxation method.

2 RESULTS AND DISCUSSION

2.1 Repair welding experiment under cyclic loads According to the results of PT and macrograph for every type specimen, the crack has been observed on the root and center of throat. It is about 0.1∼0.5 mm. And, the specimen of type-A and B, which has a groove angle of 45 degree, shows the penetrated crack. It is supposed that the weld bead is pear-shaped. The pear-shaped weld bead is easy to generate hot crack. Namely, the welding conditions with a bare possibility for a generation of pear-shaped weld bead are demanded for the repair welding.

2.2 Measuring of residual stress generated by static loads

The residual stress component (σy) of perpendicular direction of welding line generated by repair welding under compressive load is larger than that under tensile load. It is because that the applying direction of load is equal to the direction of residual stress component (σy). So, the applying load has a direct effect on the residual stress component (σy). The residual stress component (σx) of welding line direction generated by repair welding under compressive load is smaller than that under tensile load. It is also an influence of applying load. Namely, the weld metal is shrunken during a cooling after welding. The tensile residual stress is generated by this procedure at near weld metal. And, the compressive load is induced tensile stress in the applying direction of load and compressive stress in the perpendicular direction of applying load respectively. So, the magnitude of maximum tensile residual stress component (σx) of welding line generated by repair welding under compressive load is decreased. And, the area of tensile residual stress is increased for the stress redistribution. On the same principle, the magnitude of maximum tensile residual stress component (σx) of welding line generated by repair welding under tensile load is increased.

561 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Repair of a concrete bridge by composites CRFP

M. Abdessemed Civil Engineering Department, University of Blida and Ministry of Public Works, Algeria S. Kenai Geomaterials Laboratory, University of Blida, Algeria A. Kibboua, J-L. Chatelain & B. Guillier National Earthquake Engineering Center (CGS), Algeria A. Bali Construction & Environment Laboratory, Polytechnic National School of Algiers, Algeria

Among the main concern of the Algerian public works authorities, the maintenance and the con- servation of the existing bridge structures composed of more than 8000 bridges of which 53% are road bridges. The 2006 statistics showed that most road bridges are made of concrete or masonry requiring either maintenance and/or repair and strengthening. The application of reinforcement techniques such as jacketing and strengthening by composite materials makes it possible to recover the mechanical characteristics of theses structures. Hence, this will keep the structures in working state and restore their initial strength. This paper presents the results of an experimental investiga- tion of a case study assessing the behaviour of a reinforced concrete bridge reinforced by composite carbon fibre materials using ambient vibrations non-destructive tests. This case study enabled us, on the one hand, to evaluate the in-situ behaviour of the strengthened structure and the advantages of composite materials in repairing such reinforced concrete bridges and on the other hand assess the contribution of dynamic tests in the monitoring of such behaviour.

Keywords: Repair, reinforced concrete, bridge, composite materials, strengthening, ambient vibrations.

562 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Innovative rehabilitation of a damaged prestressed concrete girder bridge using prestressed CFRP sheets: Design and specification

Y.J. Kim North Dakota State University, Fargo, ND, USA M.F. Green Queen’s University, Kingston, ON, Canada G.J. Fallis Vector Construction Group, Winnipeg, MB, Canada R. Eden Transportation for Manitoba Floodway Authority, Winnipeg, MB, Canada R.G. Wight Royal Military College of Canada, Kingston, ON, Canada

ABSTRACT: This paper presents an innovative rehabilitation method for an impact-damaged prestressed concrete girder bridge. The bridge, constructed in 1963, has been damaged due to frequent collision of heavy trucks, resulting in concrete spalling and prestressing strand rupture. Prestressed Carbon Fibre Reinforced Polymer (CFRP) sheets are used to recover its inadequate flex- ural strength and serviceability. Detailed design approaches are presented, including an integrated anchorage systems. Particular attention is paid on the repair specification which may aid practicing site engineers conducting similar repair methods. A 3-dimensional finite element analysis model is discussed to demonstrate the effectiveness of the repair.

563 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Rehabilitation of bridges with concrete overlays

C.A.M. de Smet & J. Kunz Hilti Corporation, Schaan, Principality of Liechtenstein

1 INTRODUCTION

The Hilti Corporation has developed a special injection-type connector to attach new structural concrete layers to existing concrete structures. Static and dynamic research, including a fatigue approach as required in bridge construction, form the experimental basis for a comprehensive design concept. The article references the design method for concrete overlays incorporating this new type of concrete connector.

2 SYSTEM EVALUATION WITH DIFFERENT CONCRETE CONNECTORS

The simplest form of a shear connector is a straight rebar. The disadvantage of this solution is the required minimum embedment depth of 9–10 times the rebar diameter. Therefore, bent or hooked rebars are often used in practice. In order to avoid concrete crushing when using minimum values of mandrel diameters, cross bars inside the bends are required, this is however cumbersome to construct due to alignment issues. Alternate solutions exist like welded heads or threaded rods with screwed plates (Fig. 1 types (a) and (b)). Recently a new trumpet shaped hollow shear connector (the so called “HCC-B” element) made out of malleable cast steel was developed by the Hilti Corporation (Fig. 1 type (c)). The ribs develop a rigid friction hold to the borehole, thereby allowing limited immediate loading already before injection with Hilti HIT-RE 500 mortar through the shaft of the element. The protruding “head” of the connector can be utilized as a support for either the constructive- or static reinforcement The first comprehensive tests investigating the load bearing behaviour of interfaces roughened with modern equipment, in combination with post-installed standard rebars with welded heads (Fig. 1 type (a)) crossing the joint, were conducted in 1995–1997. A debonding agent was used to account for potentially unclean surfaces typically found on jobsites. Follow-up tests with the above described Hilti HCC-B elements (Fig. 1 type (c)) have recently been conducted, investigating the load bearing behaviour under static as well as dynamic loading (Figs 2–3).

Figure 1. Concrete connectors: a) Rebar with welded head, b) threaded rod with plate, c) cast steel connector Hilti HCC-B. 564 Figure 2. Testing for shear connector HCC-B. Figure 3. Dynamic testing of connector HCC-B.

565 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Bond and flexural behaviour of RC members strengthened with CFRP composites

D.S. Yang Sherbrooke University, Quebec, Canada J.M. Park, S.N. Hong & S.K. Park Sungkyunkwan University, Gyeonggi-do, Korea

This paper first investigates the interfacial bond behaviour in the shear test for CFRP laminates bonded to concrete. Two concrete compressive strengths are considered. For each concrete, the effective bond length is estimated using a linear regression analysis; the maximum bond stresses and slips at different load levels are calculated from measured strains in the CFRP laminates. Also, a simple bond-slip model for the CFRP/concrete joints, developed from the shear tests, is proposed and used in a nonlinear finite element analysis of beams strengthened in flexure with bonded CFRPs. Comparisons are made between the analytical and experimental results for the CFRP-strengthened beams. Very good agreement between the numerical predictions and test values is obtained. In this paper, a total of 7 beams were tested in flexural test and analyzed in finite element analysis, with the anchorage system, amount of prestressing used as experimental variables; one control beam, two simplified FRP-boned beams and four prestressed FRP-bonded beams. All the beams were subjected to three-point test under deflection control, with the loading, deflection and failure modes recorded to the point of failure. A nonlinear finite element analysis of the beams in the flexural test was also performed using the DIANA program, with respect to nonlinear concrete material, reinforcement and the interfacial bond-slip model of shear test between the concrete and CFRP plates. The failure mode of prestressed CFRP plate-beams was not debonding, but FRP rupture. For RC members strengthened with externally bonded prestressed CFRP plates, 1st and 2nd debonding of the composite material occurred. After the debonding of the CFRP plates occurs in the bonded system, the behavior of bonded CFRP-plated beams changed to that of unbonded CFRP-plated beams due to the fixing of the anchorage system. The flexural test results and analytical results for the RC members strengthened with CFRP plates were also compared. The ductility of the beams strengthened with CFRP plates as the anchorage system was considered high if the ductility index was above 3. The analysis results showed a good agreement with those obtained experimentally with respect to the debonding load, yield load and ultimate load.

Keywords: CFRP, debonding, shear test, flexural test, finite element analysis.

566 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Seismic performance improvement of bridges by earthquake protection systems in Korea

D.-H. Ha Konkuk University, Seoul, Korea H.-M. Koh, S.Y. Lee & H.J. Kim Seoul National University, Seoul, Korea I.J. Kwahk Korea Institute of Construction Technology, Gyeonggi, Korea

This paper describes the state-of-the-art research activities on seismic isolation systems for improving the seismic capacities of the bridges in Korea. Researches and applications of seismic protection systems in Korea have been activated since the early 1990’s, partly due to the increasing recognition of the potential seismic risk in Korea and partly due to the needs for rehabilitating the deteriorating infrastructure systems. Among various seismic protective systems, seismic isolation systems are generally accepted in Korea to be most cost-effective and efficient alternatives for seismic protection against conventional design concept, especially considering the low-to-moderate seismic characteristics of the Korea peninsula. Accord- ingly, many theoretical and experimental researches have been carried out to develop satisfactorily performing isolators in a view of seismic performance as well as economic cost. Therefore, optimal design procedure based on minimum LCC concept is presented and such approach is accepted to be more expedient for the design of seismically isolated bridges in Korea. In order to verify the adequacy of the new design concept based on the LCC minimization, experimental studies on seismically isolated bridge are introduced as well, which include pseudo-dynamic test of scaled pier, dynamic field test of full-scale and temperature dependence tests of seismic isolators. The pseudo-dynamic tests of scale-down bridge components have been conducted and the dynamic field test of full-scale seismically isolated bridge was carried out as well. Some research results of the field testing are adopted into the new bridge design specifications published in Feb. 2005. Up-to-date applications of the seismic isolation system to the bridges are reported in detail, and development of new isolation devices are introduced in this paper. In addition, various applications of isolation devices to the new and existing bridges are summarized, and major retrofitting method is described.

Keywords: Seismic isolation system, life cycle cost (LCC), field test, retrofit.

567 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Application of new prestressing method using carbon fiber plates

Takeshi Ohshiro & YoshinoriWada Nippon Expressway Research Institute Co., Ltd., Tokyo, Japan Akitaka Takeuchi Central Nippon Expressway Co., Ltd., Nagoya, Japan Kazuteru Morikita Central Nippon Expressway Co., Ltd., Tokyo, Japan Hiroshi Yasumori & Terumitsu Takahashi DPS Bridge Works Co., Ltd., Tokyo, Japan

The expressway network, which was intensively constructed during the period of rapid economic growth, now serves as a precious and matured infrastructure. However, increasingly many bridges on the expressways will need repair and reinforcement works due to deterioration and increases in size of passing vehicles, which were not predicted at the time of construction. This paper describes a new prestressing technology developed with such a background for reinforcing bridges. The new prestressing technology improves the bending strength of concrete members and involves introducing effective tensioning force of about 200 kN into carbon fiber plates (made of carbon fiber reinforced polymers) and fixing the plates onto the members by pasting with resin adhesives. Since the load applied to the anchorage zone by the work is small, no additional rein- forcement is needed to control the bearing force, cracks and the tensile force on the back surface. The method does not require stopping the traffic but can be executed from the bottom of bridges. The method can reinforce an entire continuous girder bridge just by fixing the plates onto the under- surface of the girder only at the center of the spans because the redundant force of the prestress counteracts the tensile force at the upper edges of the main girders. Today, this new technology is increasingly used for reinforcing existing bridges since works completed have shown high cost effectiveness and reinforcement effects.

Photograph 1. Applied tensioned CFRP strip. Figure 1. Mechanisms of reinforcement.

568 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Experimental study of bolted joint with ultra thick plate and M30 bolt

Jinhoon Kim Chungbuk National University, Cheongju, Chungbuk, South Korea Jongseob Byuoun Daewoo Steel Tech, Jincheon, Chungbuk, South Korea Jae Byung Jo Kyonggi University, Seoul, South Korea Kyoungsup Jung Chungbuk National University, Cheongju, Chungbuk, South Korea

The ultra thick plate and high performance steel plate have been developed recently and as a result, the actual application at the sites is expected to realize soon, and there’s also the report that the field bolting joint proved to be more economical for bridge at the coast. In view of the faulting on bolt joints, bolting distance and end distance in association with the tolerance in the thickness of ultra thick plates for applying the field bolting joint, using F10T M30 high tension bolt for newly-developed SM520-TMC proved not to cause any problem.

569 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Monitoring for fatigue crack propagation of steel plate repaired by CFRP strips

H. Nakamura & K. Maeda Tokyo Metropolitan University, Tokyo, Japan H. Suzuki Meisei University, Tokyo, Japan T. Irube TTK-Corporation, Tokyo, Japan

ABSTRACT: The aim of this study is the development of the monitoring technique for fatigue crack propagation using commercially available strain gages, when fatigue cracks are repaired by CFRP strips, and crack tips are covered with CFRP strips. First, FE analyses were carried out. The relationships between strain variations on CFRP strips and the crack progress were clarified, and the applicability of monitoring was examined. Next, fatigue test was carried out. A flat steel bar with a through crack was repaired using CFRP strips and epoxy resin adhesive. The validity of the analytical results was confirmed. Therefore, it is feasible to monitor fatigue crack propagation in bonded area by measuring of the strain variations on CFRP strips.

1 INTRODUCTION

Recently, a lot of steel bridges have been suffering from fatigue damages due to the increase in traffic and aging, etc. Then, an efficient and simple repair method is required in construction work. The authors have investigated experimentally into the repair method for fatigue cracks by adhesion of a Carbon Fiber Reinforced Plastic strip (it is hereafter called a CFRP strip). As a result, it is effective to delay the crack progress by using a CFRP strip and epoxy resin adhesives on the cracked steel plate. However, it has been a problem still pending that the crack progress in the area covered with CFRP strips cannot be evaluated. If progress of the fatigue crack is appropriately detectable after repair, it becomes possible to predict residual life with sufficient accuracy, and it is very useful for maintenance. In this study, when the fatigue cracks are repaired by CFRP strips and epoxy resin adhesives, a new trial of the evaluation of crack propagation is proposed in the area covered with CFRP strips, where it is difficult to check directly by visual detection methods, etc. Since CFRP strips as a repair material transmit the axial force, when the fatigue cracks are progressing in the area covered with CFRP strips, the relationships between the axial strains on the surface of CFRP strips and crack length were considered. The authors tried to predict the position at the crack tip during crack progress, by measuring axial strains of CFRP strips on the path of crack propagation. FE analyses and fatigue test were carried out. The relationships between the strain variations on CFRP strips and crack progress were clarified, and the feasibility of monitoring was examined. 570 Figure 1. Element divisions near the crack and Figure 2. Relationships between strains on the reference points. CFRP strip and the crack length a.

∗ Figure 3. Relationships between the strain difference ε and the crack length a in the target point of C18.

2 MONITORING BASED ON STRAIN VARIATIONS OF THE CFRP STRIP

2.1 Analysis conditions and test procedures The target fatigue crack was a through crack in a finite width plate. The examination object is the flat steel bar with a through crack and the CFRP strips, which are bonded on both sides by epoxy resin adhesives. The through crack of length 2a was set to the center of the steel plate. The case where a crack progressed under the uniform tensile stress σn of 100 MPa was analytically and experimentally examined. Figure 1 shows the element divisions near the crack and the reference points. As shown in Figure 1, the multiple points of axial strains of the CFRP strip on the path of crack propagation, namely five reference points from C10 to C18, were evaluated at intervals of 2mm, and the reference points were same as the experimental study.

2.2 Strain variations on the CFRP strip and the crack length Figure 2 shows the relationships between the strains of the reference points and the crack length a in reference points as compared with analytical results, respectively. The experimental results are qualitatively equivalent as analytical results, and experimental values are larger than analytical values in the strains on the CFRP strip. However, it is difficult to pinpoint the position of the crack tip directly from these results with properties similar to the analytical case. Then, Figure 3 shows the relationships between the strain difference ε and the crack length a ∗ in the target point of C18. This figure means that the peak has appeared clearly, that the interval of strain gage was so wide that the strain difference was large, and that the crack length was estimated to be short a little for the position of the peak. The peaks are caught clearly near the target point, and the tendency of both is very well in agreement. 571 3 CONCLUSION

In this study, a new trial of the evaluation of crack propagation is proposed in the area covered with the CFRP strips when the fatigue cracks are repaired using CFRP strips and epoxy resin adhesives. The following conclusions were obtained: (1) It is feasible to pinpoint the position of the crack tip by the strain difference between two points on the CFRP strip. (2) The crack progress can be tracked using multiple strain gages. (3) The proposed method using strain gages can be also applied to debonding detection of CFRP strips.

572 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Reevaluation of stresses and displacement of horizontally curved girders of a continuous span bridge

D.J. Kim Jacobs Edwards and Kelcey, Inc, New York, NY, USA C.P. Fan STV,Inc, New York, NY, USA B.T. Yen Lehigh University, Bethlehem, PA, USA

ABSTRACT: For the rehabilitation of the curved bridge with a series of straight girders and a curved deck, horizontally curved girders will be used. The curved girders were originally designed based on the earlier results of analyzing segments of straight beams, and then were analyzed using the conventional theory of curved beams. The traditional simple boundary condition was adopted for the curved girders. The girders adhered to all the requirements of current design rules. A recent study on the behavior of horizontally curved girders subjected to vertical loads and bending moments revealed the importance of considering large displacement, large rotation and deformation of cross section of these girders. Because a horizontally curved girder deflects laterally and rotates in addition to deflecting vertically under its own weight even before the weight of the wet concrete, the formulation of equilibrium equations for solution must consider the higher order terms of strain-displacement relationship. The lateral displacement introduces a P-Delta effect to the girder. Furthermore, the deformation of the cross section of the girder induces a relative rotation between the flanges and the web and generates additional strains in the girder. These effects have all been incorporated into the curved finite line element for analysis (Kim &Yen2007). Because of these newly available information, the displacement and stresses of the rehabilitation girders were reevaluated for safety. Three schemes of connecting a fascia girder at the mid span bracing point and at the ends of the span were developed. The schemes were restraining the lateral displacement at the connection points, restraining the lateral displacement and rotation at the mid span and the continuation of the fascia girder to the next span. These schemes are reevaluated using the three methods of analysis. First method used the closed form solution for the equations for curved beams derived by Dabrowsky (1968). The closed form solution was developed from a first order theory. The second method was by a grid analysis in which conventional straight beam element is used to model the curvature. This method incorporates large displacement and small rotation, and is permitted by the Guide Specifi- cation of AASHTO (2003). The warping moment of the curved beam are calculated by an empirical equation. The third method employed the curved finite line element. The analysis incorporates the large displacement, large rotation, P-Delta effect and cross sectional deformation of a curved member. By comparing the results from the three methods, it was found that the conventional theory of curved beams under estimated displacement, warping moment and the maximum stress of the curved bridge member. The method of grid analysis using segments of straight beams seriously over or under estimated the displacement and maximum stress of the member. The method of curved line element provided results consistent with the connection schemes. Results from this method

573 showed that even in the elastic range of material properties, the rehabilitation girder behaved nonlinearly. From the results of the analysis using the curved line element, it was possible to select an effi- cient, easy to construct and cost effective connection scheme for the safety of the bridge during constructions.

574 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Railroad bridge replacement in the US today: Current technology and future possibilities

F. Moreu ESCA Consultants, Inc. Urbana, IL, USA T. Nagayama University of Tokyo, Tokyo, Japan J. Zeman University of Illinois at Urbana-Champaign (UIUC), Urbana, IL, USA G. Rus University of Granada, Granada, Spain S. Y. Lee & T. Park Hanyang University Seoul, Korea

In recent decades loads, speed and capacities have been rapidly increasing on railroad tracks in the US, whereas the bridges carrying the trains have not changed. As a consequence, many railroad bridges need to be either repaired or replaced every year in order to maintain a reliable infrastructure for both owners and clients. Since railroads are private enterprises in the US, the profit associated with repairing or replacing bridges becomes a key element. Owners rely on economic aspects when confronting a potential bridge replacement project. The first part of this paper describes some specific railroad bridge replacement cases with regards to design and construction. This presentation further describes some of the new technology available today for structural engineering use, identifying potential areas of application for the improvement of railroad bridge replacement and maintenance. Within the world’sinfrastructure, railroad bridges are key factors which highlight the significance of a rapidly growing transportation system. During the last 50 years, loads, speed and capacities have been increasing rapidly whether the bridges carrying them have not been updated. Several railroad bridges need to be either repaired or replaced to maintain a reliable infrastructure for both owners and clients. Even though timber is an uncommon material for new bridge construction in the US and around the world, rail traffic today is sustained by a large percentage of timber trestle bridges, some of them over 100 years. In most cases, however, it is agreed upon that the timber trestle bridges in use for approximately 50 years are now at the end of their service life. In some particular cases, timber trestles do not last the anticipated 50 years of service and need to be replaced earlier than expected due to observed abnormal displacements under normal loads and traffics. In comparison with the publicly owned transportation infrastructure, railroad bridges are known within the entire structural community to be more robust against sudden catastrophic service failures. According to the GovernmentAccountability Office (GAO), there has not been a fatality associated with a railroad bridge since 1957. However, recent events such as the wooden trestle bridge in Myrtlewood, Alabama that collapsed in May 2007 while carrying segments of the space shuttle and the tragic bridge collapse in Minnesota have raised national concerns about the 76,000 railroad bridges in use today. Railroads are run by private owners who expect to make profit despite their aging infrastructure. Therefore, as presented at the beginning of this paper, old structures need to be identified and discriminated from the newer ones when replacements or improvements are needed. The owner of the structure does not want to invest in bridges that could last longer than other bridges in the network. To quantify in a fast, efficient, and quantitative (and comparable) way is a big demand on railroads today. Wireless sensing and other

575 methods of structural evaluation can provide the tools to the railroad owners to improve actual assessment of the structural integrity of their bridges. New advancements in wireless sensing make it possible to implement this new technology with more affordable budgets and soon ready-to-use technology.

576 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Assessment of repair cost and service life of repaired concrete structures after chloride attack

Ha-Won Song, Aruz Petcherdchoo & Hyun-Bo Shim Department of Civil and Environmental Engineering, Yonsei University, Seoul, Korea

This paper present an assessing method for repair cost and remaining service life of concrete struc- tures repaired after chloride ion attack. The quantitative mechanism of chloride ion diffusion in concrete structure can mathematically be described using the Fick’s second law, but an analytical solution to explain the chloride ions diffusion after concrete cover replacement in existing con- crete structures using the Fick’s second law is limited due to the complexity in solving the partial differential equation. A Crank-Nicolson based finite difference method is proposed to predict the remaining life for repaired concrete structures. Furthermore, the cost of concrete cover replacement due to chloride ion attack can also be computed using a linkage between the application time and the cost of concrete cover replacement. Finally, numerical examples are presented.

ASSESSMENT: For repair material with w/c equal to 0.3, when the chloride profile reaches the threshold value of 1.2 kg/m3 in the year 26, the first repair is applied. immediately after repair, the chloride profile decrease to zero due to removing the chloride ions together with the taken- off concrete. However, the chloride ion concentration suddenly increases, because of immediate redistribution of chloride ions from the original concrete. After a period of time of surface chloride penetration and remaining chloride redistribution, the chloride profile reaches the threshold value at the year 75, the same repair will be repeated, until the end of designed or required service life of 100 years. It can be observed that the remaining service life of the structure is 29 years after the first repair and 49 years with two repairs. The present value of cumulative repair cost is calculated at the time of repair application as shown.

Figure 1. Effect on chloride ion profile and repair cost due to different repair depths and materials.

REFERENCES

Ann, K.Y. & Song, H.W. 2007. Chloride Threshold Level for Corrosion of Steel in Concrete. Corrosion Science. 4113–4133. 577 Song, H.-W., Kim, H.-J., Saraswathy, V.& Kim, T.-H. 2007. A Micro-Mechanics Based Corrosion Model for the Prediction of Service Life in Reinforced Concrete Structures. International Journal of Electrochemical Science. 2. 341–354. Song, H., Petcherdchoo, A. & Shim, H. (submitted). Service Life Prediction of Repaired Concrete Struc- tures under Chloride Environment using Crank-Nicolson based Finite Difference Formulation. Cement and Concrete Composite.

578 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Tests on cast iron carried out to repair bearings in Tumski Bridge in Wroclaw (Poland)

Z. Manko Institute of Civil Engineering, Wroclaw University of Technology, Wroclaw, Poland

ABSTRACT: The way in which the old cast-iron bearings (damaged during World War II) in the hundred-twenty year-old Tumski Bridge in Wroclaw (Poland) were repaired is described. To select a suitable modern material for the new bearing elements, old cast iron specimens taken from the fractured plates under the fixed and expansion bearings had to be tested. The Polish cast iron of a proper grade (ZL 150) was selected for the new components of the bearings on the basis of the determined mechanical-strength properties of the old bearing material.

The subject of this paper is a two-span steel truss road bridge located on the Katedralna Street route in Wroclaw (Poland). The bridge was built in place of an old wooden bridge in the years 1888–89. It crosses an arm of the Oder river, connecting the Piasek Island with the Ostrow Tumski. For over a hundred years “the GreenTumski Bridge” with St. John’sCathedral and churches in the background has belonged to the most popular painting-photographic subjects and the most frequently visited historic sites in Wroclaw’s Old Town. But the bridge’s over one hundred years long service resulted in damage to its deck components and in the closure of the bridge to vehicular traffic. Pedestrian traffic was permitted but repairs had to be done since any further degradation of the bridge’s structural components would have resulted in their destruction. It was assumed that the bridge would not carry heavy-vehicle traffic. The fact that the bridge was historic, highly original and blended well into the surroundings – one of 17 Wroclaw bridges recognized as a monument of the art of engineering in 1976 – was taken into account in the technical-economic analysis. The results of the strength tests and the metallographic examinations show that the structural material of the old bearings in the Tumski bridge (built over 120 years ago) meets the requirements for the Polish grey cast iron grade ZL 150. The determined Young’s modulus is lower than the one required by the actual standards (120 GPa) (PN-92/H-83101 1992, PN-92/H-83123 1992). This may be due to the quite long period of aging of the material. The crystal structure of the bearing cast iron is highly homogeneous. The chemical analyses showed that the material has a low phosphorous content, which contributes to its superfluidity in the liquid state (the cast iron fills a (sand) mould better) but it slightly increases its brittleness. The microstructure examinations did not reveal any irregularities and their results are in agreement with the standards in force PN-92/H-83101 (1992), and PN-92/H-83123 (1992). The tests performed on the specimens cut out from the cast showed that the tested material corresponds to cast iron grade ZL according to the Polish Standards PN-86/H-83101 (1986), and PN-92/H-83101 (1992). This could be determined accurately after tests on a sample ingot had been carried out. An analysis of the cast iron’s basic parameters shows that the one-hundred twenty-year-old bearings are still suitable for service and that the strength of the cast iron has decreased only slightly as a result of fatigue.

579

Research & applications for bridge health monitoring

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Residual structural performance of corroded steel tubes submerged in seawater

K. Sugiura Kyoto University, Kyoto, Japan E. Watanabe Regional Planning Institute, Osaka, Japan K. Nagata Nagoya Institute of Technology, Nagoya, Japan I. Tamura Kajima Corporation, Sendai, Japan

ABSTRACT: Presented herein is a study on the residual axial strength of corroded steel tubes submerged in seawater for about 19 years to find their corrosion process, paying attention to the effect of the corroded surface profiles on the strength and deformability. From the experimental and analytical results of this study, it was made clear that the measurement of the cross-sectional profile of the test specimen was precise enough and the prediction of the strength through the finite element method taking into account the corroded profiles and their magnitude was proved to be extremely reliable.

1 INTRODUCTION

Since the most of infrastructures such as bridges and tunnels to sustain the economic and living activities have aged and damaged due to various environmental load exposure for the long use, the necessity of assessing the remaining structural performance of structures, that is the durability of structures has arisen in the recent in order to assess the life cycle cost of the infrastructures from the material stage to the demolish stage. Therefore, it is obvious that the maintenance of infrastructures have become a key issue to assure the desired performance in operational condition and to prolong the life of structures as long as possible. As for the life cycle performance of steel structures, particularly in an extremely corrosive marine environment, it is conclusive that the damage due to corrosion is crucial rather than that by fatigue. While more extensive studies have been conducted on the deterioration of steel members, few studies have been directed towards examining the residual strength and repair of steel members with corrosion damages. At first, the precise profile measurement of corroded surface of steel tubes is made and stub column tests on corroded steel tubes is carried out. In addition, the analytical investigations by nonlinear finite element analysis of the compressive strength of stub columns with patch corrosion is also conducted. Specimens are prepared from the experimental jetty consisting of RC decks and steel piles exposed in the marine environment for about 19 years, with the maximum corrosion rate of approximately 0.2mm/year, which is the largest in the splash zone.

2 CONCLUDING REMARKS

In order to develop the remaining structural performance evaluation method of corroded steel structures, the precise profile measurement of corroded surface of steel tubes and stub column 583 tests are carried out. In addition, the nonlinear finite element analysis of the compressive strength of stub columns is conducted. It is made clear from the experimental and analytical results of this study that the measurement of the cross sectional profile of the test specimen was precise enough and the prediction of the strength through the finite element method taking into account the corroded profiles and their magnitude was proved to be extremely reliable.

584 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Modal flexibility and curvature for damage assessment: Laboratory demonstrations

Mustafa Gul & F. Necati Catbas University of Central Florida, Orlando, FL, USA

ABSTRACT: In this study, different damage simulations using a laboratory model (Figure 1) are presented along with the analysis results. Two damage indicating features (modal flexibility-based displacements and curvatures) are evaluated by conducting dynamic tests on a steel grid structure. Structural behavior before and after damage is evaluated by inspecting the deflected shapes obtained using modal flexibility. In addition, the authors used modal curvature as a complementary damage index to modal flexibility where modal flexibility did not provide substantial evidence of structural damage. It is observed that curvature is advantageous for certain cases. Issues related to using curvature as a damage identification index are also addressed.

SUMMARY OF THE RESULTS: It is shown that flexibility-based displacement and curvature indices performed well as indicator and locator for damage. By inspecting the deflected shapes obtained using modal flexibility, the quantification of response under various loading such as uniform distributed load can be computed. In restrained boundary condition damage simulation,

Figure 1. The steel grid and pseudo-loading used for deflection and curvature analysis.

Figure 2. Change in the curvature after stiffness loss.

585 stiffness changes of around 30% were observed, and identification, localization and quantification of the damage could be achieved. More localized damage simulations (stiffness loss at the joint) yielded a more subtle decrease in stiffness (5.7%) where the authors investigated the use of curva- ture of the displacement shape obtained from dynamic measurements. For the same case, curvature change was observed to be 19% (seen in Figure 2). Consequently, curvature was advantageous for some cases where the modal flexibility based displacement results did not provide significant changes. However, it should also be noted that computing the second derivative for curvature might lead to numerical errors.

586 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Benchmark studies for Structural Health Monitoring using computer vision

Ricardo Zaurin & F. Necati Catbas University of Central Florida, Civil & Environmental Engineering, Orlando, Fl, USA

EXTENDED ABSTRACT: Structures are complex engineered systems that ensure society’seco- nomic and industrial prosperity. Unfortunately, they are often subjected to unexpected loading scenarios and severe environmental conditions not anticipated during design that will result in long-term structural damage and deterioration. Structural Health Monitoring (SHM) paradigm offers an automated method for tracking the health of a structure by combining damage detection algorithms with structural monitoring systems. Recently, the authors have proposed a novel frame- work for SHM of bridges by combining computer vision and a distributed sensors network that allows not only to record events but to infer about the damaged condition of the structure. Video stream is prescribed to be used in conjunction with computer vision techniques to determine the class and the location of the vehicles moving over a bridge as well as for surveillance purposes. A database with information from vehicles training sets, experimental results from the sensors net- work and, analytical models is suggested. Then, the proposed system, by interpreting the images and by correlating those with the information contained in the database, evaluates the operational condition of the bridge and/or will emit alerts regarding suspicious activities.

By adding vision capabilities to a SHM framework, correlation between moving loads (traffic) and responses (sensors readings) can be performed and cause/effects relationship can be determined However, although there are many benefits of using video in conjunction with sensing technol- ogy, there are also many issues related with this approach. Technology requirements, algorithm needs, and the amount of data to be handled are increased dramatically. For these reasons, labo- ratory validation is crucial before field deployment. The authors have designed and built the UCF 4-span Bridge. This structure is a laboratory setup used to demonstrate and test an innovative and practical integrated monitoring and analysis system, combining real-time video images with sen- sor readings. Remote controlled vehicles crawl over the bridge while a video camera supervises the structure providing traffic video stream, at the same time, a distributed array of sensors col- lects data. Correlation between moving vehicular load and structural responses is determined and presented. Description of the experimental set-up and finite element model are presented as well.

587 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Field monitoring of continuous steel-concrete composite girder during internal force adjustment of the Siyuan Bridge

Wenliang Lu Lehigh University, Bethlehem, PA, USA Beijing Jiaotong University, Beijing, China Dan M. Frangopol Fazlur R. Khan Endowed Chair, Department of Civil and Environmental Engineering, ATLSS Center, Lehigh University, Bethlehem, PA, USA

ABSTRACT: In continuous steel-concrete composite girders, the bending moment near the inter- mediate support of the continuous girder is harmful because concrete deck has a very low tensile strength. Jack-up-down at intermediate support locations is a solution to prevent or to minimize the crack width in concrete deck. This paper presents the results of field monitoring of the Siyuan Bridge girder during jack-up-down process. The Siyuan Bridge of Beijing Airport Light Rail Line was designed as continuous over three spans (38 m + 49 m + 38 m). It is a girder with variable section depth. The maximum depth is 3.2 m and the minimum depth is 1.8 m. The cross section of twin steel girders is U shape and the twin U girders are connected with each other by sev- eral diaphragms along the whole length. A total of 36 vibrating wire strain sensors and eight displacement transducers were installed. The construction sequence of the steel-concrete composite girder of Siyuan Bridge contains five stages: Stage 1 is the erection of U girders; Stage 2 is the jack-up of steel girders at two intermediate piers; Stage 3 is pouring concrete deck; Stage 4 is the jack-down of the composite girder to its design permanent position after concrete curing; and Stage 5 is stretching prestress tendons located in concrete deck. Monitoring was performed at all stages except stage 1. The performance during construction of both U steel girders and concrete deck was observed. The monitoring data showed that the value of dial indicator increased at the external side of support center and decreased at the internal side of support center. The support rotation agreed well with the theoretical predicted behavior. The monitoring strain data of steel girder varied widely and torsion phenomenon occurred during jack- up. Measured strains of concrete deck demonstrated that the expected prestressing was obtained by jack-down. From the monitoring data the following conclusions are drawn. 1. Jack-up-down at intermediate supports can produce prestress in concrete deck of continuous composite girder; the effect of this method depends on the jack-up distance. If the concrete deck is designed according to full-prestress concept, only jack-up-down is not enough and prestress tendons must be used. 2. Screw jacks are operated by workers. For this reason, it is extremely difficult to keep the same step during the jacking process. The different lifting distances of jacks introduce girder torsion and large local stresses. 3. The monitoring of Siyuan Bridge showed that the actual mechanical behavior of U girder is very complex. A detailed construction plan is necessary for jacking-up this kind of girder because its torsion stiffness is relatively low. Finally, field monitoring provided an important check of the safety of the construction process of this bridge, and also provided helpful data for the design of this type of structures.

588 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Application of GPS monitoring technology to the construction of the pylon

Jung Seok Lee Korea Infrastructure Safety and Technology Corporation, Goyang, Korea Jah Geol Yoon Daelim Industrial Company, Seoul, Korea

According to the development of construction technology, public concern of preservation of envi- ronment, human desire to esthetic civil structures, etc, recent trend of cable supported bridge is getting longer, higher and slimmer. Millau Bridge in France, recently completed, has the 250 m- high pylon. In Korea, 10 and more cable supported bridges are constructed, designed and planed. Construction of pylon and high-rise pier is more complicated than the construction of other mem- bers. Because the concrete pylon is constructed step-by-step and the shape is not simple, it is so difficult to meet the construction standard of vertical tolerance of the pylon at each stage. If a pylon has poor verticality, the safety of pylon and even the structural system is affected by the additional moment caused by the eccentricity. When constructing pylons or high-rise piers, optical surveying instruments and laser instruments are used for checking the verticality. However, higher the top end of a pylon, lager the error of verticality. Moreover, as the top end of a pylon goes high, the checking time is getting longer. In this study, we tested the availability of GPS monitoring technology in the field of the construction of the pylon by comparing the coordinates of the forms measured by optical surveying instruments with the results by GPS system. Also we developed the software for checking the coordinates of the forms and adjusting them effectively.

589 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Bridge fatigue reliability assessment and prediction

Kihyon Kwon Department of Civil and Environmental Engineering, Lehigh University, Bethlehem, PA, USA Dan M. Frangopol Fazlur R. Khan Endowed Chair of Structural Engineering and Architecture, Department of Civil and Environmental Engineering, ATLSS Center, Lehigh University, Bethlehem, PA, USA

ABSTRACT: Bridge performance may become unsatisfactory due to fatigue developed by steadily increased and repeated loads. For this reason, the assessment and prediction of bridge fatigue reliability is an extremely important issue. For the assessment of bridges, modern concepts for monitoring and maintenance programs have been developed. Monitoring programs as well as maintenance interventions are necessary to evaluate and preserve the remaining lifetime of struc- tures under uncertainties in environmental and mechanical conditions, material properties, and loading history. During the last decade, several researchers have studied the topic of design of structural mon- itoring systems to produce more reliable results. The measured data associated with an existing monitoring system should be used in a reliability assessment method with the appropriate applica- tion. Reliability evaluation based on data monitored on critical regions of a structural system can be effectively considered in prediction models since it is not practically possible to continuously monitor an entire structural system. The monitoring data may, however, contain sensor errors in measurements. Therefore, sensor errors can be treated as an error factor for the reliability evalu- ation, and the traffic increase per year can be selected as another factor for assessing remaining lifetime of structures. For this reason, the assumptions of various probability density functions for various parameters associated with the fatigue loading and the estimation of the traffic increase per year are required to predict stress ranges. On the other hand, the degradation of bridge resis- tance considering uncertainties is currently predicted based on the AASHTO Guide Specifications. The basic equations of the AASHTO Guide Specifications are used for the evaluation of fatigue reliability and remaining lifetime. The objective of this paper is to evaluate the fatigue reliability of bridges and to predict their lifetime reliability under fatigue on the basis of monitoring data. The approach is illustrated on an existing highway bridge in Pennsylvania.

REFERENCES

AASHTO Guidelines. 2002. AASHTO Standard Specification for Highway Bridges. Washington, D.C. Chung, H.Y. 2004. Fatigue Reliability and Optimal Inspection Strategies for Steel Bridges. Dissertation. Civil and Environmental Engineering Department, The University of Texas at Austin, Austin, TX. Connor, R.J. & Fisher, J.W. 2002. Report on Field Inspection, Assessment, and Analysis of Floorbeam Con- nection Cracking on the Birmingham Bridge Pittsburgh PA. Lehigh University’s Center for Advanced Technology for Large Structural Systems (ATLSS), Bethlehem, PA. Connor, R.J., Fisher, J.W., Hodgson, I.C. & Bowman, C.A. 2004. Results of Field Monitoring Prototype Floorbeam Connection Retrofit Details on the Birmingham Bridge. ATLSS Center, Bethlehem, PA. Frangopol, D. M. & Liu, M. 2007. Maintenance and Management of Civil Infrastructure Based on Condition, Safety, Optimization and Life-Cycle Cost. Structure and Infrastructure Engineering 3(1): 29–41. Frangopol, D. M., Strauss, A. & Kim, S. 2008. Bridge Reliability Assessment Based on Monitoring. ASCE, Journal of Bridge Engineering (in press).

590 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Eigenfrequency estimation for bridges using the response of a passing vehicle with excitation system

Yoshinobu Oshima Kyoto University, Kyoto, Japan Yoshikazu Kobayashi Nichizo Tech Inc., Osaka, Japan Takashi Yamaguchi Osaka City Univesity, Japan Kunitomo Sugiura Kyoto University, Kyoto, Japan

ABSTRACT: We report on the estimation method for fundamental frequencies of a bridge using the response of a passing vehicle. In this method, the bridge, vehicle and truck are thought to be a forced vibration system, where the bridge is excited by the truck periodically vibrated by excitation force, and the vehicle is a receiver to record the acceleration response to the bridge vibration. To increase the acceleration components of the bridge vibration, the truck vibrating at varying frequencies passes the bridge along with the recording vehicle. Since the dominant frequencies of the response are resonant frequencies to the excitation, the spectrum components corresponding to excitation frequencies are extracted to draw Bode diagram, which can indicate the fundamental frequencies as maximum values. Herein the proposed method was evaluated by field testing: a recoding vehicle and truck with excitation passed a test bridge with natural frequency of 3.5 Hz. The vehicle and truck moved at a speed of 20 km/h with excitation frequencies of 2.85 to 11.8 Hz, while the bridge vibration was also measured by the accelerometers installed on the bridge. In this test, time-varying AR model and Wavelet analysis were used to extract the components of the excitation frequencies, because the response is nonstationary and short. The Bode diagram was drawn by the average altitude of the components during the passing time. As a result, it was clarified that the fundamental frequencies were determined by the proposed method, based on the Bode diagram.

1 CONCLUSIONS

The following conclusions can be drawn by this study: • the monitored bridge vibrated with its eigen frequency but close to the exciting frequency by the passing truck • the estimated frequency based on the unsprung mass frequency of the vehicle agreed well with the frequency directly monitored on the bridge • to extract all the low-order eigen frequencies, Bode diagram can be applied: the monitoring vehicle and the exciting truck with several frequencies passed on the bridge several times to obtain the resonance frequencies in the Bode diagram drawn by the estimated frequencies and its intensity.

591 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Monitoring and inspection of a 30 years old prestressed concrete bridge

M. Pimentel & J. Santos LABEST – FEUP,Porto, Portugal J.R. Casas UPC, Barcelona, Spain J. Figueiras LABEST – FEUP,Porto, Portugal

ABSTRACT: In this work a 5 span prestressed concrete box girder bridge exhibiting cracking related pathologies is presented (Figure 1). Built in the late 70’s, this was the first bridge to be built by the balanced cantilever method in Portugal. Although the original design dates back to 1973, the construction only began in 1978. The bridge was opened to traffic in 1980. A detailed non destructive inspection also revealed several prestressing ducts exhibiting corrosion signs, sustaining the assumption of corroded prestressing steel in the box-girder bottom slab. In order to assess its bearing capacity, the bridge was monitored during a load test and during the 4 subsequent days under normal traffic and thermal actions. Shortly after the construction, a consistent longitudinal cracking pattern was observed in the bot- tom face of the deck top slab. In the origin of this cracking pattern was a design error leading to an underestimation of the deck slab flexural reinforcement. The bridge was designed according to the allowable stress method, and the prestressing was calculated in order to guarantee the decompression limit state under load combination equivalent to the characteristic combination of actions of the cur- rent Portuguese code (RSAEP 1983). As a consequence, and according to the current practice, the amount of longitudinal ordinary reinforcement is low. Only after a load test, the bridge was opened to traffic without axle load restrictions. The cracking pattern has been observed ever since and, not withstanding the low reinforcement content, it remains stabilized even after 27 years of service. The paper shows the results of a monitoring campaign and a detailed analysis of the bridge structural behaviour. They were used together to perform the safety assessment of the bridge exhibiting pathologies. The paper focuses mainly on the monitoring results and on the inspection procedures. With the use of a nonlinear analysis model it was possible to show that the failure mode is of ductile nature, and therefore the results of a continuous monitoring system can play an important role regarding the adequate planning of the recommended strengthening intervention according to the convenience of the bridge owner.

Figure 1. Bridge elevation and cross-section.

592 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Boundary condition parameter estimation for structural identification

Yunus Dere & F. Necati Catbas University of Central Florida, Orlando, FL, USA

ABSTRACT: The behavior of existing structures can be best identified when the current physical condition including support boundaries is accurately determined. The structural boundary condi- tions are usually idealized during modeling which may behave different for an existing real life structure. Parameter estimation using experimental data constitutes a reliable approach for veri- fying and updating a finite element model to represent actual structural behavior of an existing structure. In this study, a densely instrumented laboratory beam structure plus a moderately instru- mented grid structure are used to carryout both analytical and experimental studies for parameter estimation. During the parameter estimation process, the effect of changing boundary conditions, the numbers and locations of the sensors, as well as various loading conditions are investigated. For parameter estimation, a simple and practical computer program was developed and utilized. Then, a more sophisticated MATLAB based program, PARIS (PARameter Identification Software) was employed. The results are presented in a comparative fashion.

Figure 1. (a) Beam experimental setup (b) Grid experimental setup.

REFERENCES

Catbas, F. N. & Brown, D. L. & Aktan, A. E. 2004. Parameter estimation for multiple-input multiple output modal analysis of large structures. ASCE Journal of Engineering Mechanics 130(8): 921–930. Francoforte K. 2007. Parameter estimation using sensor fusion and model updating. M.Sc. Thesis, Department of Civil and Environmental Engineering, University of Central Florida, Florida. Frontline Systems, Inc. 2006. http://www.solver.com Fylstra, D. & Lasdon L. & Watson J. & Waren, A. 1998. Design and use of the microsoft excel solver. Interfaces 28(5): 29–55. Kassimali, A. 1999. Matrix analysis of structures. Pacific Grove: Brooks/Cole Publishing Company. 1st Ed. Sanayei, M. & Imbaro, G.R. & McClain, J.A.S. & Brown, L.C. 1997. Structural model updating using experimental static measurements. Journal of Structural Engineering 123(6): 792–798. Sanayei, M. 1998. PARIS© PARameter Identification System, Software Package. Department of Civil and Environmental Engineering, Tufts University, Medford, Massachusetts. Sanayei, M., McClain, J.A. S., Wadia-Fascetti, S., and Santini, E. M. 1999. Parameter estimation incorporating modal data and boundary conditions. Journal of Structural Engineering 125(9), 1048–1055.

593 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Long-term monitoring of stochastic characteristics of a full-scale suspension bridge

H.-B. Yun, S.F. Masri, R.D. Nayeri, F. Tasbihgoo, E. Kallinikidou, M. Wahbeh & R.W. Wolfe University of Southern California, Los Angeles, CA, USA L.-H. Sheng California Department of Transportation (Caltrans), Sacramento, CA, USA

In past decades, numerous vibration-based Structural Health Monitoring (SHM) methodologies have been deployed for civil engineering applications to improve current practices of operation and maintenance for important civil infrastructure. These SHM methodologies commonly involve data acquisition systems that record the response only of the monitored structures, due to dif- ficulties in measuring the excitation of large civil structures. Many identification methods have been developed to deal with the output-only data for nonlinear systems as well as linear ones. The fundamental assumption of these signature-based identification methods is that, if the system characteristics change, the change would be reflected in the response of the system. In practice, however, structural changes are not the only cause influencing the system response, but there could exist numerous other causes, including various environmental effects. In many cases, the environ- mental effects are not negligible, so that understanding the probabilistic properties of the identified structural characteristics, given uncertain environmental conditions, is critical for reliable condition assessment. This paper is focused on demonstrating the long-term monitoring results of the stochastic charac- teristics of a full-scale suspension bridge. TheVincentThomas Bridge, a major suspension bridge in Southern California, is used in this study.The bridge has been equipped with a state-of-the-art, web- enabled, continuous bridge monitoring system since 2004. Using a large data set, acquired by means of twenty-six strong motion accelerometers installed throughout the bridge deck and columns, the bridge was identified to investigate the stochastic properties of the identified structural characteris- tics. The statistical distributions of the identified bridge characteristics are obtained, to understand the effects of various environmental conditions on the identification results. Special events, such as small earthquakes, a ship-bridge collision, a seismic device removal, and traffic-shutdown that happened during this four-year period, were also studied to measure the detectibility of these special structural events in the identification process with uncertain environmental conditions.

Keywords: structural health monitoring, suspension bridge, system identification, stochastic change detection.

594 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

The need, challenges, and opportunities for research and application of Bridge Health Monitoring, a Turkish experience

A. Turer Civil Engineering Department, Middle East Technical University, Ankara, Turkey

ABSTRACT: Bridge Health Monitoring (BHM) is probably the frontier of Structural Health Monitoring (SHM) in civil engineering structures and gaining importance around the globe. BHM applications has a wide variety changing from ordinary but critical highway bridges to special and long span bridges in the form truss, arch, cable-stayed, and suspension types. Large or special bridges are commonly landmarks of cities and nations and deserve special attention. Usage of new materials such as CFRP in bridge industry is also potential applications for BHM due to uncertain long-term performance and behavior. Persistent research and development conducted by the universities and companies on SHM and BHM is now finding its way into pilot studies and starting to be the standard for newly constructed large or important bridges. Many countries in Asia, Europe, and America continents are eager in monitoring their important bridges. The BHM research and application differs from country to country in quantity and quality (or type) based on the Gross National Income (GNI), advancement in research, bridge stock, climate, history, etc. In this paper, a summary on BHM research and application studies and the need for SHM is given from a Turkish perspective and experience. A general assessment was made regarding the bridge type, climate, culture, history on need for research and application in Turkey.

595 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Crack detection in steel bridges

Frank Jalinoos Nondestructive Evaluation (NDE) Center, Federal Highway Administration-TFHRC, McLean, VA, USA Ali Rezai Wiss, Janney, Elstner Associates, Inc., Nondestructive Evaluation (NDE) Center, Turner-Fairbank Highway Research Center (TFHRC), McLean, VA, USA

ABSTRACT: The Federal Highway Administration is currently facilitating development and deployment of nondestructive technologies that are capable of detecting surface or subsurface growing cracks as small as 0.01 inches in length or depth in steel bridge structures. This research focuses on steel girders at fatigue prone, high stress, or critical detail locations where the presence of growing cracks are critical. In addition, locations with high concentration of stress, particularly at steel weldments and sharp corners are of interest. This study is designed to specifically focus on detection of growing cracks in steel bridge structures and thoroughly assess the capabilities and limitations of crack detection technologies. Evaluation of crack detection technologies is a multi-year program that includes laboratory testing of NDE technologies on a series of small-scale flat plate test specimens. The test specimens are fabricated with cracks which include a wide range of crack geometries, orientations, and depths in a range of material thicknesses. Test specimens also include weldments with artificially manu- factured cracks, slags, porosities, and/or lack-of-fusions in the weld. As a result of this program, reliable crack detection systems will be identified and evaluated for accuracy and reliability. The technologies under evaluation include eddy current array, phased array, acoustic emission, and electrochemical sensor.

596 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Systems-based monitoring approaches for improved infrastructure management under uncertainty: Novel approach

F. Necati Catbas Department of Civil, Environmental and Construction Engineering, University of Central Florida, Orlando, Florida, USA Dan M. Frangopol Department of Civil and Environmental Engineering, ATLSS Center, Lehigh University, Bethlehem, Pennsylvania, USA

The objective of this paper is to discuss a novel approach for improved infrastructure manage- ment, utilizing novel sensing technologies, deterioration models and decision making tools. The infrastructure owners and federal agencies in the United States broadly address the need to provide accurate and appropriate information and analyses required to effectively manage and maintain the transportation infrastructure under a complex and demanding set of conditions. The infrastructure owners have to make critical decisions under uncertainty. Currently, many decisions for infra- structure management are based on visual inspections and engineering heuristics. However, the relation between such visible signs of damage and the corresponding “condition” and/or “relia- bility” of the structure is often very difficult to establish. There are several challenges and issues to be addressed when collecting objective data by means of a continuous structural health mon- itoring system for estimating the remaining service life of structures and civil infrastructure by means of performance prediction models. By properly identifying, handling the challenges and addressing the issues, a probabilistic framework can significantly improve the use of the applied knowledge of advanced performance-based condition assessment for transportation infrastructure, mainly for highway bridges. A novel approach should focus on developing a framework to estimate existing condition, predict remaining life for safety and serviceability with significant improve- ments over the current approaches, studies and applications. This approach will be accomplished by means of integrating sensing technologies, deterioration and structural modeling, remote and automated monitoring that will enable maintenance and evaluation of structures in a timely manner. By successful realization of this novel approach, it would be possible to include other transportation network components such as pavements and traffic operations. A novel approach for systems-based monitoring is discussed along with technical strategies.

ACKNOWLEDGEMENTS

The authors would like to acknowledge the support from the U.S. Federal Highway Administration (FHWA). This material is based on the work supported by the Federal Highway Administration under Cooperative Agreement DTFH61-07-H-00040. Dr. Hamid Ghasemi of the FHWA is greatly appreciated for his support and feedback to the authors for this study. Any opinions, findings and conclusions or recommendations expressed in this publication are those of the authors and do not reflect the view of the Federal Highway Administration.

597

Seismic and dynamic analysis

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Prioritization and seismic risk assessment of bridges

D. Cardone & G. Perrone DiSGG University of Basilicata, Potenza, Italy M. Dolce Department of Civil Protection, Roma, Italy L. Pardi Autostrade per l’Italia S.p.A., Roma, Italy

Bridges and viaducts are the critical elements of a road network, due to their own characteristics and the considerable consequences, in terms of both repair costs and circulation problems, implied by their degradation and damage. Recent earthquakes have repeatedly demonstrated the seismic vulnerability of existing bridges, due to their design based on gravity loads only or inadequate levels of lateral forces. The slow degradation of the bridge structures (e.g. due to corrosion of reinforcing steel bars) can significantly change their strength and ductility, thus increasing the seismic vulnerability of the bridge as a whole. The seismic zonation of many European countries, moreover, has been recently revised, prescribing more severe peak ground accelerations in several regions, thus further emphasizing the seismic inadequacies of many bridges and viaducts, due to the gravity load based design. In this paper, two different approaches for the seismic risk assess- ment of highway bridges are proposed. The first approach is especially suited for the prioritization and screening of a large stock of bridges. The seismic risk derives from the combination of three components: (i) the seismic hazard of the site, (ii) the seismic vulnerability of the construction and (iii) its exposure or importance. To quantify the aforesaid three components, suitable ratings are assigned to a number of characteristics of the bridge, which affect its seismic performances. The seismic hazard is estimated based on the seismic zone and soil conditions at the bridge site. The vulnerability is evaluated by considering 12 different modes of crisis, corresponding to dif- ferent damage/collapse mechanisms for each structural element of the bridge (Foundations, Piers, Abutments, Bearings and Joints). The exposure of the bridge is evaluated with reference to the consequences that the expected earthquake can produce in terms of human losses and interrup- tion of service of the highway and of all the economical and social activities for which the full serviceability of the bridge, in normal running conditions and/or in a post-earthquake scenario, is required. The second approach has been developed to be applied to a limited number of bridges, pre-selected based on the first approach. It is inspired to the principles of the Capacity Spectrum Method, reviewed under an adaptive perspective. The seismic resistance of the bridge is evaluated through an adaptive pushover analysis. The pushover analyses for the evaluation of the global non- linear (monotonic) behavior of the bridge is carried out in the longitudinal and transverse direction separately. First of all, a moment-curvature analysis of each pier section is performed. After that, the force-displacement behavior of each pier is derived. The force-displacement relationships of the pier-bearings systems are then obtained, by summing the displacements of pier and bearings under the same horizontal force. Finally, the response of the bridge in each direction (longitudinal and transversal) is examined through “adaptive” pushover analysis. The first output of the accurate procedure are a series of fragility curves, which describe the seismic vulnerability of the bridge associated to different Performance Levels (PL’s), under a probabilistic point of view. The seismic risk is evaluated with the use of hazard maps, known the GPS location of the bridge site. The hazard maps provide the PGA values at the bridge site having a given probability of exceedance in a given interval of time (e.g. 50 years), i.e. a sort of hazard curve. The seismic risk for a selected PL is 601 obtained by calculating the convolution integral of the product between the seismic vulnerability of the bridge (expressed by the fragility curves) and the seismic hazard of the bridge site (expressed by the hazard curve). In the paper the main aspects relevant to both approaches are presented. Par- ticular attention is focused on the basic assumptions and most important steps of the two numerical procedures.

602 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Ambient vibration test and seismic evaluation of steel-deck truss bridge

S. Jung Structural Dept., Pyunghwa Engineering Corp., Anyang, Gyeonggido, Korea S.-T. Oh & S. Kim School of Civil Engineering, Seoul National University of Technology, Seoul, Korea Y.-W. Shim Jungam Engineering & Construction Corp., Seoul, Korea S.S. Chen Department of Civil Structural and Environmental Engineering, State University of New York, NY, USA

This study describes an analytical and experimental investigation of the pedestrian steel-deck truss bridge in the City of Rochester, New York, USA (Figure 1). An ambient vibration experiment on full-scale structures is a way of assessing the reliability of the various assumptions employed in the mathematical models used in analysis. It is also the most reliable way of determining the

Figure 1. Evaluation view of Pont de Rennes.

Figure 2. Capacity spectrum curve.

603 structural parameters of major importance in structural dynamics. In the analytical modeling, three- dimensional finite element analysis was used. In this investigation, pedestrian-induced vibrations have been measured on the bridge to determine the displacement and the vertical and transverse vibration characteristics of the steel-deck truss. Experimentally obtained natural frequencies and mode shapes have been compared with the analytical results. With the resulting analytical three- dimensional finite element model, non-linear static (pushover) analyses are conducted (Figure 2). After seismic vulnerability assessment, the results show that end sway frame, bottom lateral system, and cross bracing systems are vulnerable.

604 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Finite element model updating of a concrete arch bridge through static and dynamic measurements

H. Schlune, M. Plos, K. Gylltoft, F. Jonsson & D. Johnson Chalmers University of Technology, Structural Engineering, Göteborg, Sweden

ABSTRACT: Finite Element (FE) analyses are increasingly used to make decision whether strengthening, repair or replacement of bridges are needed. These decisions can only be as good as the results from the analyses, and therefore a reliable and accurate model is decisive for good assessment and management. FE models verified and updated through measurements performed can fulfil this demand. This paper describes how field test and monitoring results were used in a combination of manual model refinements and non-linear optimization to obtain a more realistic FE model of one of the world’s largest single arch bridges, the New Svinesund Bridge. Measured strains, forces and deflections during a load test and measured eigenfrequencies under ambient excitation were used to update the initial model of the bridge. Manual model refinements included among other things, a stiffness increase of the concrete arch, a consideration of the asphalt concrete layer and a more realistic representation of the bearings, which especially improved the agreement of the FE model with the measurements. Before automatic FE model updating was performed, a parameter study was conducted to determine the sensitivity of the bridge response to each uncertain structural parameter. That allowed to exclude some possible updating parameters from the automatic updating algorithm and made it possible to assign good stating values for the remaining updating parameters. The remaining parameters were than updated using the downhill simplex method, see Nelder and Mead (1965), which led to further improvements of the model accuracy. It was shown that different types of measurements are needed for updating a FE model that should be generally improved for dynamic and static analysis. Manual model refinements turned out to be indispensible before model updating through non-linear optimization. The final updated model could reproduce all measured parameters with significantly improved accuracy.

REFERENCES

Nelder J. A. and Mead R. 1965: A simplex method for function minimization. The Computer Journal, Vol. 7, No. 4, pp. 308–313.

605 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Seismic response of highway viaducts under design live load considering vehicle as a dynamic system

Mitsuo Kawatani, Chul-Woo Kim, Shimpei Konaka & Rie Kitaura Department of Civil Engineering, Kobe University, Nada, Kobe, Japan

ABSTRACT: This study is intended to investigate the seismic response of a steel plate girder bridge under vehicle loadings during moderate earthquakes of high probability to occur. The high- way bridge design codes of Japan Road Association (JRA code) do not consider the live load in the seismic design of highway bridges because of the low possibility of the event that both of the live and seismic loads occur at the same time. However, frequent traffic jams in urban areas indicate a high possibility to encounter an earthquake during rush hour, but seismic response of bridges during a traffic jam has not been fully investigated yet. Two simply supported steel girder bridges, which are parts of a viaduct, with steel bearings are adopted as analytical bridge model as shown in Figure 1. Each bridge consists of five girders. The offset beam is applied to consider the geometric difference among the neutral axes of girders, those of bearings and pier caps. An internal element is used for bearing at each support. The pier bottom is assumed to be fixed. Two Level 1 ground motions of stiff and moderate soil sites from the JRA code are used in the seismic response analysis. The ground motion in the vertical direction is also considered during the analysis. To investigate the effect of vehicle’s dynamic system on the seismic response of the steel plate girder bridge, following four scenarios are investigated: CASE1 is the bridge model without con- sidering design live load; CASE2 is the bridge model considering vehicles occupying two lanes as additional mass; CASE3 is the scenario considering the interaction between a bridge and stationary

Figure 1. Bridge model.

606 Figure 2. Design live load.

Figure 3. Bar graph of RMS and peak values of seismic responses of bridge. vehicles also occupying two lanes; CASE4 is the scenario considering the interaction between a bridge and moving vehicles also occupying two lanes. During analysis, the running speed of the design live load is assumed as 10 m/s. Vehicle is idealized as the dynamic system with 12-DOFs. The CASE2, CASE3 and CASE4 are the situation of traffic jam. Figure 2 shoes the design live is considered in the analysis. Comparison between effect of vehicles dynamic system and the bridge model disregarding vehicles shows that the tendency to reduce the bridge response under the ground motion observed on moderate soil sites. On the other hand, under the ground motion taken from stiff soil sites, the vehicles on the bridge tend to amplify the seismic response of the bridge (see Figure 3).

607 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

A substructures approach in the dynamic analysis of continuous beams under moving oscillators

V.De Salvo & G. Muscolino Department of Civil Engineering, University of Messina, Italy A. Palmeri School of Engineering, Design & Technology, University of Bradford, UK

When dealing with the vehicle-induced vibration of continuous bridges, the dynamic excitation is usually modelled through moving forces or masses, in so completely neglecting the bridge- vehicle interaction. On the other hand, this phenomenon should be carefully taken into account, since it is directly related to many engineering issues, e.g. fatigue damage of bridges or riding comfort of vehicles. Unfortunately, most of the conventional approaches available in the liter- ature exhibit severe theoretical and practical limitations, so that their extension to cope with realistic cases of dynamic interaction between running vehicles and continuous bridges is quite problematic. In this contribution, aimed at overcoming these limitations, a novel technique for studying the dynamic response of multi-span beams traversed by moving vehicles, modelled as SDoF oscillators of given mass, stiffness and damping, is proposed and numerically validated. In order to cover the most frequent cases of engineering interest, the continuous beam in our formulation is inhomoge- neous, and the position of the intermediate supports is arbitrary, while the oscillator can accelerate and decelerate along the beam. The structural damping is also included. The basic idea is to take advantage of the Component Mode Synthesis (CMS) method, which in the past proved to be very versatile in treating continuous structures coupled with moving systems. To do this, in a first stage the continuous beam is ideally decomposed in a series of alternate primary and secondary spans, for which convenient boundary conditions are selected; the compatibility at the interface between adjacent spans is then re-established, and a set of efficient shape functions for the continuous beam as a whole body is defined. In a second stage, starting form the total kinetic and potential energies of the beam-oscillator dynamic system, the Lagrange’s equations of motion are derived. In a third stage, finally, the equations of motion are arranged in a compact state-space form, which is particularly efficient for numerical implementations. Among the advantages, the proposed technique lends itself to be easily automated, since the approximate modal shapes of the continuous beam are deduced from the well known eigenfunctions

Figure 1. Three-span haunched beam considered in a numerical application.

608 of pinned-pinned, pinned-clamped and clamped-clamped homogeneous beams. Moreover, being based on a substructure approach, the extension to more sophisticated models of vehicle is quite straightforward. The numerical applications, e.g. to the bridge structure depicted in Figure 1, confirm that accurate predictions of the dynamic response may require to include the interaction between running vehicles and continuous bridge.

609 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Improvement of seismic analysis concerning the characteristic difference of HDR-S between the stages of design and inspection

Hishijima Yosuhisa Kawaguchi Metal Industry, Inc, Japan Moon-Sig Yoo K E&C, Inc, Korea Dong-Ho Ha Konkuk University, Korea Ki-Young Kim K E&C, Inc, Korea

This paper describes the reliability of the seismic isolators employed in Machang Bridge. Machang Bridge, which is to be opened this year (2008), employs a seismic isolation device, the HDR-S (High Damping Rubber bearing – Super). HDR-S is a new version of HDRB (High Damping Rubber Bearing). HDR-S, with high damping ratio of LRB(Lead Rubber Bearing) and durability of HDRB, has been widely used recently. To validate the reliability of HDR-S, the bridge was re-analyzed according to HDR-S’s inspection characteristics. Then, the results analyzed with the inspection values were compared with the original design results. To confirm and validate the actual safety of manufactured seismic isolation system, such a procedure should be performed. In Machang Bridge, due to the higher inspection damping ratio than the design damping ratio, the re-analyzed reaction results declined by 4%∼9% from the design results. In addition, the scope of HDR-S’s characteristic tolerance was compared using Guide Specifi- cations for Seismic Isolation Design by AASHTO (American Association of State Highway and Transportation Officials). The risk due to the variation of the seismic isolation device’s charac- teristics was analyzed and assessed through the bridge. There was a slight difference, only 0.4%, between design results and analysis results using the tolerance. The slight difference from the design results could reduce the excessive margin of a sub-structure. Consequently the regulations for HDR-S would be extremely fastidious. And finally, the performance of HDR-S and HDRB was compared, to determine the relative mer- its of HDR-S. Damping ratio is 20.5% for HDR-S and 10%∼15% for HDRB. The seismic response characteristics of a bridge with HDR-S and HDRB was also introduced. The result indicates HDR-S has proven to be a more effective seismic isolation system than HDRB.

610 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Vulnerability assessment of an existing highway bridge by 3-D nonlinear time history analyses and proposing its retrofit design

M. Hosseini International Institute of Earthquake Engineering and Seismology (IIEES), Tehran, Iran S.R. Khavari Science and Research Branch of the Islamic Azad University (IAU), Tehran, Iran

Various techniques have been proposed by researchers for seismic retrofit of bridges, among them two ones have been used widely in recent years: 1) altering the bearing fixities, which can result in elimination of the need for costly retrofitting of substructures, and 2) using braces as retrofit elements for decreasing the lateral or longitudinal motions. In this paper the seismic vulnerability of an existing two span reinforced concrete slab bridge has been studied for proposing its retrofit design. In this paper the seismic vulnerability of an existing two span Reinforced Concrete (R/C) slab bridge with round columns has been studied. At first, the bridge design has been checked by the seismic provisions of the new version of AASHTO (2000) code, by which the weakness of bridge in bending resistance and particularly shear resistance of its piers columns has been realized. In the next step, regarding the high Demand over Capacity Ratios (DCRs), some Push Over Analyses (POA) have been performed, by which the ultimate resistance as well as the rotations of plastic hinges in columns have been obtained. The results have confirmed the high seismic vulnerability of the bridge. On this basis, two retrofit designs: 1) adding lateral bracings between piers columns, and 2) fixing the lateral deck-to-abutment connection, both along with adding diagonal members in longitudinal direction have been proposed. The adequacy of both retrofit designs has been shown by some Nonlinear Time History Analyses (NLTHA) by using some accelerograms, which have frequency contents compatible with the conditions of the bridge site. Based on the results of POA and NLTHA it can be concluded that: • The existing bridge is highly vulnerable, and the main cause of its vulnerability can be claimed to be the slenderness of its columns on the one hand, and the low ratio of steel bars in the columns on the other. • Of the two proposed retrofit designs the second one which leads to lower lateral displacements and does not need any bracing element in the lateral direction seems more economical and practical. However, in this case the deck-to-abutment connections, which should be fixed to prevent the lateral displacements, may need special design considerations. It should be noted that there are some other retrofit techniques such as increasing the columns’ cross-sectional areas and steel bars, using FRP, using seismic isolators and so on, which may be more appropriate for the studied bridge, however, to realize the most economical and practical retrofit technique some more investigations are necessary.

611 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Shake table studies of scaled reinforced concrete bridge piers subjected near-fault ground motions

Y.-S. Chung, C.-Y. Park & H.-K. Hong Chung-Ang University, Ansung-Si, Korea D.-H. Lee Gyeongdo Provincial College, Yechon-Gun, Korea C.-S. Shim Chung-Ang University, Ansung-Si, Korea

Since 1995, a comprehensive study of the seismic response of conventional and composite rein- forced concrete bridge piers has been in progress at Chung-Ang University. Previous experimental investigations for the seismic behavior of concrete bridge piers have been evaluated by the Quasi- static test and the pseudo-dynamic test considering far-fault ground motions. As shown Figure 1, this research aims to investigate the seismic response of Reinforced Concrete (RC) bridge piers subjected to impulsive near-fault ground motions by the shaking table test, four reinforced concrete (RC) bridge piers with a diameter of 400 mm and height of 1600 mm have been constructed for the shake table test and been evaluated under near-fault ground motions. One reference RC bridge piers have been comparatively made for the seismic response by the pseudo-dynamic test. Test parame- ters are lap-splice of longitudinal reinforcing steels, the space of transverse reinforcing steels, and test method. Test result showed that RC specimens under the shake table test failed at relatively low displacement ductility, compared to that of RC bridge piers subjected to the pseudo-dynamic test, and that large residual displacement was observed under impulsive near-fault ground motions, and that RC piers with lap-spliced longitudinal reinforcement exhibited much lower displacement ductility.

Figure 1. Test setup.

612 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Computer wind investigations for long bridge crossings

D. Janjic & A. Domaingo TDV Technische Datenverarbeitung Dorian Janjic & Partner GmbH/Bentley Systems, Graz, Austria

ABSTRACT: Wind impact is becoming more and more important for safety considerations of bridges, especially if crossing large rivers, estuaries or sounds with larges spans. These structures are usually very slender and therefore very susceptible to wind induced vibrations. Many critical aero-elastic phenomena have been observed on existing bridges, and considering the risks in the design phase, as well as continuously observing related phenomena throughout life- time, have become a must for all major bridge structures. These phenomena include vortex shedding and the lock-in phenomenon, across-wind galloping and wake galloping, torsional divergence, flutter phenomena and wind buffeting etc. Being able to deal within one software package with all related problems eases the required investigations. Once an appropriate mathematical model has been established, respective analyses for design purposes or for verifying observed phenomena and calibrating the model assumptions can be easily performed at any time. This ability was recently included in a renowned structural engineering program. The first step is to determine the relevant aerodynamic coefficients with an integrated CFD func- tion. Up to now, these coefficients have typically been measured in time consuming wind tunnel tests. Extensive comparisons of the CFD results with those of wind tunnel tests were used to cali- brate the used discrete vortex method (DVM). A typical set of steady state coefficients is shown in Figure 1. The CFD analysis also allows for investigating the vortex shedding phenomenon. The pro- gram calculates the vortex shedding frequency as a function of the wind speed, and comparing it with the natural frequencies yields the critical wind speed value. Within further CFD runs with dynamic cross sections flutter derivates can be calculated. The full set of wind data can thus be prepared based on numerical simulation. The aero-elastic stability checks can be evaluated and reported. The wind buffeting analysis is based on the above-mentioned aerodynamic coefficients and their derivatives, and on the relevant wind profile parameters. A full set of statistical wind data contains speed, direction, turbulence intensity, power spectrum and coherence data. The analysis includes aeromechanical admittance functions to define interaction between bridge and wind. The results of a wind buffeting analysis are presented in Figure 2. Aero elastic damping and stiffness are fully included in analysis. The outlined solution procedures are implemented in commercial software package.

Figure 1. Steady state coefficients of Hardanger Figure 2. Tower and deck bending moment of bridge. Stonecutter bridge due to wind buffeting.

613 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Identification of the dynamic characteristics of long span bridges using ambient vibration measurements

A.L. Hong & R. Betti Department of Civil Engineering and Engineering Mechanics, Columbia University, New York, NY, USA

ABSTRACT: In this paper, the results of a study on the identification of the dynamic character- istics of long-span bridges using ambient vibration measurements are presented. The methodology used in this study is based on the data-driven Stochastic Subspace Identification (SSI) algorithm, which identifies a stochastic state-space model by projecting the future output onto the past output. The SSI technique can be adjusted to different formulations such as Principal Component algorithm (PC), Canonical VariateAlgorithm (CVA),and Unweighted Principal Component algorithm (UPC) by using proper sets of weighting matrices. For a comparative study of these three formulations, their theoretical differences were first studied and then their effectiveness in the identification of real structures was tested using ambient vibration measurements of two major suspension bridges, the Vincent Thomas Bridge (VTB) in Long Beach, CA and the New Carquinez Bridge (NCB) in Vallejo, CA. In the identification analysis, the stabilization diagram that represents the identified frequencies as a function of the model’s order was used. In the state-space system identification using noise-polluted measurements, such a diagram plays a crucial role in highlighting struc- tural modes against numerical modes in an overestimated dimension of the order of the system’s model. Although higher-order dynamic models could provide more modes being stabilized in the stabilization diagram, an excessively large system order is not desirable in practice because the computational effort and the numerical errors grow quite rapidly with the system’sorder. This study focuses on the capacity of the PC, CVA, and UPC methods to identify modal parameters (natural frequencies, damping ratios, and mode shapes) of the two bridges with a state-space dimension of less than 100. The difference between the PC and UPC methods is represented by the orthonormal constraint on the state. It was found, from the comparison of the PC and UPC methods, that the orthonormal constraint on the state does not provide any improvement in describing the structural dynamic behavior for the modal parameter identification while does make the UPC method more sensitive to measurement noises. In the comparison between the CVA and UPC methods, the ver- tical and transverse vibration analyses of the VTB and the vertical vibration analysis of the NCB showed that the process of normalizing the future output, which is involved in the CVA method, caused an increase of the state dimensions for the structural modes to be stabilized. As a result, the PC method, where neither of the constraints is held in defining the state, showed the best performance among the three methods. In the transverse vibration analysis of the NCB, however, the CVA method performed better than the other two methods. Additional analyses are ongoing to explain this discrepancy within a more general framework.

614 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Application of indexing and detailed seismic risk assessment approaches to existing bridges

D. Cardone & G. Perrone DiSGG University of Basilicata, Potenza, Italy L. Pardi Autostrade per l’Italia S.p.A., Roma, Italy

The Italian highway network consists of almost 5000 bridges, mainly built between 1960 and 1980. The seismic safety of the majority of the existing bridges is rather uncertain, being based on old seismic codes, relying on the elastic design philosophy. Recent earthquakes clearly pointed out the seismic vulnerability of existing bridges, especially due to piers and bearing devices. In order to make a rational decision about the need of retrofitting (or not) an existing bridge, reliable tools for the seismic evaluation of bridges are needed. They should relate the seismic risk of the bridge to given structural Performance Levels (PL’s). Current approaches in seismic vulnerability and risk assessment of existing structures (buildings and bridges) can be divided in three main groups, based on their level of complexity. The first, most simple, level is based on individual judgement. It does not require any analysis and its goal is to determine the priority levels of structures that require retrofitting interventions. The second level methodologies are applied when a more accurate evaluation of a building/bridge stock is required. They are based on simplified linear analyses performed following different approaches. The third level methodologies employ nonlinear (static or dynamic) analysis of the structure, thus requiring detailed information on the as-built characteristics of all their structural members. Two procedures for the seismic risk assessment of bridges have been developed within the SAGGI (Advanced Systems for the Global Management of Infrastructures) research project, in which the DiSGG of the University of Basilicata is involved, for what concerns the seismic aspects of the project. The first proposed procedure is based on the assignment of proper ratings to different characteristics of each structural element (piers, abutments, bearing devices, etc.). The second procedure is based on the use of Fragility Curves, associated to given PL’s of the bridge, which are then combined with a representation of the seismic hazard of the bridge site. The first procedure is primarily aimed at screening and prioritizing a large stock of bridges. The second procedure is mainly addressed to an accurate assessment of the bridges with higher risk indexes. The backgrounds and most important features of the two procedures are described in a companion paper presented at the same Conference. In this paper, the two approaches are applied to a number of existing bridges and the results compared. A set of 10 existing bridges of the A16 Napoli-Canosa Italian highway have been selected as case-study. The selection resulted from a preliminary examination of the A16 Highway, aimed at identifying bridge types and characteristics really representative of the whole bridge inventory of the A16 Highway. All the selected bridges are multi-span simply supported bridges, with span lengths of approximately 33 m. All of them were built between 1969 and 1971, according to the pre-1971 Italian Regulations for RC structures. As a consequence, though being built in seismic areas, they were not designed to resist any horizontal action. In addition, neither specific design criteria (capacity design) nor specific rules for seismic detailing were adopted. The comparison of results shows a good accordance (at least qualitative) between the two procedures: bridges with higher probability of collapse according to the accurate procedure occupy the first positions in the ranking of the simplified procedure and vice-versa: bridges with lower probability of collapse in the accurate procedure occupy the last positions in the ranking of the simplified procedure. This

615 clearly proves the suitability of the ratings assigned to the different characteristics and categories of the bridge within the simplified procedure. Obviously, some differences are found for structures with same types of piers and bearing devices. As a matter of fact, the indexing procedure cannot distinguish structures having same types of piers and bearing devices. Similarly, some discrepancies arise from the different accuracy of the two procedures in the evaluation of the seismic hazard. To conclude, the proposed indexing procedure can be reliably used for prioritizing and screening a great stock of bridges. The proposed accurate procedure can be subsequently used to examine a pre-selected limited number of bridges, in order to make a rational decision about the need of retrofitting or replacing them.

616 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Effectiveness of rupture controllable steel side blocks for elevated girder bridges with isolation bearings

N. Asada, M. Matsumura & T. Kitada Osaka City University, Osaka, Japan M. Sakaida Teikoku Engineering Consultants Inc., Gifu, Japan M. Yoshida Kawaguchi Metal Industries Co., Ltd., Osaka, Japan

1 INSTALLATION OF RUPTURE CONTROLLABLE STEEL SIDE BLOCK

Steel side blocks, which are generally designated a joint protector to restrain the transverse displace- ment of an isolated viaduct under consideration, are set near both sides of the isolating bearings. In considering more effective and rational use of the side blocks, the side blocks can be designed to have the following two functions; the joint protector against the Level 1 Earthquake with the maximum acceleration of 150–200 gals at the surface level of the ground as well as a knock-off member to provide isolating effect against the Level 2 Earthquake like the Hyogo-ken Nambu Earthquake. Then, a slit type steel side block (CSB), which surely breaks at the intended strength, is developed by the authors (Sakaida et al., 2004) and the shape and the breaking procedure of CSB is illustrated in Figs. 1 and 2. The CSB is designed by Eqs. (1) and (2).

Figure 1. Breaking procedure of CSB.

Figure 2. Shape of CSB with symbols for design formulae.

617 The breaking procedure of CSB is as follows; (a) CSB is set aside the top plate of the laminated rubber bearing, (b) and (c) when the top plate on the bearing shifts small, CSB deforms elastically and restrict the horizontal displacement of the superstructure against the Level 1 Earthquake, (d) when the top plate shifts large against the Level 2 Earthquake, CSB deforms and (e) CSB breaks at the slit when working shear stress in the connected part predominates and then isolated condition is promptly provided to the viaduct.

2 STATIC AND DYNAMIC BREAKING TEST

Static breaking test was carried out by using a downscaled specimen of the practical CSB, 1/2 scale of an actual dimension. Test results indicate that the breaking load of the CSB specimens take almost the same value with the design break load. The fluctuation of the breaking load among 4 CSB specimens of the same shape is just within about 10% from the design breaking load. Then, the CSB with the shape of (A − C)/A = 0.85, A/B = 4 and C/B = 0.6, is recommended in controlling the breaking load calculated by the design formulae. Dynamic breaking test was carried out by using 6 specimens of CSB to verify the influence of the loading speed on the breaking load. The breaking loads in the dynamic loading test increase about 10% than that in the static breaking test and the breaking displacement of the CSB decreases as the loading speed increases. Then, the correction factor considering dynamic loading effect, β in the design formulae of the CSB and the value is decided 0.9.

3 SHAKING TABLE TEST

A shaking table test of a system, consisting of a superstructure, a rubber bearing and a bridge pier, is carried out to confirm the application effect of the CSB to a viaduct experimentally. The specimen setting on the shaking table is shown in Fig. 3. A gradually increasing sinuous horizontal displacement is input to the shaking table. The frequency of the input displacement is the same as the natural frequency of the system and the amplitude is gradually increased to 1 mm. As shown in Fig. 4, the horizontal displacement is not become large in case of no_SB, without the knock-off function. When SB, the side block without breaking, is applied, the horizontal displacement increases gradually as the amplitude increases. The similar tendency is observed up to the breaking of the CSB but after the breaking the horizontal displacement decreases. Therefore,

Figure 4. Time history of horizontal displace- Figure 3. Specimen set up on shaking table. ment of pier.

618 the side block with knock-off function like the CSB can provide isolated condition promptly after the breaking of the CSB.

4 CONCLUSIONS

Design formulae of the CSB are proposed by referring to the static and dynamic loading test results. A correction factor, β = 0.9 is employed into the design formula to consider the loading speed influence on the breaking load. Dynamic behavior of a viaduct with isolation bearings and the CSB and the effectiveness of the CSB are verified through the dynamic loading test and the seismic response analysis. It can be concluded that the CSB can be designed as a joint protector against the Level 1 Earthquake and as a knock-off member against the Level 2 Earthquake to provide isolating effect in the transverse direction. The contacts and/or collisions between the expansion joints should be considered in further investigation.

REFERENCE

Sakaida, M.,Yoshida,M., Matsumura, M., & Kitada, T.2004. Static BreakingTestof Rupture Controllable Side Blocks of Bridge Bearings, Proceedings of the Second International Conference on Bridge Maintenance, Safety and Management, IABMAS04, Kyoto, Japan, 393–395.

619 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Influence of bullet train as dynamic system on seismic performance of Shinkansen viaducts

M. Kawatani Department of Civil Engineering, Kobe University, Kobe, Japan X. He Research Center for Knowledge Networking, Kansai University, Osaka, Japan K. Shinagawa Graduate School of Engineering, Kobe University, Kobe, Japan S. Nishiyama Nikken Sekkei Civil Engineering, Ltd., Tokyo, Japan

1 INTRODUCTION

Japan is located in an earthquake-prone region. The earthquake-proof capacity of Japan’shigh-speed railway system, so-called Shinkansen, is always a concern considering the extremely high-speed of the bullet trains. The Kobe earthquake, which occurred on Jan. 17, 1995, brought severe damage in the southern Hyogo Prefecture. The railway viaducts were collapsed by the earthquake. The bridge structures, especially the piers, were damaged severely in a wide range. Considering the earthquakes expected to occur predictably in the future, it is urgent to examine the seismic performance of the existing bridge structures. Although the live load of train is considered in the Seismic Design Code for Railway Structures (RTRI 1999) in Japan, the train is merely treated as subordinate variable load to the bridge structure. The upper limit of the train’s response for arbitrary ground motions is specified, considering the possibilities that the train may act as damper and the train can vibrate with different phase from the bridge. However, such specifications are rather ambiguous for general cases. It is not completely rational to treat the train merely as additional mass to the bridge because the train is a complicated dynamic system. In the recent performance-based design practice, to satisfy both safety and economy demands, the dynamic effect of the train on the bridge response subjected to earthquake should be investigated further. In this study, assuming that the structure remains in elastic domain during moderate earthquakes, an approach to examine the seismic performance of the Shinkansen bridge-train interaction system is established. The seismic performance of the pier is confirmed by comparing the cross-sectional forces resulted from the earthquake load with the maximum strengths of the pier. The dynamic effect of the train on the seismic performance of the bridge is also evaluated.

2 ANALYTICAL PROCEDURE

Viaducts with track structures are modeled with three-dimensional (3-D) finite elements. A 15-degree-of-freedom (15-DOF) sprung-mass dynamic bullet train model that can appropriately express the lateral, vertical and rotational motions of the car body and bogies is developed for seismic analysis. The dynamic differential equations of the bridge system are derived using modal analysis technique (Kawatani et al. 2000, Kim et al. 2005). The earthquake excitation 620 is considered as inertial force acting simultaneously on all DOFs of the bridge and train models. Newmark’s β direct numerical integration method is employed to solve the coupled differential equations of the bridge-train-earthquake interaction system. The parameter β, which controls the variation of acceleration within the time step, is set as 1/4. The rate of convergence is set as 0.001 for the acceleration response to insure the analytical accuracy. The structure is assumed to remain in elastic domain during a moderate earthquake. To examine of seismic performance, the maximum capacities of the pier should be calculated at first. In this study, the maximum strengths for the bending deformations of the pier are calculated according to the specifications of the Seismic Design Code for Railway Structures (RTRI 1999), which will be referred to as the seismic design code.

3 EVALUATION OF SEISMIC PERFORMANCE

To examine the dynamic effect of the train on the seismic response of the bridge, four analytical cases are chosen for the seismic analyses. For all cases, both EW and UD components of four ground motions are applied. For the four analytical cases used above, the bending moments and the flexural strengths of the basal part of pier L1 are shown in Figure 1. In all cases they are larger than the flexural strengths for cracking of concrete. Under Level 1 ground motion: In Case-1 and Case-2, the bending moments exceeded the strengths for yielding of reinforcing bars, while they are smaller than the limit states of failure. Under Ground motion 1: In all cases, the cross-sectional forces are larger than the maximum capacities. The results in this analysis indicate that even for rather moderate ground motions, according to the relation between the predominant frequency components and the natural period of the bridge, the seismic response of the bridge may be unexpectedly intense. Therefore for such cases, due to the evaluation of the seismic performance of the bridge-train interaction system, it may be necessary to consider countermeasures to reduce the bridge response, such as to raise the stiffness of the piers or reinforce the bridge by steel damping braces.

Figure 1. Bending capacity examination.

621 4 CONCLUSIONS

In this study, an analytical approach to evaluate the seismic performance of the Shinkansen viaducts subjected to running train loads is established. Then the seismic responses of the bridge are simu- lated and evaluated. The seismic performance of the pier is evaluated by comparing the maximum strengths with the cross-sectional forces under three ground motions. This study also laid a foundation for further research on the seismic performance of Shinkansen bridge-train system, such as to reveal the mechanism of bridge collapse under strong earthquake motions, evaluate the running safety of the bullet train, or discuss countermeasures against excessive seismic response of the bridge structure, etc.

622 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Vibration-based tension identification of ultra long stay cables

Jianfeng Liu & Ning Fang Department of Bridge Engineering, Tongji University, Shanghai, China Qiwei Zhang State Key Laboratory for Disaster Reduction in Civil Engineering, Tongji University, Shanghai, China

Tension estimate of ultra long stay cables is an important issue during the construction and service stages of long span cable-stayed bridges. Stay cables are susceptible to environmental excitations. Considering the advantages and convenience of application, the technique of tension estimate through the measurement of vibration of the cables is particularly focused. Based on the measured natural frequencies, four main methods for cable tension estimate are developed according to the consideration of the sag-extensibility and bending stiffness or not. The first is based on the vibration chord theory and ignores the effects of sag-extensibility and bending stiffness. The second method deduced by Irvine takes account of the sag-extensibility without bend- ing stiffness. The third method considers the bending stiffness and neglects the sag-extensibility. The fourth method takes sag-extensibility and bending stiffness into account simultaneously. In this paper, vibration acceleration responses were measured on site from the cables of the Sutong Bridge in China, which is the longest cable-stayed bridge with a main span of 1088 meter. The selected two cables’ material and geometry properties are listed in the following table. The two ultra long cables’measured natural frequencies were identified by peaking picking method. Higher order natural frequencies of the ultra long cables can be easily identified. The finite element method was applied to the cables’analysis when the stress of the cables is 20 and 50 percent of their limiting stress. Combining the experimental measurement results and the finite element analysis results, it can be drawn that natural frequencies of the cables increase with the mode number by approximate integer multiples except the first-order in-plane frequency when the stress of the cables is relatively lower. Four aforementioned theoretical methods are used for the cable tension estimate and the identified results are compared with each other. The discrepancies between estimate results from different methods are all less than 1%. Therefore, the tension of ultra long span stay cables can be identified by the referred approaches with measured vibration frequencies.

Table 1. Material and geometry properties of the identified cables.

Cable Young’s Horizontal Height Diameter Mass Angle of No. modulus(MPa) length(m) (m) (m) (kg/m) inclination(◦)

A34 1.95 × 105 489.915 232.725 0.124 99.92 25.41 J34 1.95 × 105 533.950 221.304 0.124 99.92 22.51

623 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Calculation of the influence line of a bridge using a moving vehicle

A. González & E. J. OBrien University College Dublin, Dublin, Ireland

ABSTRACT: The influence line of a bridge depends on the load effect being sought, the section under study, the travelling path of the moving load and the geometry, boundary conditions and stiffness of the bridge structure. Its shape and magnitude characterizes the static behaviour of the bridge and it can be employed in a number of applications: collection of traffic data using a bridgeWeigh-In-Motionsystem, determination of the characteristic load effect, soft load testing and bridge health monitoring among others. The influence line can be estimated crossing a vehicle of known weights over the bridge if the vehicle axles are accurately located on the bridge at each point in time. The procedure is based on a least squares fitting technique that takes into count linearity and the principle of superposition of the individual contributions of each axle to the total static response. But bridge measurements contain additional dynamic and noise components that make them differ from the theoretical total static response, and they prevent an exact calculation of the influence line. In the case of a stiff short-span bridge with a smooth surface, the dynamics caused by a moving vehicle on the bridge response will be negligible and the ordinates of the influence line can be approximated using an optimisation procedure. Nevertheless, there are cases where the dynamic component contained in the signal can not be neglected, i.e., medium and long-span bridges or short- span bridges with a rough surface. In these cases, if a least-squares procedure based on superposition was employed, the resulting influence line could vary depending on a number of parameters, namely vehicle speed, vehicle initial condition or level of vehicle-pavement-bridge interaction. It is clear that the safe removal of the dynamic component of the signal prior to the application of any optimization technique will lead to an improvement in the accuracy of the estimated influence line. In this paper, the authors employ wavelet analysis to extract the total static component from the measured bridge response. The wavelet domain representation of the signal facilitates the suppres- sion of noise and dynamic frequency components that interfere with the static response. In contrast to the Fourier transform, that provides information about the frequency domain, the wavelet trans- form allows localization of both time and frequency domain information. One step in the wavelet transform calculates a low pass result (approximation) and a high pass result (detail), where the low pass result represents a smoother version of the original signal. The low pass result recursively becomes the input to the next wavelet step, which calculates another low and high pass result. So, the measured signal is decomposed into a number of approximations and details. Details and lower levels of the approximations contain noise and vehicle and bridge dynamic components, while higher approximation levels provide a signal representation, mostly, if not entirely, static. This high-level approximation can be used to obtain the influence line more accurately than using the original signal through a least-squares fitting technique. The approach is tested with numerical sim- ulations of a bridge response when crossed by a two-axle truck. The simulated response is further contaminated with Gaussian noise. Then, the influence line is calculated using a variety of wavelet functions for two bridge spans and two levels of noise. Results show that wavelet decomposition of the signal allows removing undesired frequency components without significant loss of static information and it can be used to provide an accurate influence line even in the case of significant low-frequency bridge dynamics and high levels of noise.

624 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Seismic assessment and evaluation of 520 highway bridges in Western Kentucky

C.C. Choo California State University, Fresno, California, USA I.E. Harik University of Kentucky, Kentucky, USA W. Zatar Marshall University, WestVirginia, USA H.S. Ding Southeast University, Nanjing, China

The New Madrid and Wabash Valley are two of the most influencing seismic zones to the west and northwest of Kentucky. The New Madrid Seismic Zone (NMSZ) is particularly active; in the month of September 2007 alone, there have been twenty recorded earthquakes with magnitudes ranging from a low of 1.3 to a high of 2.6. Following a seismic event in or in the vicinity of western Kentucky, it is essential that high priority and emergency routes remain functional and operational. This requires ensuring the ability of all the bridges along these routes to withstand future major earthquakes. Interstate-24 and five other parkways have been designated as high priority routes; combined they have more than five hundred bridges on and/over these routes. Due to the sheer volume, and limited funds and personnel, it would not be feasible to assess the safety of each bridge with adequate details. The objective of this study is to provide and apply a seismic risk, rating, and assessment system for the 520 bridges and their embankments. The result is the identification the most vulnerable bridges susceptible to damage during a major seismic. The investigative effort begins with a preliminary assessment and ranking of these bridges based on the 50-year and 250-year time histories developed specifically for Kentucky; following the collection of information and a comprehensive inventory of these bridges. The outcome from the preceding step helps identify the most seismic vulnerable bridges, which are evaluated in subsequent step. A more in-depth assessment and evaluation is then carried out for the selected bridges. The detailed evaluation is based on a capacity/demand ratio method performed on various components of a bridge. 3-D finite element model of each of the selected bridges is constructed using commercially available software to facilitate the evaluation. Deficiencies of the components, if any, are identified and documented. A preliminary assessment of the stability of the bridge embankments and the liquefaction poten- tial of the foundation soil is also performed. Though detailed analysis was not carried out, the preliminary assessment helps prioritize and identify the ones that are most vulnerable and rank them in accordance with their needs for further evaluation or remediation. The outcomes of these investigations indicate that the rating methods provide an effective mean of identifying high seismically vulnerable bridges, and that helps authorities prioritize bridges for further evaluation, retrofit, or remediation.

625 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Cross-sectional stress distribution of short suspenders in arch bridges

Y.B. Li Department of Bridge Engineering, Tongji University, Shanghai, China Q.W. Zhang State Key Laboratory for Disaster Reduction in Civil Engineering, Tongji University, Shanghai, China

The structural behavior and the non-uniform cross-sectional stress distribution of vertical short suspenders in arch bridges are studied in this paper. A short suspender of arch bridges can usually be considered as an end-fixed beam because of its flexural and shear rigidity. The deformation of a short suspender under external loads consists of axial extension, longitudinal and transverse flexure. Since the influence of shear stress can be ignored due to its small value, the cross-sectional stress of suspender is regarded as the combination of axial tensile stress and flexural normal stress. The following assumptions is made in this study: (1) the suspenders are fixed at the connections with the ribs and the deck, the deformation of suspender meets the condition of the Bernoulli’s principle; (2) the ribs is infinitely rigid but the deck is flexible, and the relative displacement of suspender is mainly subjected to vibration of deck; (3) the deck can be considered as Bernoulli- Euler beam, and the influence of shear deformation and the moment of inertia is neglected; and (4) The frictions between wires are not counted. The temperature distribution along the circumference and radius is continuously measured in three days by an in-house system. Thereby the mathematical model of the circumferential temper- −a(R−r) ature difference distribution of the suspender is proposed as Trθ =T0(1 + cos θ)e /2, where T0 is 10 according to measurements. Formulations for calculating the self-restraint stress and the secondary temperature stress of short suspender are proposed. The Cross-sectional stress of suspender due to vertical and longitudinal vibration of deck is investigated by considering a simply-supported deck subjected to a moving harmonic exciting force. The enhancement coefficients of stress subjected to vertical and longitudinal vibration of deck, λv and λMy, are defined to reflect the non-uniform stress distribution. For illustration, a built Concrete-Filled Steel Tube (CFST) arch bridge is studied. The vertical and longitudinal frequencies of vehicles are input as 3 Hz and 1.8 Hz, respectively. It is shown that the non-uniform stress distribution under local temperature load is significant, the maximum amplitude of local temperature stress is generally 24 MPa and the enhancement coefficient of stress is 0.14. The stress of the short suspender caused by the vibration of deck is governed mainly by the vehicle speed, vehicle frequency, and driving frequency with shift. It can also be seen that the higher modes may generally be neglected when computing the cross-sectional stress of short suspenders. The non-uniform stress distribution of the short suspender subjected to the temperature load and the vibration of the deck is significant, but approaches to be constant when the distance to the surface of suspender increases. Results also show that the stress cycle of short suspender due to these sources is evident.

626 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Experimental investigation on the Bi-lateral seismic behavior of a two-span bridge model isolated by rolling-type bearings

Kuo-Chun Chang Department of Civil Engineering, National Taiwan University, Taipei, Taiwan Meng-Hao Tsai Department of Civil Engineering, National Pingtung University of Science & Technology, Pingtung, Taiwan Zheng-Yi Lin Department of Civil Engineering, National Taiwan University, Taiwan

ABSTRACT: Seismic isolation technology has been applied to protect civil structures from earth- quake damage for years. On the growing need of practical application, many seismic isolation devices have been developed in the last few decades. Among those developed devices, the Rolling- Type Bearing (RTB) has been proved to be effective in reducing earthquake-induced demand on structures. In this paper, shaking table tests are conducted to investigate the bi-lateral seismic behavior of a two-span, simply supported bridge model isolated by the RTBs with or without vis- cous dampers. Seismic responses of the isolated bridge model under the unilateral and bi-lateral excitations are obtained to investigate the response difference of the two vibration units and the bi-directional interaction effect of the RTB system. Test results show that those two vibration units isolated by RTBs may exhibit different earthquake responses, even though they are designed and constructed in the same way. Also, the bi-directional interaction effect is apparent for the RTB system without viscous dampers, especially for the longitudinal seismic responses. Nevertheless, the added viscous dampers are effective in reducing the response difference of the vibration units and suppressing the bi-directional interaction effect of the RTB system.

Keywords: Seismic isolation, rolling-type bearing, shaking table test, viscous damper.

627 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Modal analysis of corrugated steel flexible shell bridge structure before backfilling

D. Beben Civil Engineering Department, Opole University of Technology, Opole, Poland Z. Manko Institute of Civil Engineering, Wroclaw University of Technology, Wroclaw, Poland

ABSTRACT: This paper is presented a modal analysis of a steel flexible shell structure made of corrugation plates which is a main load-carrying element in soil-steel bridge situated in Giman (Sweden). Comparison and parametrization were carried out on two calculation models: a plane (2D) one and a spatial (3D) one. The rigid connection between the steel shell elements and the articulated (hinged) joints employed in the foundation are notable features of the bridge structure. The numerical analysis was performed by means of a FEM based software package. Two kinds of finite elements were used: beam and shell ones. The natural frequencies of the bridge shell structure were compared by means of subspace and reverse iteration methods. The steel flexible shell model was parameterised taking into account the effect of the variable (because of joints) characteristics of its shell elements connected with a foundation on the eigenvalues under the deadweight. In addition, the optimum number of finite elements in the models and the convergence of the results depending on the number of iterations were estimated. Conclusions were drawn on the basis of analysis.

The growing popularity of soil-steel bridge designs in the world in recent years and the increasing length of the main spans used in these structures have inspired searches for new, detailed and extended formats of their analysis which would take into account adverse (extreme) load conditions, the safety and strength of such bridges and their economical design. The long spans and the peculiar features of soil-steel bridges, particularly the complicated modes of vibrations, put such structures in a special category. The natural frequencies and the corresponding to them modes of free vibrations are dynamic quantities crucial for corrugated steel plate bridge characterized by considerable differences in the rigidity of their particular structural members (the shell elements, ribs, the substructure layers, the soil-steel system, etc.). The natural frequencies of corrugated steel plate bridge structures are usually low and as a result, the latter are more vulnerable to the dynamic effects of soil loads than, for example, simply supported beam structures. Therefore it becomes necessary to subject such structures to modal analysis which is an essential part of the theoretical considerations, the design calculations and, above all, the normal service of such bridges. A numerical analysis of a Corrugated Steel Plate (CSP) bridge, concerning its dynamic aspect, was carried out. This served as a basis for a general analysis of the modes and frequencies of free vibrations. An appropriate plane (2D) and spatial (3D) models made it possible to determine the particular dynamic properties of this structure in the range of modal frequencies. The dynamic response of the corrugated steel shell structure under its deadweight before backfilling could be verified better by means of the spatial model. The analysis was expanded through four versions of discrete models within the 2D and 3D systems, with shell elements rigidity, variable along the span support length depending on the use or elimination of articulated joints in the connection with foundation, being the principle parameter. In addition, the convergence of solutions was determined for the plane and spatial models and the optimum number of finite elements used in the calculation models was also given.

628 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Experimental study on the shear characteristics of seismic isolation bearings

I.J. Kwahk, C.B. Cho & Y.J. Kim Korea Institute of Construction Technology, Goyang-shi, Korea

Researches attempting to minimize seismic damages and reduce seismic loading through the instal- lation of seismic isolation bearing in structures or in bridges have been and are being continuously performed in the scope of aseismic technology. For the design and construction of such seismic isolation bearings to gain reliability, securing uniform quality during their manufacture as well as performance under environmental changes all along their service life are priorities. Accordingly, this study conducted tests on lead rubber bearings (LRB) and rubber bearings (RB) produced with respect to the seismic isolation design in order to evaluate their shear characteristics standing among the basic performances required for the assessment of the seismic performance and quality of seismic isolation bearings. The shear performance of the specimens were evaluated through comparison of the experimentally measured shear stiffness and equivalent damping ratio with the design values, and the quality level was assessed by analyzing the deviation of the design values due to fabrication error. The so-measured level of quality exhibited considerable deviation regard

Table 1. Design conditions and number of specimens.

Design Seismic Design Dead Live Total conditions displacement displacement loads loads load Number of Designation (mm) (mm) (kN) (kN) (kN) specimens

LRB1 240 240 5919.3 985.6 6905 18 LRB2 240 240 5356.4 884.6 6241 20 LRB3 240 240 3526.5 545.2 4072 8 LRB4 (Br) 40 40 3405.5 2270.7 5676.2 3 LRB5 (3T) – – ––– 3 LRB6 (7T) – – ––– 3 RB2 240 240 5574.1 964 6538 20 RB3 240 240 3682.4 636.5 4319 20 Total – – –––95

Figure 1. Manufactured specimens.

629 to the quality specifications of seismic isolation bearings proposed by the Korea Highway Bridge Design Code with distribution very close or sometimes exceeding the specified design limit values. Moreover, seismic isolation bearings are likely to experience additional changes of their shear characteristics according to temperature, aging, creep, amplitude and frequency of the seismic load occurring during their service life. Accordingly, the fact that most of the seismic isolation bearing is showing quality level close to the error limits since their fabrication is a crucial problem which needs stringent management of the performance evaluation of the bearings to be delivered and used on field.

630 Seismic design and performance issues for highway bridges

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Post-earthquake evaluation of reinforced concrete bridge columns

A. Vosooghi & M. Saiidi University of Nevada at Reno (UNR), Reno, Nevada, USA S. El-Azazy California Department of Transportation (Caltrans), Sacramento, California, USA

ABSTRACT: The serviceability of bridges is often in question after earthquakes. Based on the inspection of the damaged RC columns, engineers have to determine if the bridge is sufficiently safe to be kept open to traffic or if it is repairable. The decision has to be made quickly.Yetit is a critical decision because it could vastly affect the operation of the transportation network and could impact the emergency response operation and economic well being of the community with potentially far reaching effects. To assist bridge investigation teams, it is necessary to define logical damage states that can visually be identified in the field. These damage states should provide information about the structural soundness and internal stress and deformation history of the bridge. The assessment of the status of the structure can then be followed with a reliable practical decision about any repair strategy. Extensive shake tables testing of a variety of columns in the past decade enable researchers to correlate the observed damage with internal stress and deformation history of the structure. Five distinct apparent damage states are defined for RC columns. By investigation of detailed data from 17 bridge column models, mostly tested on shake table, correlations have been established between damage states and internal and external seismic response parameters. The models from which the data are obtained consist of single columns, two-column piers, and a 2-span bridge model. The five damage states are flexural cracks (DS-1), first spalling (DS-2), extensive spalling (DS-3), visible bars (DS-4), and imminent failure (DS-5). Note that “failure” has not been included in the list because a column with ruptured reinforcement is considered to be irreparable. The response parameters are the maximum drift ratio, maximum longitudinal bar strain, and maximum lateral bar strain. The drift represents the maximum external response of the column, and strains represent the maximum internal response of the column. Based on shake table test data for 17 columns, the correlation between each damage state and internal and external response parameters of the columns was investigated. Although there was significant scatter in the data, the response parameters showed a clear trend in terms of sensitivity to different damage states. The results show that if some of the bars are visible in the damaged column, the lateral bars have experienced yielding, therefore their effectiveness to provide confinement is questionable.

633 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Identification of effective seismic retrofits for common bridge classes on the basis of failure probability

J.E. Padgett Rice University, Houston, TX USA R. DesRoches Georgia Institute of Technology, Atlanta, GA USA

EXTENDED ABSTRACT: The vulnerability of non-seismically designed bridges and potential for strong ground shaking in the Central and Eastern United States (CEUS) is becoming more readily recognized. As a result, several states are considering or have established seismic retrofit programs to evaluate and retrofit seismically vulnerable bridges. However, few studies have assessed retrofit measures for the bridges common to this region, or provided effective means for comparing alternate retrofit strategies. There is a need for improved understanding of the viability of different retrofit measures for typical CEUS bridges in order to support such retrofit decision making and mitigation programs. To this end, this work has developed a methodology to assess the vulnerability, or fragility, of common CEUS bridge classes with a range of potential retrofit measures. The tools developed are known as retrofitted bridge fragility curves, which state the probability of meeting or exceeding various damage states over a range of earthquake intensities. The fragility curves offer insight as to the most effective retrofit measure in reducing the failure probability for different bridge types and damage states. This paper summarizes some of the key findings of the research in terms of which retrofit measures were found to be most effective for common CEUS bridge classes, and the insights gained as to why some measures are not ideal on the basis of system failure probability. The different bridge types evaluated as a part of this work include the Multi-Span Simply Supported (MSSS) concrete girder bridge class and the Multi-Span Continuous (MSC) steel girder bridge class. Retrofit strategies considered include seven different combinations of measures including transverse shear keys, steel column jackets, elastomeric isolation bearings, steel restrainer cables, and seat extenders. The findings of the study indicate that the steel jackets are particularly effective for the MSSS concrete bridge class, whose system fragility is strongly dependent upon the vulnerability of the columns (mainly at the higher damage states). While the MSC steel bridge also has problems with column vulnerability, these bridges also tend to use highly vulnerable steel fixed and rocker bearings, which cannot be improved by the use of steel jacketing alone. Therefore, for the MSC steel bridge, the elastomeric isolation bearings tend to be superior in improving the performance for every damage state, because they not only isolate the deck from the substructure and limit column demands and active abutment demands, but they also replace the existing bearings with less vulnerable ones. It is noted that the elastomeric isolation bearings do tend to increase the passive deformations or transverse deformations of either bridge type, resulting in a more vulnerable abutment. In some cases, as illustrated above, this may be overshadowed by the other positive effects of the retrofit. Further insights gained from comparison of the conditional system failure are detailed in the paper. The results are intended to provide guidance for bridge owners and engineers and improve the state of practice of seismic retrofit in the region.

634 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Enhancement of axial ductility of circular concrete bridge columns

L.A. Marvel, J.C. West & R.A. Hindi Bradley University, Peoria, Illinois, USA

ABSTRACT: The compressive strength and ductility of a concrete column is mainly due to the lateral confining steel surrounding the concrete core, as shown in countless research projects. When a concrete column is subjected to axial load it expands in the transverse direction. ACI 318-05 specifies a minimum spacing of 25 mm for spirals in all reinforced concrete col- umns. For the design of important structures that are located in areas of high seismic activity where maximum ductility is required, the 25 mm spacing may limit the ductility. In addition, ACI 318-05 has also implemented a maximum spacing of 75 mm for spirals. This maximum requirement can cause problems when constructing large diameter columns, this maximum spacing creates a cumbersome construction process trying to maintain constant spiral spacing. Hindi (2005) proposed the use of two opposing cross spirals in lieu of the conventional single spiral which can be manipulated to either increase the characteristics of the column or the enhance the constructability of the member. This paper experimentally investigates the axial behavior of high strength concrete bridge columns using the cross spirals confinement technique. Twenty-one circular concrete columns with various spiral spacing’s and various longitudinal steel ratios were built and tested. Twelve columns, which have the same longitudinal reinforcement ratio of 2% and different spiral con- figurations and spacing, are presented. The columns were loaded with monotonic axial loading while monitoring the axial displacement to study the behavior of the new confinement technique compared to the conventional single spiral confinement technique. The columns considered in this research were based on a typical bridge column that satisfies ACI 318-05 specifications. Four of the twelve specimens were constructed using the conventional single spiral used as control spec- imens and the remaining eight columns were constructed using the cross spirals. All specimens had the same overall dimensions with the only variation being the spiral configuration and the spiral spacing. A diameter of 350 mm and length of 1000 mm was used. 10 − 15.9 mm longitu- dinal reinforcing bars were used with a longitudinal steel ratio of 0.02. The confining spiral was 9.5 mm with four different spiral spacings 40, 50, 55, and 60 mm. A loading rate of about 45 kN per second was maintained, and continued until complete failure was reached. During the testing the applied load and vertical deformation of the column was recorded. All of the bridge columns displayed visible signs of spalling at about 2.7 MN and all failed due to rupture of the spiral(s) which led to buckling of longitudinal reinforcement and ultimately crushing of the concrete core. Several conclusions were drawn: – Columns confined with regular spirals and cross spirals that have similar volumetric ratios of confining steel performed very similar during the testing. – Columns confined with cross spirals with twice the volumetric ratio of confining steel as compared to the regular spirals had significantly enhanced strength and strain capacity. – The new cross spiral technique of confinement may be used to increase the strain capacity or to increase the constructability of bridge columns. – With the addition of the cross spiral and keeping the same spiral spacing, strength and strain capacity can be increased. Alternatively, by using the cross spiral technique with double the spacing, congestion in heavily reinforced sections can be avoided without a reduction in the strength and strain capacities.

635 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Retrofitting structures with a combination of seismic isolation and attenuation

A. Caner Middle East Technical University, Ankara, Turkey M.J. Abrahams & E. Dogan Parsons Brinckerhoff, New York, USA C. Ozkaya Middle East Technical University, Ankara, Turkey

ABSTRACT: Seismic requirements were upgraded in most of the bridge design specifications in the last thirty years. The bridges constructed thirty years ago or more were mostly designed for gravity loads only and could be vulnerable at new levels of high risk earthquakes defined in the specifications. Typical seismic retrofit measures can be performed in different ways such as to strengthen the columns with steel jacketing, or to wrap columns with composite materials, or to make the decks or spans continuous by eliminating existing joints, or to strengthen the foundation with mini piles, or to improve the stability of the approach fill and abutments, or to replace the old bearings with seismic isolation bearings. Out of all these options, seismic isolation appears to be the most common scheme if the objective is to minimize the seismic forces and to avoid costly substructure retrofits. Moreover, the foundations and columns are usually difficult to retrofit due to limited accessibility during service. Deck elements may become effective attenuators of seismic forces. Permitting deck joint colli- sion during an extreme seismic event while possibly causing localized and repairable damage to deck elements makes retrofitting easier and can result in reduced cost. A case study is used to illustrate how a half a century-old bridge was retrofitted using this approach. The 102nd Street bridge is a 65 year old, 2 lane, 250 meter long, viaduct over the Grand Central Parkway, Queens, New York.The bridge has 11 spans. Bridge superstructure is composed of steel girders with a concrete deck, and supported by reinforced concrete columns and abutments at each end. A rehabilitation design was prepared by Parsons Brinckerhoff, New York. The bridge rehabilitation project included replacement of the concrete deck, replacement of the existing rocker bearings with friction pendulum bearings for a better seismic performance, strengthening of the existing steel girders with additional girders, replacement of the deck drainage and deck joints, adding steel framing for future access ramps and repainting. The bridge could not be closed during the rehabilitation as it was needed to provide access to the east end terminals of the LaGuardia Airport. Nonlinear time history analysis of the bridge was conducted using the ADINA software. The hysteretic force-displacement behavior of the friction pendulum bearings was modeled per the properties given by the bearing manufacturer. A soft bumper was placed at the expansion joint gaps to minimize the local damage caused by pounding. The pounding element was selected to be a nonlinear compression only gap spring with an initial stiffness properties of the soft bumper followed by the more rigid deck stiffness. The soil-structure interaction was modeled by the ground springs per the geotechnical recommendations. Each of the three earthquake records was made compatible to the design response spectrum of New York City Seismic Code (1998) with a 2500 year return period.

636 It is found that seismic isolation can be effectively used to retrofit the existing bridges so that the substructure does not need to be retrofitted. The local damage caused by pounding force can be minimized at expansion joints by adding relatively soft bumpers at the joints. Furthermore, seismic retrofit design is not only controlled by force demands but also by displacement demands.

637 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Supplemental device to improve the performance of seismic-isolated bridges in near-fault zones

M. Dicleli Middle East Technical University, Ankara, Turkey

For SIBs subjected to far-field ground motions, the Isolator Displacements (IDs) generally have manageable magnitudes. However, for SIBs subjected to NF ground motions with directivity effect, the IDs tend to be considerably large. Thus, isolators, expansion joints and substructures with very large dimensions may be required for SIBs located in NF zones to accommodate such large IDs. Consequently, the application of seismic isolation to bridges in NF zones becomes virtually impractical as a stand-alone seismic mitigation technique. Although it may be possible to reduce the IDs and Substructure Forces (SFs) to manageable ranges by providing additional seismic control devices, most of such devices are generally expensive and are not commonly used for seismic protection of bridges, mainly due to the lack of experience with such devices and the associated maintenance cost. Thus, a rational solution to the problem associated with large IDs for SIBs subjected to NF earthquakes is required. To address the problem stated above, the efficiency of using Elastic Gap Devices (EGDs) in SIBs for reducing the IDs while keeping the SFs at reasonable ranges for a wide range of NF earthquake magnitudes is investigated. Elastomeric bearings placed in parallel with isolators between the superstructure and substructures that are engaged upon closure of a gap may be used for this purpose. Elastomeric bearings have already been used over many years by state departments of transportation and require only minimal initial cost and maintenance compared to most seismic control devices. Thus, they can easily be used for seismic design and retrofitting of SIB located in NF zones. Nonlinear time history (NLTH) analyses are conducted in two phases to study the effect of EGDs on the performance of SIBs. In the first phase, a parametric study, involving more than 400 NLTH analyses of simplified structural models representative of typical SIBs, are performed using simulated NF ground motions. As the response of SIBs subjected to pulse type excitations are governed by the mass, m, of the bridge, the properties, of the isolator and the properties of the NF ground motion, the effect of EGDs on the performance of SIBs is considered in relation to these parameters and the EGD properties (gap and elastic stiffness). For this purpose, a total of ten different combinations of isolator-EGD stiffness values and gap openings are considered. Furthermore, the analyses are performed for various characteristic strength values of the isolators. In the second phase of analyses, the results obtained from simplified structural models and simulated NF ground motions are verified. For this purpose an actual bridge with (i) regular elastomeric bearings (ii), with isolators and (iii) with isolators and EGDs is modeled and analyzed using five recorded NF ground motions. The analyses results revealed that providing EGD is beneficial for decreasing the isolator dis- placements to manageable ranges. Furthermore, using EGDs produced smaller isolator velocities and a faster decay of the displacement and velocity amplitudes of the isolator. This is very benefi- cial for the cyclic performance of the isolators as the heat generated by the isolators under cyclic motion will be dramatically reduced due to the presence of the EGDs.

638 Smart sensing and monitoring technologies for bridge maintenance, safety and management

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Baseline knowledge discovery from one-year structural monitoring measurements of Donghai Bridge

Z. Sun & Z.H. Min Department of Bridge Engineering, Tongji University, Shanghai, China Z.F. Zhou Shanghai Juyee Technology Development Ltd. Company, Shanghai, China

Donghai Bridges is a linkage connecting the Luchao Port in Shanghai and the Yangshan Island deep water Port in Zhejiang Province. A Structural Health Monitoring System for Donghai Bridge (SHMSDH) was designed to be able to automatically collect the data of displacement, force, stress and so on of the structure through the sensors embedded in or installed on the surface of the structure. Since the formal opening to use on December 10th 2005, Donghai Bridge has operated normally for almost two years. The SHMS installed on the bridge has collected millions of valuable data to describe the healthy condition of the bridge under different types of loadings (wind, ground motion, wave and etc), different types of environmental variations (such as temperature, moisture and etc). It is thus the time to analyze the signal, extract meaningful features, and accumulate the baseline knowledge about the bridge. So the research group downloaded 8-months (from January 1st 2007 to August 31st 2007) structural acceleration data from the SHMS of Donghai Bridge to discover the baseline information about the variation of the modal characteristics of the bridge. Figure 1 shows the Probability Density Function (PDF) for the fundamental symmetrical bending mode of the bridge in these three seasons. According to these PDFs, the thresholds concerning temperature effect for damage alarming in winter, spring and summer can be computed. The results shown in fig. 1 tell that structural modal frequencies seem to be correlated to the magnitude of the air temperature.

Figure 1. The probability density function (PDF) for the fundamental symmetrical bending mode.

641 This observation is quite reasonable since the variation of the temperature will induce the change of structural geometrical stiffness and thus in turn the shift of structural modal frequencies. These observations also tell that: 1) the influence of the temperature variation on structural stiffness of a cable-stayed bridge is complex; 2) the air temperature effect mainly dominant the uniform change of structural global stiffness but not the change of local stiffness; 3) the temperature effect is an important factor for the variation of modal frequencies. If modal frequencies are used as the condition indices for damage assessment, this effect must be calibrated.

642 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Test-bed implementation of piezopaint-based acoustic emission sensor for crack initiation monitoring

Y. Zhang University of Maryland, College Park, USA X. Li Lehigh University, Bethlehem, USA

ABSTRACT: Fatigue-induced crack monitoring and corresponding retrofit actions will lead to a prolonged life and enhanced reliability of structural systems. Manual inspection is plagued with problems such as intensive labor, time consuming process, and subjective results; therefore it is unsuitable for rapid assessment of structural conditions. Acoustic Emission (AE) is the elastic wave generated by sudden energy release within a material, it provides real-time information on damage progression within a structure. Different from ultrasonic test, which excites elastic waves into a solid, AE sensor passively listens to the signals generated by crack initiation and progression with the monitored structure. Due to recent growing use of piezoelectric materials for continuous structural health monitoring applications, AE sensor based on lead-zirconate-titanate (PZT) wafer operating in d31 mode has attracted attention from some researchers. This paper is focused on the analytical study of piezopaint that is proposed for use as close-range acoustic emission sensor in facture monitoring. Models for key material properties of piezopaint that are most important to ultrasonic sensing are first reviewed. The results of a parametric study, which is conducted to provide guidance on how to tailor the material properties of piezopaint to match the special needs of ultrasonic sensing applications, are presented here. The effect of piezoelectric ceramic volume fraction on paint sensor response is examined. The results of this analytical study results prove to be very useful to the optimization of piezopaint formulation as well as provides guidance to the design of piezopaint sensor for fracture monitoring applications. Experimental data is also provided to validate the analytical result. The use of piezoelectric mate- rials for ultrasonic signal measurements is discussed along with a series of ultrasonic tests performed to verify the ultrasonic sensing capability of piezopaint. Field implementation of piezopaint sensors on a steel bridge test-bed in Korea is planned.

643 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Development of experimental benchmark problems for international collaboration in structural response control

Chin-Hsiung Loh Department of Civil Engineering, National Taiwan University, Taipei, Taiwan Anil K. Agrawal Department of Civil Engineering, The City College of New York, USA Jerome P. Lynch Department of Civil & Environmental Engineering, University of Michigan, Ann Arbor, USA Jann N. Yang Department of Civil & Environmental Engineering, University of California, Irvine USA

ABSTRACT: In the past two decades, many control algorithms and devices have been proposed for civil engineering applications. Each control strategy, including algorithm, device configuration, etc., have their own merit, depending on the particular application and desired effect. The ability to make comparisons for different control strategies on a common basis will enable us to focus future efforts in most promising directions and to establish performance goals and specifications. One approach to achieve this goal is to establish large experimental models that allow researchers in structural control to test their algorithms and devices and directly compare the results. Several benchmark structural control models have been developed during last decade through the sponsor- ship of ASCE Committee on Structural Control and International Association of Structural Control and Monitoring (IASCM). The main objective of developing these models has been a standardized evaluation of the performance of various control systems/algorithms when applied to various types of structural systems. A systematic and organized investigation of various forms of structural con- trol devices/algorithms based on these benchmark models has led to a significant advancement on the understanding of the performance of various devices/ algorithms and the dissemination of out- comes to structural engineering community. An extensive analysis of these benchmark structural control problems formed the basis for a special issue of Earthquake Engineering and Structural Dynamics (Spencer et al. 1998a,b). Recently, well-defined analytical benchmark problems have been developed for bridge structures subjected to seismic excitation through the sponsorship of the ASCE structural control committee (Agrawal et al. 2004). In order to truly demonstrate the capability of various structural control systems in protecting the integrity of buildings during earth- quakes, we propose to develop experimental benchmark models for: (i) a 3-story steel frame with dimensions 3.0 m by 2.0 m in plane and 3.0 m height for each story, and (ii) a 6-story steel frame with dimensions 1.0 m by 1.2 m in plane and 1.0 m height for each story. These two experimen- tal benchmark models will be made available to the international structural control community for testing various control algorithms and devices. The objective of this paper is to: (i) introduce previous studies conducted by different organizations using these two benchmark structures, and (ii) propose the test-bed structures for an international collaboration on structural control through benchmark studies.

Keywords: Structural control benchmark, semi-active control, wireless sensing and control system

644 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Advance sensor technologies on Korean Bridges: Field benchmark opportunities

J.P. Lynch & J.H. Kim Department of Civil & Environmental Engineering, University of Michigan, Ann Arbor, MI, USA Y. Zhang Department of Civil & Environmental Engineering, University of Maryland, College Park, MD, USA M. Wang Department of Civil & Materials Engineering, University of Illinois-Chicago, Chicago, IL, USA H. Sohn & C.B. Yun Department of Civil & Environmental Engineering, KAIST, Daejeon, Republic of Korea

The United States and Korea have vast national bridge inventories that require vigilant inspec- tion and repair. With a new generation of sensing technologies and computational tools emerging from interdisciplinary research between civil engineering and other engineering disciplines, bridge management strategies are rapidly improving. This paper describes a new US-Korean joint col- laboration focused on the technical and scientific challenges associated with fusing many of these emerging sensors into a comprehensive Structural Health Monitoring (SHM) system for bridges. Four fully operational bridges managed by the Korea Highway Corporation (KHC) have been offered to the research team for the installation of the sensors under investigation. The KHC has recently constructed a redundant stretch of the Jungbu Inland Expressway near Icheon, South Korea. Along the length of this road are three bridges (the Geumdang, Yondai and Samseung Bridges as shown in Figure 1) that carry traffic over irrigated agricultural flood plains. These three bridges are selected as testbed bridges in this study; in addition, a more complex long-span cable stay bridge is selected as the study’s fourth testbed structure. A variety of new sensor technologies have been selected for integration including wireless sensors, EM stress sensors, vision displacement systems and piezoelectric active sensors. Using these sensors as building blocks, the first phase of the study focuses on the design of a comprehensive SHM system that will be deployed upon the highway bridges in Korea. With permanently installed SHM systems in place, the second phase of the study provides open access to both the bridges (e.g. to test new sensors) and the response data continuously collected (e.g. to test interrogation algorithms) as an international testbed for struc- tural health monitoring. Concurrent to the research described herein, the US-Korea collaboration is also focused on using the testbed opportunity to accelerate the development of smart structure curricula at the undergraduate and graduate education levels.

Figure 1. Three test-road bridges: (a) Geumdang Brdige; (b) Yondai Bridge; (c) Samseung Bridge. 645 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Remotely controllable structural health monitoring systems for bridges using 3.5 generation mobile telecommunication technology

K.Y. Koo, J.Y. Hong, H.J. Park & C.B. Yun Department of Civil and Environmental Engineering, KAIST, Korea

One of fundamental obstacles in SHM system implementation is the network implementation which connects a control office and a bridge site. Most of conventional wired internet services through telephone lines or cable TV lines are usually unavailable in rural area where bridges are commonly located. Mobile internet service using CDMA mobile telecommunication network is available in virtually everywhere but the maximum data transmission speed is low so that it can be applied to only limited cases of small amount of data transmission. Construction of a private network line can be an alternative way. But, the construction and maintenance cost may be significant and inappropriate. Recently, a new 3.5 generation mobile telecommunication technology changed the network implementation easy and tractable. HSDPA (High Speed Downlink Packet Access) technology commercialized firstly in South Korea, 2007, is currently available in 55 countries. HSDPA shows national-wide service area so that virtually any bridge site can be accessed with high speed data transmission rate. In this study, a new SHM system framework using HSDPA is presented incorporating remote desktop software. The presented SHM system has grate advantage that it can be controlled and maintained remotely by utilizing remote desktop software through Internet implemented by HSDPA. The proposed framework and its feasibility are tested and demonstrated using a field appli- cation on the Geumdang Bridge in operation. It can be expected that advances in the mobile telecommunication industry continuously benefit the network implementation of SHM systems for bridges.

Figure 1. Architecture of the remotely controllable SHM framework.

646 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Structural damage assessment using optical Fiber Bragg Grating vibration sensing system

R.J. Sun, Z. Sun & L.M. Sun Department of Bridge Engineering, Tongji University, Shanghai, China

1 A FBG ACCELEROMETER

As we known, the reflection wavelength which is called Bragg wavelength is related to strain and temperature changes. Mechanism of the proposed FBG accelerometer is depicted in Fig. 1. A prototype FBG accelerometer was fabricated accordingly (Fig. 2). Its performances were exam- ined by a series of shaking table tests. The tests verified the sensitivity and linearity of the sensor.

2 INTEGRATED FBG SENSING SYSTEM

For structural health monitoring, this novel sensor is integrated with the FBG strain gauges and FBG thermometers and a distributed sensing system is setup. To verify the efficiency of the proposed distributed sensing system for structure health monitoring, an experimental study is conducted (as shown in fig. 3). The results (as shown in fig. 4) verifies that the integrated system is efficiency to collect reliable data for structure health monitoring and damage detection.

Figure 1. Mechanism of the new FBG accelerometer. Figure 2. A prototype FBG accelerometer.

Figure 4. The measured strain mode difference for Figure 3. Experimental setup. damage locating.

647 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Self-sensing and power harvesting carbon nanotube-composites based on piezoelectric polymers

K.J. Loh, J. Kim & J.P. Lynch Department of Civil & Environmental Engineering, University of Michigan, Ann Arbor, MI, USA

Recently, the need for long-term Structural Health Monitoring (SHM) has warranted the design of novel sensing transducers. In this study, single-walled carbon nanotubes (SWNT) are used as filler material to tailor the sensing and power harvesting performance of poly(vinylidene fluoride- trifluoroethylene) [P(VDF-TrFE)] nanocomposites. Six different P(VDF-TrFE)-based thin films of different SWNT concentrations are fabricated via a thermal evaporation and annealing procedure for performance comparison. Preliminary experimental results shown in this study will highlight three performance attributes of SWNT-P(VDF-TrFE) thin films: 1) high deformation tolerance for robust field applications, 2) linear strain sensing performance, and 3) potential power harvesting capabilities. First, strain sensing using SWNT-P(VDF-TrFE) nanocomposites can be realized by taking advantage of the fact that applied strain induces deformations to the planar β-phase molecular conformation of P(VDF-TrFE) chains. The deformations in individual chains induce changes in the material’sdielectric properties or capacitance. Results from Figure 1 show that the nanocompos- ite’s capacitance varies linearly with applied strain (to ±2500 µε). In general, these strain sensors exhibit no drift over time and possess capacitive strain sensitivities between 1.2 and 2.5, comparable to traditional metal-foil strain gauge factors of 2.0. On the other hand, the piezoelectric nature of these SWNT-P(VDF-TrFE) films can be uti- lized for harvesting energy from structural vibrations. Upon mounting various specimens onto a cantilevered aluminum plate excited at a sinusoidal frequency of 20 Hz via a modal shaker, the self-generated voltages from each specimen is compared to a commercially-poled P(VDF-TrFE) film (Fig. 2). Experimental results suggest comparable power harvesting capabilities between the proposed low-voltage poled (300–500 V) and commercially poled (2500–3500 V) nanocomposites. In addition, due to the strong piezoelectric effect of these nanocomposites, and that no power supply is required for sensor interrogation, a simple data acquisition system can interrogate multiple films simultaneously and in real-time by recording self-generated output voltages.

Figure 1. (Top) Performance of an SWNT-P(VDF- Figure 2. Measured piezoelectric output voltages of TrFE) capacitive strain sensor and (Bottom) its cor- P(VDF-TrFE)-based thin films mounted onto a thin responding linearity plot showing a strain sensitivity aluminum plate excited at a sinusoidal frequency of (SC = (C/C0)/ε) of 1.5. 20 Hz.

648 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

A nonlinear impedance method and its potential application in baseline free crack detection in metallic structures

D. Dutta Department of Civil and Environmental Engineering, Carnegie Mellon University, Pittsburgh, PA, USA H. Sohn Department of Civil and Environmental Engineering, Korea Advanced Institute of Science and Technology, Daejeon, Korea

An electro-mechanical impedance method using the principles of nonlinear acoustics is developed to detect cracks in metallic structures. Lead-Zirconate-Titanate (PZT) transducers are used for exciting and measuring acoustic waves in a structure. Cracks in a structure give rise to a set of non- linear boundary conditions in the acoustic field that leads to the production of acoustic harmonics. This phenomenon affects the coupled electro-mechanical impedance of the PZT-structure system in a way that the measured voltage across the external impedance of a voltage divider circuit con- tains super harmonics of the input frequencies. Furthermore, only the PZT-patches near the crack seemed to be affected by acoustic nonlinearity. The above observations are exploited to detect as well as localize a crack in a metallic structure. The technique also has potential for baseline-free crack detection (i.e. detecting cracks without referring to data from undamaged condition of the structure), provided comparison is allowed between the data obtained from PZT patches near the crack with those away from it.

Figure 1a. A schematic of a cracked beam with a Figure 1b. A schematic of the voltage divider cir- surface mounted PZT. cuit.

Figure 3. Variation of damage index (D) with dis- tance from crack (where applicable) for different Figure 2. Experimental setup for detecting cracks specimens (undamaged: US, damaged: DS-1 & on aluminum beam. DS-2).

649 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Piezoelectric sensor system for structural health monitoring

B. Kim & Y. Roh Kyungpook National University, Daegu, Korea

This paper presents the feasibility of a sensor system consisting of piezoelectric oscillator sensors to detect local damages and ultrasonic sensors to detect global damages in a structure. The oscillator sensor is composed of a feedback oscillator circuit and a piezoceramic lateral mode vibrator to be attached to a structure. Damage to the structure causes a change in the resonant frequency of the piezoceramic vibrator. The oscillator circuit instantly detects the frequency change and configures the damages. However, the response of the oscillator sensor is limited to the area near the sensor, thus local measurement. An ultrasonic sensor generates Lamb waves and the waves traveled over a long distance are received by another piezoceramic patch on the structure. The received wave form reflects all the defects encountered during the propagation, thus global measurement is possible. The two sensor types can be combined as a sensor network to work as a piezoelectric sensor system for structural health monitoring. In this paper, numerical and experimental performance of the oscillator sensor system was analyzed with a sample aluminum plate where artificial cracks of different lengths and number were imposed in sequence. Characteristics of the Lamb wave propagation across an artificial damage were also investigated through theoretical analysis. The reflection and transmission coefficients of S0 and A0 mode Lamb waves were analyzed in relation to the geometry of a rectangular crack when the wave propagated across the crack in an elastic plate.

650 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Fatigue reliability updating through inspections and monitoring data of steel bridges

Chunsheng Wang, Xin Yu,Yacheng Feng & Xin Liu Key Laboratory for Bridge and Tunnel Engineering of Shaanxi Province, Highway College, Chang’an University, Xi’an, Shaanxi Province, China

The safety state of existing steel bridges in China is not optimistic while these bridges are carrying an increasing volume of traffic. So methods for evaluating and updating the actual fatigue life of such structures using monitoring data are urgently required. Because fatigue flaws growth is random, and the practical load histories also have great random, the repeated on-site detection of structures has become a very important link of the evaluation. The development of modern NDI techniques may facilitate early detection of flaws and allow more economic inspection, strengthening and help to ensure the safe condition and extend the service life of such structures. So how to use the results of NDI to evaluate the fatigue reliability has become the research focus of fatigue life evaluation. The Bayes theorem can use the results of NDI techniques, which make the fatigue reliability models be updated and be more effective. The development of modern NDI techniques may facilitate detection of crack initiation and growth for existing steel bridges in operating period. An improved method for the estimation of the remaining fatigue life and safety of existing steel bridges is required in order to make their continued safe use possible while avoiding unnecessary repairs or replacements. In current paper, based on the probabilistic fracture mechanics, the probabilistic characteristics for NDI techniques and the Bayes theorem, the fatigue reliability updating assessment model for existing steel bridge components is presented, which develops the fatigue life and service safety evaluation methods for existing steel bridges. In the case study, the curves of the third suspender (M3L3) of Zhejiang Road Bridge in Shanghai can be figured out after inspected and updated, using the fatigue-updating models. The results showed that the fatigue reliability will increase with no flaws detected, and decrease with unknown flaws detected, and change remarkably with flaws detected.

ACKNOWLEDGEMENT

The writers gratefully acknowledge the financial support provided by National Natural Science Foundation of China (Grant No. 50408028).

651 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Advanced signal processing for ultrasonic structural monitoring of waveguides

M. Cammarata & P. Rizzo Department of Civil and Environmental Engineering, University of Pittsburgh, Pittsburgh, PA, USA D. Dutta Department of Civil and Environmental Engineering, Carnegie Mellon University, Pittsburgh, PA, USA H. Sohn Department of Civil and Environmental Engineering, Korea Advanced Institute of Science and Technology, Daejeon, Korea K.A. Harries Department of Civil and Environmental Engineering, University of Pittsburgh, Pittsburgh, PA, USA

ABSTRACT: Ultrasonic Guided Waves (UGWs) are a useful tool in those structural health mon- itoring applications that can benefit from built-in transduction, moderately large inspection ranges and high sensitivity to small flaws. This paper describes a damage detection method, based on wavelet transform and outlier analysis for structural waveguides. The method combines the advan- tages of UGW inspection with the outcomes of the Discrete Wavelet Transform (DWT) that is used for extracting defect-sensitive features that can be combined to perform a multivariate diagnosis of damage. In particular, the DWT is exploited to generate a set of relevant wavelet coefficients to construct a uni-dimensional or multi-dimensional damage index. The damage index is then fed to an outlier analysis to detect anomalous structural states. The general framework presented in this paper is applied to the detection of fatigue cracks in a W6x15 steel beam. The probing hardware consists of Lead Zirconate Titanate (PZT) materials used for both ultrasound generation and detec- tion at chosen frequency. The effectiveness of the proposed methods for the structural diagnosis of defects that are small compared to the waveguide cross-sectional area is discussed.

652 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Development of an optical fiber corrosion sensors based on light reflection

H. Huang & N. Gupta Department of Mechanical and Aerospace Engineering, University of Texas at Arlington, US

ABSTRACT: This paper presents the development of an optical fiber corrosion sensor based on laser light reflection. Principle of operation, fabrication technique, and experimental characteriza- tion of the corrosion sensor are presented.

1 INTRODUCTION

One of the primary damage mechanisms that limit the life of metallic structures exposing to harsh environment is corrosion. Structurally embedded sensors that detect corrosion development without disassembly could result in tremendous maintenance cost reduction and safety assurance. In this paper, we describe the development of an all-optical fiber sensor that detects corrosion based on the surface reflectivity change of a sacrificial metallic thick film. The advantages of the corrosion sensor include simple construction, inexpensive components, and compact size.

2 PRINCIPLE OF OPERATION

The optical fiber-based corrosion detecting system consists of a laser source, a sensor probe, a sacrificial film, and an optical power detector. The sensor probe is constructed from two optical fibers placed adjacent to each other. One fiber delivers the laser light to the sacrificial film and another collects the reflected light. At the presence of the corrosion, the reflectivity of the sacrificial film is reduced, resulting in a decrease in the power of the reflected light. Therefore, the sensor output servers as a good indicator for corrosion development. The sensor is packaged in such a way that only one side of the sacrificial film is exposed to the corrosive environment while the other side of the sacrificial film is finely polished and is isolated from the environment. The corrosion sensor will only detect the corrosion when it is severe enough to penetrate through the thickness of the sacrificial film. Corrosion progression can be tracked by employing corrosion sensors with different film thicknesses.

3 SENSOR FABRICATION

The sensor probe was constructed by packaging two optical fibers in a stainless steel tube and polishing it following the same preparation procedure for optical fiber connectors. The sacrificial thick film was polished and glued to the end of another stainless steel tube. The tube/film sub- assembly and the sensor probe were assembled together after properly aligning the sensor probe and the sacrificial film using translation stages.

4 EXPERIMENTAL CHARACTERIZATION

The performance of the optical fiber based corrosion sensor was validated in two steps. First, steel disks with alternating corroded and uncorroded regions were characterized using the sensor probe 653 with a benchtop set-up. The decreases in the sensor output correlated well with the positions of the corroded region. Secondly, packaged corrosion sensors were fabricated and submerged in saline solution and the sensor output was monitored on a daily basis. The sensor output was stable through the experiment until it dropped drastically on the 17th day. On dismantling the corrosion sensor, water condensations were found on the polished side of the metallic thick film. Some corrosion pits were also observed on the polished surface under a microscope.

5 CONCLUSIONS

An optical fiber corrosion sensor was developed based on the measurement of surface reflectivity change. Benchtop tests were performed and demonstrated that corrosion decreases the sensor output drastically. A flexible fabrication technique was also invented to package the sensors. The packaged sensors are made of inexpensive components and are suitable for field test. Initial test of the packaged sensor demonstrated that it is capable of detecting water ingress. Future work includes better packaging of the sensor to make it water proof and miniaturizing the corrosion sensor.

654 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Stochastic Subspace Identification (SSI) model analysis using wireless data logger system in grand bridge under wind load effect

Yu-Seung Kim & Soo-Yeol Park Korea Maintenance & Control Co., Ltd, Korea Chung-Bang Yun Korea Advanced Institute of Science and Technology, Daejeon, Korea Jun-Sung Choi Korea Maintenance & Control Co., Ltd, Korea

Although there were many monitoring systems suggested for health monitoring of Bridge, but these systems using wire cable for sensing have brought new problems as like Disturbance and Noise since the span of Bridge has been extended and the size of structure has been growing more. Specially, if the length of cable is longer than the permitted length, Line impedance would be measured according to the length of cable and signal disturbance level. The wireless data logger will show that it is the optimal solution for the long span of Bridge as compared to the wire data logger. Proving that, Acceleration data in Grand Bridge under wind load effect to get the wire and wireless data logger were compared as performing modal analysis by Stochastic Subspace Identification (SSI) method. For performing this analysis, three tests was progressed. First test is the measurement of acceleration on a free-fixed beam using the Wire and the Wireless Data Logger under the Impact input. And Second test is confirming the effect of the line impedance in the signal cable through changing the cable length. And the third is the modal test in the grand bridge. Authors have confirmed the effect of the line impedance to mode shape changing through above tests and the need of the wireless system to the measurement in field. Specially, there is requested to using the wireless system in the structure analysis under the ambient vibration more and more. The analysis in time domain base will be more affected by the noise and the line impedance and if SSI method will be applied, the wireless system must be used.

655 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Sequential health monitoring in steel plate-girder bridges by using combined vibration-impedance signatures

D.S. Hong, J.H. Park, J.T. Kim & W.B. Na Pukyong National University, Busan, Korea

In this paper, a hybrid health monitoring system for steel plate girder bridges by using sequential vibration and impedance approaches is newly proposed as shown in Figure 1. Damage types of inter- est include damage in support and flexural stiffness-loss in girder. First, theoretical backgrounds of the hybrid health monitoring system are described. The hybrid scheme mainly consists of three sequential phases: 1) damage alarming, 2) damage classification, and 3) damage estimation. In Step 1, the occurrence of damage is alarmed in global manner by using frequency response changes in the target structure. The Frequency-Response-Ratio Assurance Criterion (FRRAC), described in Eq. (4), is used to alarm damages in the entire structure. In Step 2, the alarmed damage is classified by using the change in impedance signatures. The self- diagnostic electro-mechanical impedance change, described in Eq. (6), is used to detect damage in local area near a PZT sensor which is locally sensitive to the sensor-vicinity zone. In Step 3, the classified damage is examined in detail for damage localization and severity estimation. For the ‘stiffness-loss’ case, the modal strain energyd-based damage index method, described in Eq. (12) and Eq. (13), is used to locate and estimate severity of damage. But, the algorithm is not properly selected to estimate the severity of ‘support-damage’ in support of the steel plate girder. So, this part will remain on future works. The feasibility of the proposed system is evaluated on a laboratory-scaled steel plate girder model for which hybrid vibration-impedance signatures were measured for several damage scenarios of support-damage and flexural stiffness-loss.

Figure 1. Hybrid health monitoring scheme for steel girder bridges.

656 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Smart bearings for structural behavior monitoring

Falah M. Wegian College of Technological Studies, Safat, Kuwait Gongkang Fu Wayne State University, Detroit, MI, USA Changjiang Scholar Chair Professor, Tongji University, Shanghai, China Jihang Feng WSP Cantor Seinuk, New York, NY, USA Yizhou Zhuang & Pang-jo Chun Wayne State University, Detroit, MI, USA

Bearings are important components in a structure, such as a building and a bridge. They transfer the superstructure loads to the substructure.They are also designed to accommodate relative movements between the super- and sub-structure, so that such movements or tendency of movements would not cause damage to the structure system. In addition, bearing units are fabricated in the shop not in the field, and preparing them for measurement can be simpler and easier compared with field installation of sensors. This paper presents a successful attempt to accomplish just that for structure behavior monitoring. A number of attempts have been reported in the literature towards developing smart bearings. The concept of sensored bearings was first tried in 1996. On the other hand, the instrumentation is not portable with the bearing but fixed to the structure component outside the bearing. Another research effort included fiber optic sensors in a metallic plate inserted under the bearing, to mea- sure the vertical load on the bearing. Measurements were obtained but the bearing’s monolithic configuration was changed. Another attempt developed a load cell made of a composite plate with fiber optic sensor, and then sandwiched it with two elastomer pieces. This approach separates the “bearing” into three pieces, apparently not suitable for practical application. These reported efforts failed to produce a practically useable and thus viable smart bearing to function as regular ones. This paper presents a prototype smart bridge bearing with fiber optic sensors imbedded in the elastomers. It can be used in bridges directly without additional installation and erection efforts. The prototype smart bearing was load tested in the laboratory for compression and shear. In general, an almost linear behavior between the applied force and the measured strain was observed in the tests. A highway bridge of the Michigan Department of Transportation has been planned to receive four of such bearings as part of its replacement project. A finite element model of the structure has been developed and analyzed. Results show that the strain effects of dead loads and live (truck) load are within the range and resolution of the sensors imbedded in the bearings. During the construction stages, strain readings of the bearings for beam placement, deck placement, superimposed dead load application, and truck load application will be taken and long term monitoring has been planned. The finite element analysis results show that certain reaction components are sensitive to the considered deterioration scenarios in the deck and prestressed concrete beams.

657 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Smart structural elements for the condition monitoring of bridge structures

D. Zonta, M. Pozzi & H.Y. Wu DIMS, University of Trento, Trento, Italy D. Inaudi Smartec SA, Manno, Switzerland

This paper introduces a concept of smart structural elements for the real-time condition monitoring of bridges. These are prefabricated Reinforced Concrete (RC) elements embedding a permanent sensing system and capable of self-diagnosis when in operation. Sensors are not just applied to the member, but are an integral part of the prefabricated element, influencing its design criteria, performance and detailing. The real-time assessment is automatically controlled by a numerical algorithm founded on Bayesian logic. The basic idea is that response measurements depend, on the one hand, on external actions such as temperature and loads; on the other on long term effects, such as dead load redistribution, creep, shrinkage, strand relaxation. Long term effects produce slow changes in the structural response. Therefore, the identification procedure consists in two steps: first the strain measurements recorded by the sensors are compensated by environmental actions using a probabilistic linear filter; then, the compensated strain feeds a non-linear iterative identification algorithm. The concept is to assume a set of possible scenarios for the state of the structure, including an intact condition and various combinations of damage, such as cracking or loss of prestressing. Based on these scenarios, a set of potential structural responses is defined, each described by a vector of parameters and by a theoretical model. Given the prior distribution of this vector, the method processes the measurements and assigns posterior probability to each scenario as well as updated probability distributions to each parameter. To verify the effectiveness of the technology, we produced and tested in the laboratory a reduce-scale prototype of smart beam. The specimen is 3.8 m long and has a 0.3 by 0.5 rectangular cross-section, and was prestressed using a Dywidag bar, in such a way as to control the preload level. The prototype is equipped with traditional sensors, including 12 metal-foil strain gauges and 2 thermocouples, as well as with two novel types of Fiber Optic Sensors (FOS). The first type of FOS is a multiplexed version of the standard SOFO (Surveillance d’Ouvrages par Fibres Optiques) interferometric sensor packaged in a 3-field scheme onto a composite bar. The second type of FOS is based on the direct time-of-flight measurement of pulses travelling into a fiber optic coil. Different levels of cracking were produced in the specimen through the application of a vertical load action using an hydraulic actuator: the protocol included a sequence of load-unload cycles of the same amplitude, repeated for decreasing values of prestressing. Aim of the identification was to recognize the loss of prestressing based only on the strain measurements and vertical load data. The experiment clearly shows that parameters such as loss of prestressing can be identified with a high degree of reliability. Despite the fact that the laboratory example provided is limited, procedure is not problem dependent and can be extended to a broader class of problems, including manifold scenarios, model or material uncertainties, prior knowledge of parameter distribution. With respect to classical damage detection methods, its merit is to provide not only information on the damage, but also the degree of confidence of this information. This is of paramount importance when the results of damage assessment serve as an input in a decision-making process.

658 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Sensing capability of electromagnetic induction system for vibration control of structures

H.J. Jung, D.D. Jang & H.J. Lee KAIST, Daejeon, Korea S.W. Cho Samsung SDS, Seoul, Korea J.H. Koo Miami University, Oxford, OH, USA

Semi-active control systems have recently received considerable attention for vibration mitigation of structures in the field of civil engineering.A magnetorheological (MR) damper-based semi-active control system can be considered as one promising candidate for large-scale structures. Recently, the smart passive system, which consists of an MR damper and an electromagnetic induction (EMI) system consisting of permanent magnets and a coil, has been developed for replacing a feedback control system including a power supply, a controller, and sensors with the EMI system. The EMI system converts mechanical energy (i.e., reciprocal motions of an MR damper) into electrical energy (i.e., electromotive force or induced voltage) according to the Faraday’s law of electromagnetic induction. It is expected from several previous investigations that the smart passive system is suitable for mitigating the vibration of civil engineering structures. From a different perspective, on the other hand, the EMI system could be a kind of response sensing devices for the MR damper-based semi-active control system as well as an alternative power supply. That is because the induced voltage from the EMI system is linearly proportional to the relative velocity across the MR damper according to the Faraday’s law. In addition, some control algorithms used in the MR damper-based semi-active control systems require the measurement information on the response related to the relative velocity of the damper. In this study, the sensing capability of the EMI system is preliminarily examined for an application to the MR damper-based semi-active control system. To do this, a series of shaking table test are carried out using the EMI system designed by the authors. Various base excitation inputs are considered, and the relative velocity of the damper and the induced voltage form the EMI system are compared. Through analyzing the experimental results, its feasibility is also discussed. From the preliminary shaking table tests,

Figure 1. EMI system. Figure 2. Relative velocity versus induced voltage.

659 it is verified that the EMI system could be used as a velocity-sign sensor needed for the MR damper-based semi-active control system employing some specific control algorithms such as the maximum energy dissipation algorithm. The additional research on the applicability of the EMI system to actual civil engineering structures is in progress by carrying out the extensive shaking table tests.

660 Special session on Incheon Bridge

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Incheon Bridge project outline

J.H. Yang Son, M.G. Yun, H.S. Kim, H.Y. Shin & I.S. Shim Samsung Engineering & Construction, Seongnam-si, Korea

Incheon Bridge with a total length of approximately 12 km is the second marine bridge link- ing Incheon Free Economic Zone (Songdo New City) in Incheon city to Youngjong Island where Incheon International Airport is located, followedYoungjongBridge which opened to traffic in year 2000. Currently in Songdo, many of infrastructure construction works including skyscrapers are in progress to establish the foundation in turning Songdo into an international city. The Cable- Stayed Bridge (CSB) to be located in the navigational channel (Width 625.5 m, Height 74.0 m) is designed as the dominant large-scaled construction well-assorted to the Incheon City’s aim to be one of the well advanced high-tech international cities, and its detailed design work in stages is making steady progress as well as construction works. (Please refer to the figures below, Figure 1 – Incheon Bridge Project, Figure 2 – Incheon Bridge Overview, Figure 3 – Sectional View of Incheon Bridge) This project composes of the Cable-Stayed Bridge with 800 m main span of five(5) continuous steel box girders, Approach Bridge of seven(7) continuous PSC box girder (length of girder, 145 m) in Rahmen type, and Viaduct of five(5) continuous PSC box girder (length of girder, 50 m), as well as Toll Plaza in Youngjong Island and Ship Impact Protection (SIP) to secure the safety of passing vessel. This project commenced in July 2005 from the piling work, and will be completed in October 2009. The pile foundation and pile cap work for CSB section is completed in year 2006, and currently the pylon/pier work and steel box girder manufacture work is ongoing. For Approach Bridge and Viaduct, the superstructure work such as girder manufacture work and girder installation is in progress. Meanwhile, at the Toll Plaza, the earthwork is started and being carried out, and the design work for other ancillary works such as access to be necessary for inspection is in progress. In this paper, the outline of Incheon Bridge and the project implementation system are briefly presented.

Figure 1. Incheon Bridge Project.

663 Figure 2. Incheon Bridge Overview.

Figure 3. Sectional View of Incheon Bridge.

664 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Design and construction of approach bridge in Incheon Bridge project

Jong-Young Song, Kyu-Yong Choi, Hyun-Yang Shin & Wan-Su Lee Samsung Corporation, Korea Byeong-Cheol Cho & Dae-Won Hwang Seoyeong Engineering, Korea

The Incheon Bridge will link Incheon International Airport located on Yeong-jong Island to New Songdo City. The 12 km sea crossing bridge, the longest in South Korea, consists of three different types of bridge. The main bridge is a 1.5 km Cable Stayed Bridge with steel deck; Approach Bridge is a 1.8 km precast PSC box girder bridge, and Viaduct is an 8.7 km precast PSC box girder bridge. The main span of the Approach Bridge is the longest 145m, precast PSC box-girder bridge. Preliminary and detail design were carried out to meet the requirements of both Korean Bridge Design Code and AASHTO LRFD. Before final closure, jacking force is applied longitudinally to counteract the deformation-induced forces due to creep and shrinkage. Total 836 numbers of small segments ranging from 7.2 m to 3 m in depth will be erected using balanced cantilever method by specially designed derrick crane for this project. Also, pier table of superstructure was designed as 20-meter long precast unit to reduce construction time and erected by 3000 tonnage capacity floating crane. Precise geometry control in both fabrication and erection is an important requirement for bridge constructed by balanced cantilever method. In this paper, the key aspects of design and construction of Approach Bridge are introduced.

Figure 1. Profile of Incheon Bridge.

Figure 2. Key segment installation.

665 Figure 3. Recent scene of West Approach Bridge.

666 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Development of geometry control system for cable-stayed bridges and application to the Incheon Bridge

K. Jung & H.S. Lee Seoul National University, Seoul, Korea

A geometry control system for cable-stayed bridge, which can effectively control the cable tension and geometry during construction, is developed and proved its practical validity by applying to Incheon Bridge. The structural system of cable-stayed bridge changes according to the progress of construction. Then errors in geometry and cable force may be accumulated and amplified through complicated construction steps. Although the design is accurate and correct construction is done, a certain amount of errors in cable tension and geometry is inevitable. Thus, the feedback process of measurement and control system is required for more accurate construction. In this study, the feedback process of measurement and control system is referred as geometry control system. The purpose of the geometry control system is eliminating deviation between the object and actual structure. By applying the geometry control system to the cable-stayed bridge during the construction and completion, the structure satisfies the target configuration. To control the config- uration and cable tensions under the construction of Incheon Bridge, two different methods will be applied. The first one is cable length adjustment. Since the cable-stayed bridges resist the exter- nal force almost by the cable member force, adjusting the length of cables is one of the simplest ways to control the configuration of the bridge. The amount of cable length adjustment is calculated using an optimization method, which minimizes the errors between the measured configuration and target configuration. The adjustment for minimizing configuration errors may yield meaningless solutions in optimization process due to the instability, which is triggered measured configuration polluted by noise. The regularization technique is considered to overcome the instability. And the other one is system identification. In case the only cable adjustment system can not control the cable tension and geometry within the allowable limits due to the inaccuracy of the assumed analysis model of the bridge, the system identification scheme is utilized to update the properties of analysis model. The error function is defined as the least square errors between calculated configuration by assumed analysis model and measured ones. A regularization scheme is adopted to alleviate the instability of minimization problem. The sensitivity analysis for optimization method and inverse analysis is evaluated by the direct differentiation of equilibrium equation. The proposed method is applied to the Incheon Bridge to demonstrate its validity and effectiveness.

Keywords: geometry control, optimization, system identification, sensitivity analysis

667 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Seismic design and performance assessment of pile-bents in Incheon Bridge viaduct

Hyeok-Soo Son & Myung-Seok Oh Seoyeong Engineering Co., Ltd., Seoul, Korea Kyoung-Lae Park & Jong-Ho Yang Samsung Engineering & Construction, Seongnam-si, Korea

An alternative to the conventional pile-supported foundation, which has three kinds of elements; a column, a pile-cap, and a pile, is the integral pile shaft-column, so called “pile-bent”, which has a continuous single column-pile without a pile-cap. This system can have significant construction cost savings and reduce some construction problems such as reinforcement congestion and compli- cated detailing at column-footing connection when compared to groups of smaller piles integrated with a pile cap. The main objective of this paper is to represent the seismic design of pile-bents in Incheon Bridge Viaduct, which is under construction in Korea. Especially, for the pile-bents, the overall safety against seismic events will be governed by the structural displacement rather than strength capacity. Therefore, it is very important to estimate its seismic performance based on dis- placement design approach. To estimate the seismic performance of pile-bents, nonlinear push-over analysis on simplified model based on equivalent cantilever column approach was performed. And then, seismic performance assessment was conducted by using ADRS (Acceleration Displacement Response Spectrum) method. From the analysis and assessment results, pile-bents show adequate performance and safety under the given seismic conditions. In addition, the structural analysis method applied to the pile-bents, reinforcement details in accordance with AASHTO LRFD Bridge Design Specifications, and brief introduction of pile-bent construction are introduced.

668 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Geometry control for the concrete pylon of Incheon Cable Stayed Bridge

D.K. Im, J.G. Yoo, C.H. Kim & H.S. Kim Samsung Corporation (Incheon Bridge Project), Incheon, Korea

Incheon Bridge is the longest cable-stayed bridge with 800 m center span in Korea. The pylon is shaped of Inverted Y Shape with 238.5 m height and made of concrete. 1- tie and 3 nos. of struts are installed and stressed by 2000 KN to reduce the bending moment of pylon leg as cantilevered erection. The lower cross beam is precasted at casting yard and installed on the temporary bracket attached to the pylon leg. The geometry control of pylon consists of stage-by-stage analysis, camber evaluation, alignment monitoring and adjustment. In this study, the geometry control applied to the pylon construction of Incheon cable stayed bridge is presented. Stage-by-stage analysis is carried out by using PCCAP-II software which was developed by Samsung Corporation for analysis of cable stayed bridge and suspension bridge. The camber of the top of pylon is 86 mm in vertical and max. horizontal camber is 88 mm in the middle legs. The survey for geometry control is normally performed at night time or before dawn to minimize the effect of the temperature gradient. The geometry control for the pylon of Incheon Bridge are carried out successfully within the tolerances under the effectively reduced cycle time.

669 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Case study of Osterberg-Cell pile load test on large diameter drilled shaft in Incheon Bridge project

Seung-Hyun Shin, Yang-Ku Lee, Zu-Cheol Kim, Jeong-Hwan Kim & Hyun-Gun Lee Samsung Corporation, Incheon, Korea

Incheon Bridge Project is the construction of marine highway bridge with total length of 12.343 km, linking New Songdo City to the northen east side of connecting road of Incheon International Airport over the navigational channel of Incheon Harbor. The project involves a 1.48 km-long cable-stayed bridge section where the section has the highest clearance of navigational channel, a 1.78 km-long approach bridge section where the section is to connect viaduct with CSB section with ascended vertical curve, and 8.4 km-long viaduct where the section is connected to approach bridge from both coastal sides with lowest level height. For foundation piles, RCD Pile was applied as being supported by bearing capacity of weathered rock, soft rock or hard rock. The pile diameter was different for each section, i.e. 1.8 m for Viaduct, 2.4 m forApproach Bridge and 3.0 m for CSB section. In accordance with the CSR (Concessionaire Supplementary Requirements), 4 real-scaled load testing piles were applied to the pile load test and its results was included in the design. Full-scaled load testing was carried out with Osterberg Cell Load Testing Method. Maximum load of 28,958 ton (World Record) was loaded on 3 piles with 2.4 m diameter and 1 pile with 3.0 m diameter. This study is to introduce the process of construction of real-scaled load testing pile/performance of full-scaled load testing on the marine site which were successfully carried out.

670 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Design of ship impact protection in Incheon Bridge

J.H. Kim & H.Y. Shin Samsung Corporation, Seoul, Korea H.T. Kim & S.H. Lee Seoyeong Engineering Co., Ltd, Seoul, Korea

The Incheon Port area, where Incheon Bridge is being constructed, is the key to Seoul, the capital city of Korea. In 2002, the vessel traffic volume passing through this area reached more than 40,000. In designing this sea-crossing structure, therefore, it is critical to secure safety of Incheon Bridge by protecting from risk of vessel collision. For Incheon Bridge, several large diameter protection dolphins would be placed around the pier so that the structure can be present in safety from the ship impact. Each dolphin consists of driven straight web steel sheet piles, filled with crushed rock and topped by a concrete cap. The dolphin is considered as the plastic/large deformation sacrificial structures which have to be replaced after a major collision. In this paper, detailed design of Ship Impact protection for Incheon Bridge is introduced. The energy absorption ability of the dolphin, which is an essential capability for SIP design, is investigated using ABAQUS of general-purpose finite element analysis program and the numerical results from ABAQUS is verified by Centrifuge Testapplying a geometric model scale of 1:200. Finally, the response characteristics of the protection structure obtained for the FEM analysis are incorporated in simulation and the appropriateness of the protection layout is verified by the simulation.

671

Structural health monitoring on cable supported bridges

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Development of bridge WIM systems without axle detector using artificial neural network

Min-Seok Park ETRI, Korea Expressway Corporation, Gyeonggi-do, Korea Jungwhee Lee Research Institute of Industrial Science and Technology, Gyeonggi-do, Korea Byung-Wan Jo Hanyang University, Seoul, Korea Sungkon Kim Seoul National University of Technology, Seoul, Korea

This paper describes the analysis of vehicular loads reflecting the domestic traffic circumstances is necessary for the development of design live load models in the analysis and design of highway bridges or the development of fatigue load models to predict the remaining lifespan of the bridges. This study intends to develop a ANN (Artificial Neural Network)-based method for the analysis of data to obtain information concerning the loads and running conditions of vehicles crossing bridge structures exploiting the signals measured by strain gauges installed at the bottom of the bridge superstructure. This study relies on experimental data corresponding to the crossing of hundreds of random vehicles rather than on theoretical data obtained through numerical simulations to secure training data for the training and test of theANN. In addition, data acquired from 3 types of vehicles weighted statically at measurement station and crossing the bridge repeatedly were also exploited to examine the accuracy of the trained ANN.

675 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Development of local live load truck model for long span bridges based on BWIM data of Seohae cable-stayed bridge

Min-Seok Park ETRI, Korea Expressway Corporation, Gyeonggi-do, Korea Chan-Hee Park & Jungwhee Lee Research Institute of Industrial Science & Technology (RIST), Gyeonggi-do, Korea

Using Bridge Weigh-In-Motion (BWIM), unbiased sample truck data were collected for trucks on a cable-stayed bridge. Collected data from BWIM system located at Seohae Cable stayed bridge were statistically analyzed. Measured trucks were segmented into 3, 4, 5 axle trucks and for each truck category statistical distribution of axle weight and distance between axles were analyzed. Load effect (moments) from trucks obtained from BWIM are calculated. Each truck is loaded on Seohae bridge and maximum moment from structural analysis using SAP 2000 was obtained. From

Figure 1. Distribution of moments and maximum mean moments due to trucks on Seohae bridge.

Figure 2. Truck Model (Sample) for Seohae Bridge.

676 the analysis, it is shown that distribution of total weight differs from that of load effect of trucks. Therefore load model should be derived from load effects of the measured trucks. Extreme distribution is assumed for upper level of load effects distribution. A procedure is given to evaluate extreme load effect from limited collected data. Load effects for each collected truck are calculated and cumulative distribution function of the load effect is determined. Maximum mean load effect is determined for upper level of load effect and load model will be derived. Example is shown to demonstrate the model developing process based on sample BWIM data at Seohae bridge.

677 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Long-term structural behaviors of Seohae cable-stayed bridge based on results from SHM and surveys

J.C. Park, C.M. Park, M.S. Park & I.K. Lee ETRI, Korea Expressway Corporation, Gyeonggi-do, Korea B.W. Jo Hanyang University, Seoul, Korea

In this paper, the long-term structural behaviors of the Seohae cable-stayed bridge are presented. Specially, it focuses on the inclinations of pylons, the geometrical variations through extensive surveys, and the dynamic characteristics. These data have been collected for 7 years with the SSHMS and surveys. To extract trend components from the measured inclinations, Singular Spectrum Analysis (SSA) was applied. From the trend curves, it is clear that the long-term behavior of pylons shows a tendency to incline toward the main span. The variation in the longitudinal displacement at the top of the pylon was +79.5 mm at PY1 and −29.1 mm at PY2 for 7 years. Geometry surveys together with the SSHMS had been performed in 2001, 2002, 2006 and 2007 extensively. Overall heights at the main span tended to low and the peak point moved from No. 45 (2001) to No. 48 (2007). The approximately +30 cm upward camber in 2001 at the mid-span – compared to design – showed to low to +20 cm by −10 cm in 2007. It is judged that the trend variations of inclinations and camber mainly induced by the effect of creep and shrinkage of the structure. Structural analysis and design information conformed this tendency. Using measured natural frequencies and temperatures, the long-term variations of dynamic characteristics of the bridge were estimated. As the first natural frequency, fv1, averages 0.260 Hz ranging from 0.260 ∼ 0.261 Hz and the second natural frequency, fv2, averages 0.328 Hz ranging from 0.328 ∼ 0.329 Hz, there were no yearly variations of natural frequencies. It means that the bridge has been shown a healthy structural behavior although the inclinations of pylons and the camber of superstructures were slightly changed. Through relations both the monthly mean natural frequencies and temperatures, it can be known that there was a very high linear relation with negative slope between the two variables. To evaluate analytically the long-term structural behaviors considering the effects of temperature, creep and shrinkage, several researches to construct the analysis model have been performed now.

Figure 1. Longitudinal displacements at top of pylon. Figure 2. Comparisons of camber variations.

678 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Modal parameter extraction of Seohae cable-stayed bridge using TDD technique

Byeong Hwa Kim Department of Civil Engineering, Kyungnam University, Masan, Kyungnam, Korea Jong Chil Park, Min Seok Park & Il Keun Lee ETRI, Korea Expressway Corporation, Gyeonggi-do, Korea

The TDD technique that is an output-only modal parameter extraction method has been applied to the extraction of modal parameters of Seohae cable-stayed bridge. For a set of acceleration responses measured from the total 72 sensor points on the deck plate of the bridge in the vertical direction during 3 hr, the 24 high resolution mode shapes are extracted by the TDD technique. The results have been compared to those of previous study. Based on the results, the following three conclusions can be made. First, the important lower modes of deck plate on the most of long span bridge lies in the range of 1 Hz ∼ 2 Hz while the cable frequencies lies in the range of 1 Hz ∼ 20 Hz while they depends on many design parameters. Thus, in order to extract the accurate modal parameters from ambient modal testing, the sampling rate should be differently selected with respect to each structural member, and the sampling dura- tion should be very large enough. Second, the measurements of high resolution mode shapes are needed in the long span bridges because such mode shapes can be used to examine the working condition of supports, to estimate healthy condition, and to update the numerical model. Third, the TDD technique is most suitable for the extraction of the high resolution mode shapes in real-time condition. As presented in the paper, the TDD technique is distinguishable among many existing ambient modal parameter extraction techniques due to its robust and efficient algorithm.

679 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Analysis model updating of the Seohae cable-stayed bridge

H.K. Kim, S.D. Park & K.T. Kim Mokpo National University, Jeonnam, Korea W. Park, S.H. Lee & J.F. Choo Seoul National University, Seoul, Korea J.C. Park & M.S. Park Korea Highway & Transportation Technology Institute, Gyeonggi, Korea

The Seohae Bridge is a composite cable-stayed bridge with main span of 470 m. The bridge was opened in 2000 to traffic in Korea. In order to secure a more reliable dynamic analysis model for Seohae Bridge, a model updating procedure was presented. The procedure used a carefully constructed baseline FE model based on various measured data under construction as well as design data. Modal information based on acceleration and displacement measurement was also used for the reference values for the updating procedure. For a baseline FE model, grillage model was adopted with three dimensional frame elements. In this study, a sensitivity based automatic model updating procedure was presented, which solves an optimization problem for model error minimization. Ten vertical vibration modes and nine design parameters was selected for the problem. Updated results showed that the model error could be reduced from 5 ∼ 15% to 0 ∼ 6% in terms of natural frequency ratio between the model and measured data. During the optimization procedure, the target error bounds were 3% for the lower vertical modes and 6% for the horizontal modes, respectively. In order to prevent mode interchange due to the closely spaced frequencies of large three dimensional FE model, Modal Assurance Criteria (MAC) were also introduced to verify the updated results through the proposed optimization procedure.

Table 1. Comparison of natural frequencies between baseline and updated FE model for Seohae Bridge.

Baseline model Updated model Measured No. frequency (Hz) Frequency (Hz) Diff. (%) Frequency (Hz) Diff. (%) Mode shape

1 0.2557 0.2383 −6.80 0.2582 0.98 1st vertical 2 0.3256 0.2909 −10.66 0.3190 −2.03 2nd vertical 3 0.3725 0.3197 −14.17 0.3502 −5.99 1st lateral 4 0.4555 0.4268 −6.40 0.4566 0.24 1st torsional 5 0.5290 0.4896 −7.45 0.5336 0.87 3rd vertical 6 0.5762 0.5235 −9.15 0.5767 0.09 4th vertical 7 0.6250 0.5721 −8.46 0.6277 0.43 5th vertical 8 0.7171 0.6798 −5.20 0.7592 5.87 6th vertical 9 0.8140 0.7403 −9.05 0.8081 −0.72 7th vertical 10 0.8333 0.7571 −9.14 0.8255 −0.94 3rd torsional

680 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Development of prediction method of non-linear observed data from cable-stayed bridge using support vector regression

Mi-Yun Park Infra Asset Management Co, Seoul, South Korea Hyo-Nam Cho Hanyang University, Seoul, South Korea Kwon-Woo Park Unicons Co, Seoul, South Korea Jong-Chil Park, Min-Seok Park & Il-Keun Lee Expressway & Transportation Tech.INs, Seoul, South Korea

ABSTRACT: The SVR is “SRM (Structural Risk Minimization)” technique which could reduce the generalization error and especially can be easily extended to the non-linear properties of data using radial kernel function. Moreover, RBF (Radial Basis Function) in kernel function is applied in order to improve the performance of the generalization error. In the paper, the SVR method is applied to the observed data from a real monitoring system of a cable-stayed bridge for the prediction of deflection of girders and displacement of towers due to the change of temperature over a number of years, which were performed by using tíme-series analysis about the real observed data. Finally, the comparison between the observed data and the prediction data was made by evaluating the error ratio, and the efficiency with the applicability of the proposed SVR method is demonstrated.

1 EAMPLE TEST OF GSL VOLUME USING SVR

SupportVectorRegression (SVR) is a regression analysis in the SVM field and it has been proposed by Vladimir Vapnik ,Harris Drucker, Chris Burges, Linda Kaufman, Alex Smola, etc. in 1996. The proposed model of Support Vector Classification (SVC) relies only on partial set of a training data because of the high expense on composing the model. In addition, the model created with only SVR (Support Vector Regression) also relies on partial set of a training data due to free charge of composing the model which can also it can easily make model estimation. Diverse loss function

Figure 1. Prediction of GSL volumn using SVR.

681 Table 1. Comparison of SVM and SVR to predict great salt lake volume.

4-step interval prediction SVM SVR Difference(%)

Coefficient of efficiency 0.981 0.989 0.82(%) Index of agreement 0.920 0.952 3.48(%) Correlation Coefficient 0.977 0.987 1.02(%)

Figure 2. Monitoring data in Expansion Joint. Figure 3. Prediction data in Expansion Joint.

Figure 5. Atmosphere temperature correlation of Figure 4. Normalized error of prediction data. prediction data and real data.

can be used for SVR. Loss Function here is the function that tells which value to insert when there is value difference between the estimation value and real value. Therefore, in this research, through examples of SVR method, estimation efficiency was investigated and made a prediction on the Seohae Grand Bridge monitoring data. The applied example is to prove the efficiency of SVR method by using the time series data of GSL (Great Salt Lake) Volume. As GSL (Great Salt Lake) Volume possessed as a low dimensional non-linear kinetic behavior, a non-linear study and a short term test possibility has been examined.

2 THE PREDICTION OF MONITORING DATA SEOHAE CABLE-STAY BRIDGE

In order to estimate the displacement and conduct according to temperature variance of cable stayed bridge, SVR method was applied. Prior to the execution of this technique, relations between the monitoring value and estimation value of the previous temperature variation has been gathered as well as the error rate. The monitoring data has eliminated the noise signal in which the omitted portion of the connection on the displacement quantity according to temperature variation can be seen. Therefore in the similar state of prediction data it can be known by prediction error. 682 Table 2. Result of prediction of displacement of expansion joint 39.

RMSE 4.057 Index of agreement 0.985 MAE 1.891 Coefficient of efficiency 0.997

3 CONCLUSION

This research has been carried out in 3 aspects; short-term data prediction, long-term data pre- diction, and prediction made on outdoor temperature and its structural variation. Summing up the previous result, prediction method using SVR related to monitoring data of Saehae-bridge proved more effective in aspect to correction.

683 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

SHMS and wind engineering on the Busan-Geoje Fixed Link bridges

Y.M. Kim, D.Y. Kim & C.H. Kim DAEWOO E&C, Suwon, Korea A. Galmarini, P.D. Frederiksen & J.E. Andersen COWI A/S, Lyngby, Denmark

The Busan-Geoje Fixed Link is a major infrastructure development in the south-eastern part of Korea. The link will – when completed in 2010 – provide easy access between metropolitan Busan and Geoje Island. The Busan-Geoje Fixed Link has a total length of 8.2 km and comprises three major elements: two cable stayed bridges and an immersed tunnel. In addition, roads and bored tunnels are provided on the two intermediate islands to connect the major elements as well as buildings for operation and maintenance and toll stations. The project is developed as a Public- Private-Partnership project, where GK Fixed Link Corporation has been awarded the concession to design, construct and operate the Link for 40 years. DAEWOO E&C Co., Ltd. is the leading company of the concessionaire. This paper describes the design of the bridges with special focus on SHMS regarding wind loadings including wind tunnel tests. SHMS (Structural Health Monitoring System) is planned to be installed to monitor the short- term and long-term behaviours during construction and operation. The sensory system consists of 262 sensors and their relevant interfacing units. The sensors include anemometers, temperature detectors, accelerometers, strain gauges, level sensing stations, displacement transducers etc. They measure everything from temperature and strains in structural members to wind speed and the deflection of the cables and any movement of the bridge decks and towers. These sensors are the early warning system for the bridges, providing the essential information to accurately monitor the general health monitoring conditions of the bridges. The overall shape of the deck, a concrete slab resting on a longitudinal I-beam on the either side and a set of cross beams was chosen for easy fabrication and erection. However, such plated girder designs are known to be susceptible to flutter instability and vortex shedding. Therefore, a series of wind tunnel tests and numerical analysis were carried out to determine if countermeasures were required. The acquisition of SHMS data is necessary to verify stochastic load parameters and structural responses in comparison with calculated response since the aerodynamic behaviours induced by wind can be difficult to predict during the design stage.

684 Structural robustness

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Robustness investigation of long suspension bridges

F. Bontempi & L. Giuliani Department of Structural and Geotechnical Engineering, University “La Sapienza” of Rome, Italy

ABSTRACT: A structural system can be defined as complex, if its behavior is influenced from nonlinearities, uncertainties or interactions. In order to govern the complexities described above, it is important to define a conceptual framework that can be adapted from the concept of dependability and dependable system defined in the field of electronic and systems engineering and extended to the structural field. A comprehensive array of definitions is given and the role played by dependabil- ity in structural engineering is discussed. Particular attention is given to the dependability attribute of robustness that indicates the ability of a structure to preserve a certain level of structural integrity after a failure. Within a robust structure the damage is a bounded damage and has no propagation, i.e. the entity of damage is proportional to the amplitude of its cause. With particular reference to the evaluation of the structural robustness, the exposed concepts of structural dependability are applied to the case of a long span suspended bridge of unique characteristics. A set of nonlinear static analyses are performed on different damaging conditions affecting the bridge suspension system. The considered load scenario refers to the presence of a LM71 train. This is a heavy weight freight train that has been modeled by means of a distributed load of 88 kN/m for a length of 750 m, disposed around one third of the main span. The aim of the analysis is to evidence the structure strength decrement (represented by the collapse load multiplier) for a loss of hangers starting from the one around the mid length of the train. The following contingency scenarios of an asymmetrical loss (i.e. affecting one side of the bridge only) of 1, 2, 3, 4, 5, 6 hangers, and of a symmetrical loss (affecting both sides of the bridge) of 1 + 1, 2 + 2, 3 + 3 hangers, are considered. A review of the investigations outcomes permit to remark the following points: 1. the load multiplier for the nominal configuration is coherent with the assumed limit state design philosophy that requires the contemporaneous presence of two LM71 with a safety factor of 1.5: in fact, 2 × 1.5 = 3 that is less that the obtained load multiplier of 3.2313; 2. the system appears reasonably robust for the first damage levels in fact, it can absorb a single hanger loss without essential decrease of structural strength, while a remarkable decrease of capacity can be observe with the loss of at least three hangers; 3. the behavior of the single hangers, i.e. ductility/fragility, is small until the case of loss of 5/6 hangers: in these cases, the effect is noteworthy; 4. from the general point of view, the torsional stiffness of beams of the gridwork is remarkable. Furthermore some evaluation on the nature of the structural failure can be deduced: in partic- ular, from an observation of bending moment distribution on the longitudinal beams and on the transversal one, the collapse mechanism can be deduced. In case of a single hanger failure, the collapse mechanism involves locally the railway girder under the location of the LM71 train, while the longitudinal beams are still resisting: it seems like that the LM71 punches the compounded deck. With an asymmetrical loss of more hangers, the picture is somewhat altered, since the longi- tudinal beams are more involved in the failure. Eventually, in the cases of a symmetrical damage the involvement of the longitudinal beams is more marked.

687 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Dynamic analysis for structural robustness evaluation

L. Giuliani & F. Bontempi Structural and Geotechnical Engineering Department, University of Rome “La Sapienza”, Italy

ABSTRACT: The robustness of a structure, intended as its ability not to suffer disproportionate damages as a result of a limited initial failure, is an intrinsic requirement, inherent to the structural system organization. A robust design cannot therefore be limited to design the structure just for additional accidental load cases, but should take into account the response of the structure to different damage conditions. Complex structural systems like the long suspended bridge considered in this paper, in case some failures involving the suspension system occur, can be inclined to the propagation of the damage, due to the high continuity of the long central span. Some specific investigations aimed to robustness assessment have therefore to be carried out on the structure: following a bottom-up approach the critical event modeling can be neglected and an initial failure has to be a-priori assumed on the structure, in order to evaluate if and how much the overall response of the structure is influenced from a local failure. The investigations presented in this paper start from assuming an abrupt rupture of some hanger ropes of the suspension system and investigate the response of the system to different damage levels and locations. This kind of investigations require quite complex nonlinear dynamic analyses to be performed, in order to account for the dynamic amplification caused from such abrupt initial failures and for the plastic reserves of the materials that the system can exploit during the collapse. Furthermore, the triggering of local mechanism should be evaluated, since it would not necessarily indicate a complete loss of resistance of the structure as a whole and could possibly even lead to a collapse standstill. The review of the results provided from the performed investigations permits to identify some characteristics in the behavior of the bridge after a damage in the hanger suspension system, that are intrinsic to the structural system: the bridge result to be more sensible to the damage at mid-span, where a lower number of failed hangers is needed in order to trigger a chain rupture with respect to the bridge sides. This higher damage sensibility of the bridge central zone counterpoises a lower acceleration of the collapse progression triggered by central ruptures, with respect suspension system that requires a growing hanger length from the centre to the sides of the bridge: when a chain rupture trigger, the ultimate elongation required to the hangers adjoining the failed ones increases as the collapse propagates (because the unsupported deck length also increases). If the initial damage occurs at mid-span, it involves the shortest hangers and the collapse propagation is partially slowed down from the growing element ductility of sideward hangers. On the contrary, a more intense initial damage is required sideways to trigger chain ruptures, but then the hanger breakdowns speeds up when moving toward the centre, where the hanger length decreases. If in the first case a closer increment in the section of the hangers (that remain instead the same for about 5/6 of the span length) could possibly provide for a collapse standstill, in the second case the progressive collapse shows a preferential direction, making probably less effective such a measure. Another consideration about the possible collapse standstill concerns the higher susceptibility of the bridge to an unsymmetrical hanger failure than to a symmetrical one: in the last case the symmetrical hinge formations provide for a symmetrical moment increment on the deck box-girders, thus possibly allowing for an early deck segment detachment that would arrest the collapse.

688 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Measure of structural robustness under damage propagation

Fabio Biondini & Stefano Restelli Department of Structural Engineering, Politecnico di Milano, Milan, Italy

ABSTRACT: Structural robustness is generally defined as the ability of a system to suffer an amount of damage not disproportionate with respect to the causes of the damage itself. However, the concept of robust structures is still an issue of controversy, since there are no well established and generally accepted criteria for a consistent definition and a quantitative measure of structural robustness. In general, a measure of structural robustness should arise by comparing the structural performance of the system in the original state, in which the structure is fully intact, and in a perturbed state, in which a prescribed damage scenario is applied (Frangopol and Curley 1987, Restelli 2007, Biondini et al. 2008). To this purpose, performance indicators associated with ulti- mate conditions are of great importance in robustness evaluations associated with damage suddenly provoked by accidental actions, like explosion or impacts. However, damage could also arise slowly in time from aging of structures, as induced for example by environmental aggressive agents. In this context, performance indicators associated with serviceability conditions may become of major importance in life-cycle robustness evaluations. In this paper, a general approach to robustness analysis of structural systems undergoing damage is presented. The deterioration effects on the system performance are evaluated with reference to suitable performance indicators identified with meaningful parameters of the structural response. The variation of these indicators with respect to the values associated with the performance of the undamaged system is used to formulate dimensionless measures of structural robustness. Moreover, an index of structural integrity aimed to quantify the severity of the structural failure with regards to its consequences, is proposed. In the proposed approach, damage is viewed as a progressive deterioration of the material properties and its amount is specified at the member level by means of a damage index associated with prescribed patterns of cross-sectional deterioration. Starting from this local definition of damage, a model of damage propagation at the system level is also developed by using a damage- sensitive fault-tree analysis. In such a way, all the possible damage paths associated with the actual topology of the whole structural system are described by branched networks where the degree of activation of each nodal connection is properly tuned to account for the prescribed level of local structural damage. The effectiveness of the proposed measures of structural robustness is shown through the appli- cation to a truss system. An application to a bridge structure is also presented to highlight the usefulness of the proposed approach also in the context of robust design of bridges.

REFERENCES

Biondini, F., Frangopol, D.M., Restelli, S., (2008). On Structural Robustness, Redundancy and Static Indeterminacy, ASCE Structures Congress 2008, Vancouver, Canada, April 24–26. Frangopol, D.M., Curley, J.P., (1987). Effects of Damage and Redundancy on Structural Reliability, Journal of Structural Engineering, ASCE, 113(7), 1533–1549. Restelli, S., Measure of Structural Robustness of Deteriorating Systems, Degree Thesis, Politecnico di Milano, Milan, Italy, 2007 (In Italian).

689 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Robustness assessment of a cable-stayed bridge

M. Wolff & U. Starossek Structural Analysis and Steel Structures Institute, Hamburg University of Technology, Germany

The failure of one structural element can lead to the failure of further structural elements and thus to the collapse of large structural sections or the entire structure. Such disproportionate collapses have been discussed and investigated for some years, but mainly for buildings. Generally, structures can be made collapse resistant by ensuring a high level of safety against local failure or by increasing their robustness. Robustness has been defined as the insensitivity to local failure where the size of the local failure has to be determined by the design objectives (Starossek 2006). For cable-stayed bridges, collapse resistance is primarily achieved by increasing the robustness. The loss of cables must be considered as possible local failure since the cross sections of cables are usually small and therefore possess a low resistance against accidental lateral loads stemming from vehicle impact or malicious action. Current recommendations for cable-stayed bridges constitute the sudden loss of one single cable; other authors assume the sudden loss of all cables in a 10 m range. The loss of cables can lead to overloading and rupture of adjacent cables. A collapse progressing in such a way is called a zipper-type collapse. Because the stiffening girder is in compression and a cable loss reduces its bracing against buckling, the collapse is accompanied and reinforced by a progressive destabilization of the structure. To trace the collapse progression following the rupture of one or more cables, dynamic non-linear analyses in the time domain are necessary. When designing for the loss of a cable, only the maximum responses are of interest. For this, quasi-static analyses using a dynamic amplification factor to account for the dynamic effects are advised. This approach is also advised by current recommendations for cable-stayed bridges. This paper examines the structural response of a cable-stayed bridge to the loss of one cable by means of dynamic analyses including large displacements. The effects of cable sag, transverse cable vibrations and structural damping are determined. Dynamic amplification factors are computed and limits of the quasi-static approach are outlined. Finally, conclusions are drawn as to the collapse behavior of cable-stayed bridges. Ultimate states are calculated taking into account large displacements and non-linear material behavior.

REFERENCE

Starossek, U. 2006. Progressive collapse of structures: Nomenclature and procedures. IABSE, Structural Engineering International, 16(2). http://www.sh.tu-harburg.de/starossek/Index.htm

690 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Approaches to measures of structural robustness

U. Starossek & M. Haberland Hamburg University of Technology, Hamburg, Germany

SUMMARY: The topic structural robustness attracts increasing attention in context of progressive or disproportionate collapse. In this paper, basics for the development of measures to quantify structural robustness are formulated; requirements for measures as well as possible applications are suggested. Some simple formulations of measures are presented and discussed.

EXTENDED ABSTRACT: Robustness has been recognised as a desirable property of structural systems which mitigates their susceptibility to progressive disproportionate collapse. At present, there is neither a uniform theory of robustness and progressive collapse nor any agreement on terms and nomenclature. Codes and other relevant publications require robust structures. However, mainly qualitative and hardly quantitative recommendations are provided. In this paper, robustness is defined as the insensitivity of a structure to local failure. To examine a structure in terms of its robustness, a quantitative description by means of a measure would be useful. The measure should quantify the structure’s robustness with one single value. It should express how and to what extent design objectives are influenced by local failures. The measure could be used for evaluation, optimisation and regulation of the robustness. Furthermore, future generations of standards could be supplemented by a system partial safety factor based on that measure. To achieve these tasks, the measure must be expressive, objective, simple, calculable and generally applicable. A variety of approaches to the quantification of robustness, vulnerability or comparable charac- teristics have so far been published. None of these emerge as distinctly superior or preferable. Some simple formulations of stiffness, damage or energy-based measures of robustness developed by the authors are presented and discussed regarding the suggested requirements and possible applica- tions. This investigation indicates that the requirements are partly in conflict with each other, so it may not be possible to fulfil them all to the same level at the same time. Hence the requirements may have to be limited. This seems plausible at least with regard to the generality requirement. Different types of structures have a tendency for certain significant mechanisms of collapse. Measures that are specific to collapse mechanisms appear to be favourable for a realistic description of structural behaviour. In principle, it is possible to distinguish between measures based on structural behaviour and those based on structural attributes. Measures based on structural behaviour often appear expressive but can hardly be calculated because of the extensive structural analyses that are required. On the other hand, measures that are based on structural attributes are usually easier to calculate while their expressiveness is not (yet) adequate. Despite continually growing numerical calculation capacities, a realistic analysis of structural behaviour as a result of an assumed initial damage will require great effort, even in the future. Under this premise, the development of expressive measures based on structural attributes for the prediction of structural behaviour appears to be desirable for practical applications.

691 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Collapse resistance and robustness of bridges

U. Starossek Hamburg University of Technology, Hamburg, Germany

ABSTRACT: As evidenced by failures, disproportionate or progressive collapse triggered by an initial local failure is an important aspect in bridge design. This question is first addressed in terms of structures in general. Definitions for the terms robustness and collapse resistance are proposed and the shortcomings of current design methods are addressed. A pragmatic approach for designing structures against disproportionate collapse is outlined and a list of design methods is presented. If collapse resistance is to be based on the presumption of local failure, the alternative-paths method or the segmentation method can be pursued to make the structure robust and to limit an incipient collapse to an acceptable extent. On the other hand, if the design is based on reducing the probability of local failure, the specific-local-resistance method and non-structural protective measures can be considered. These methods do not aim at enhancing robustness and thus seem less attractive. Nonetheless, they must be considered at least in cases where other methods are impracticable or overly expensive. In addition to these direct design methods, indirect design methods (prescriptive design rules) are known and in use for buildings but seem inappropriate for bridges. Bridges are primarily horizontally aligned structures. Impact loading produced by falling struc- tural components or debris should therefore be less of a concern for bridges when compared to buildings. It follows that there is less need to provide alternative paths in bridges, and to tie together structural elements, in order to prevent the fall and impact of components. At the same time, it is often difficult to provide alternative paths in structures that have one main axis of extension such as bridges. Various bridge systems are considered in more detail. It is found that continuous girder bridges can be made collapse resistant by the segmentation method—which might require a selective elimination of continuity at segment borders—or by reducing the probability of the local failure of key elements. In the case of cable-stayed bridges, the stay cables are key elements that seem particularly vulnerable. The scenario of a zipper-like collapse triggered by the rupture of one or a few cables is countered by providing alternative paths, i.e. by designing the bridge for the corresponding loss-of-cable load cases. This approach can be complemented by non-structural measures that protect the cables against vehicle impact, malicious action and corrosion. Similar conclusions are drawn for suspension bridges and their hangers. Hanger rupture can be made less probable but, at the same time, it should be designed for such a scenario. A further option, in the case of an earth-anchored suspension bridge, is to pursue the segmentation method and to isolate a collapse initiated by hanger rupture by hinges inserted into the stiffening girder. The primary load-bearing system of large suspension bridges could be made more robust by raising the number of suspension cables which facilitates the formation of alternative paths. A particular concern is the possible exposure of unprotected suspension cable strands to malicious action. For bridges of high significance or exposure, appropriate shielding and security systems should be provided. Arch bridges bear similarities with suspension bridges and much of the respective statements apply. Arches are prone to global stability failure, though, and they are often made of thin-walled cross sections—two particularities that warrant further consideration.

692 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Evaluation of the dynamic amplification factor for cable breakage in cable-stayed bridges

Y. Park, H.-M. Koh & J.F. Choo Department of Civil and Environmental Engineering, Seoul National University, Korea H. Kim Department of Civil Engineering, Mokpo National University, Korea J. Lee Department of Civil and Environmental Engineering, Seoul National University, Korea

Cables being essential structural components in cable-stayed bridges, the effect of the breakage of a cable should be necessarily evaluated during design in order to assess the safety of the structure under such eventuality. Accordingly, major codes are suggesting threshold values of the Dynamic Ampli- fication Factor (DAF) to be used to multiply the static load simulating the dynamic effect induced by the cable breakage. However, the currently proposed thresholds of the DAF are often leading to excessively conservative design that affects the economy and slenderness of the bridge. On the other hand, despite of the importance of the evaluation of the effect of cable breakage during design, a very few studies has been devoted to derive thoroughly appropriate DAF. Therefore, this paper intends to investigate systematically the dynamic response provoked by the loss of cable in a cable-stayed bridge with focus on the dynamic amplification through case studies on an actual bridge. Time history analyses are performed for various cable loss scenarios considering different modal damping ratios of the structural system. In addition, two patterns are considered for the loss of cable under dead and live load : instantaneous loss of cable and progressive loss of cable. Progressive loss of cable which corresponds to more realistic cable rupture pattern is simulated by decreasing the tension to zero in a short period. Dynamic responses are compared to static responses to quantify a dynamic amplification factor for pylons, cables and points on the deck. It is expected that further applications of the proposed evaluation procedure to a larger number of bridge will provide appropriate values of the DAF for cable breakage meeting the demands for both safety and economy.

Keywords: Cable-stayed bridge, Loss of cable, Dynamic amplification factor, Progressive rupture

693

Sustainable bridges

Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Test of a concrete bridge in Sweden. – I. Assessment methods

A. Puurula, O. Enochsson, H. Thun, B. Täljsten & L. Elfgren Luleå University of Technology, Luleå, Sweden J. Olofsson Skanska Sweden, Göteborg, Sweden B. Paulsson Banverket, Borlänge, Sweden

ABSTRACT: Field tests have been carried out on an existing reinforced concrete railway bridge in the Research Project “Sustainable Bridges”, see www.sustainablebridges.net. The project was supported by the European Union 6th Framework Program during 2003–2007. Procedures have been tested for inspection and condition assessment, load carrying capacity, monitoring, and strengthening, SB-Guide (2007). This paper, which is the first in a series of four, presents assessment methods. Swedish and European Code methods were compared to finite element models and the Modified Compression Field Theory, MCFT. A failure in combined shear, bending and torsion was reached for an applied mid span load of 11.7 MN. This was close to what was predicted by the methods in the project and 20 to 50% higher than other predictions based on common codes and models.

Figure 1. Photo looking north of the railway bridge in Örnsköldsvik during testing. A mid span load is applied with a steel beam anchored to the bedrock some 10 m below the bridge foundation slabs. A temporary by-pass for the traffic in the right lane can be seen to the right of the bridge.

697 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Test of a concrete bridge in Sweden. – II. CFRP strengthening and structural health monitoring

B. Täljsten Luleå university of technology, Luleå, Sweden and Denmark Technical University, Lyngby, Denmark M. Bergström, H. Nordin, O. Enochsson & L. Elfgren Luleå university of technology, Luleå, Sweden

This paper presents a full-scale test on a strengthened railway concrete trough bridge. A unique opportunity came up. The existing railway line was going to be replaced with a new one and the bridge became obsolete. The purpose of the project was to investigate the shear bearing capacity of the bridge. To avoid bending failure, the soffit beams were strengthened with Near Surface Mounted Reinforcement (NSMR) consisting of Carbon Fibre Reinforced Polymers (CFRP). The project was a part of the European funded research project Sustainable Bridges (www.sustainablebridges.net). The bridge was tested to failure to demonstrate and test new and refined methods developed in the project regarding procedures for condition assessment and inspection, load carrying capacity, measurement and strengthening, SB-Guide (2007). The bridge was built in 1955. It has two spans of 12 m. In one of the spans a loading beam made of steel was placed in the centre of the span. The loading beam was then pulled down with cables injected to the bedrock beneath the bridge. The bridge was heavily monitored with many different types of sensors, for example; electrical strain sensors on concrete, steel and CFRP rods, LVDT (Linear Variable Displacement Transform- ers) for measuring the displacements at various locations and curvature, laser deflection meters for measuring the mid-displacement, accelerometers, fibre optic crack sensors and fibre optic strain (Bragg) sensors. For the performance of the strengthening system four quantities were analyzed; load, strain, curvature and stiffness. The strain distribution was established by applying strain gauges both on the compressed concrete, tensile steel reinforcement and the CFRP rods. The chosen strengthening method was Near Surface Mounted Reinforcement (NSMR) rectan- gular bars of Carbon Fibre Reinforced Polymers (CFRP) which were mounted by bonding cut out groves in the slab, the size of the grooves where 15 × 15 mm. This configuration does not interfere with the existing steel reinforcement. Research at LTU has shown that this method is superior compared to externally bonded FRP plates and the method has also been used in several field applications in Sweden, both in buildings and bridges. In this particular case rectangular bars with a cross section of 10 × 10 mm were used. The rods chosen where provided by Sto Scandinavia AB with the brand name Sto FRP Bar M10C. The modulus of elasticity for the rods was 250 GPa with a strain at failure of 11 ‰. The adhesive used for bonding was Sto BPE Lim 567 (A + B) a cold cured two component tixotropic epoxy adhesive with a modulus of adhesive of 6.5 GPa and a bond strength of approximately 20–22 MPa. The strengthening of the bridge was very successful and a stress of approximately 1950 MPa was calculated from strain readings in the CFRP rods. Furthermore, very high shear stresses, approximately 10 MPa were transferred from the CFRP rods to the concrete in the bonded slots. At failure a very distinct fish bone pattern had developed in the concrete and the end location of the rods. It was found from the test that it is very difficult to predict the ultimate behaviour of the bridge even though it was mapped in detail before the moni- toring and testing was carried out. The study stresses the importance of using SHM for evaluation of existing design models and the behaviour of real structures and it will be very interesting and challenging to evaluate the result from the bridge further.

698 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Test of a concrete bridge in Sweden. – III. Ultimate capacity

O. Enochsson, A. Puurula, H. Thun, L. Elfgren & B. Täljsten Luleå University of Technology, Luleå, Sweden J. Olofsson Skanska Sweden, Göteborg, Sweden B. Paulsson Banverket, Borlänge, Sweden

ABSTRACT: A reinforced concrete railway trough bridge has been loaded to failure in the Euro- pean Research Project “Sustainable Bridges”. Procedures have been tested for inspection and condition assessment, load carrying capacity, monitoring, and strengthening. This paper, which is the third in a series of four, discusses the ultimate capacity of the bridge. A failure in combined shear, bending and torsion was reached for an applied mid span load of 11.7 MN. This means that the bridge could carry eight trains with an axle load of 250 kN instead of the one it was designed for.

Figure 1. View of the bridge in Örnsköldsvik during testing.

Figure 2. Ultimate failure caused by stirrups rupture after yielding in longitudinal reinforcement.

699 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Test to failure of a railway reinforced concrete through bridge in Örnsköldsvik, Sweden. – IV.Evaluation of damage detection methods

P.J.S. Cruz & R. Salgado ISISE, University of Minho, Department of Civil Engineering, Guimarães, Portugal

ABSTRACT: The important advances in vibration monitoring in the last years have encouraged the scientific community to develop methods based on vibration monitoring for detecting damage. In fact, the current structural evaluation of bridges needs more sophisticated methods in order to detect damage in the earliest steps. These damage detection methods have to handle several problems in order to be applied to bridges successfully. For instance, the dynamic response and modal parameters of the bridge after damage have small changes if it is compared to the dynamic parameters of the previous structural condition. Furthermore, these methods are very sensitive to noise and temperature effects. These factors may hide these small changes in the vibration parameters. Recently, a new vibration based damage detection methods have appeared. These methods try to magnify the differences in vibration parameters caused by damage. For the evaluation of the vibration based damage detection methods some bridges have been tested in the last years. However, no final conclusions about damage detection methods have been obtained so far. This was the result of the accuracy of these methods which change with the ambient conditions, the bridge geometry, the material and the quality of the gathered information. Therefore, until a certain level of maturity in these damage detection methods is achieved, it will be necessary to carry out several deliberated damage tests on bridges. In this study, the modal and damage identification of the Övik Bridge was carried out through different deliberated damaged phases. The structural behaviour of this bridge was evaluated in two damage tests and its dynamic response was acquired after each one. Afterwards, the most promising damage detection methods (the Curvature, the CWT, the CWA and the WPS methods) were compared and evaluated through the dynamic simulation acquired after the damage tests. Comparison of the modal parameters between the two damage tests indicated close similitude among the first three mode shapes. A special mention has to be done to the damping ratio for the first mode shape. Its value increased more than two times from the first to the final test. Damage present in the bridge might increase the damping in the lateral movement related to this mode. To exemplify the implementation of the damage detection methods, they were applied to the first two mode shapes and acceleration response along the sidewalks of the bridge. Damage was successfully identified and localized by all the involved methods in the two analyzed damage scenarios. The Curvature, DWA and CWT methods could detect damage close to sections where severe damage was localized. Nevertheless, other peaks were also localized near the supports, outside the damage zone. With the WPS method, a better damage identification was obtained. In summary, the damage detection methods used in this investigation allowed to successfully identify the damage. A close grid of sensors near the probable damage zone is required in order to acquire the local disturbance in mode shapes.

700 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Assessment and monitoring of an old railway steel truss bridge in northern Sweden

Ola Enochsson & Lennart Elfgren Department of Civil, Mining and Environmental Engineering, Luleå University of Technology, Luleå, Sweden Anders Kronborg & Björn Paulsson Banverket, Luleå & Borlänge, Sweden

ABSTRACT: Measurements have been carried out in order to improve an assessment of a riveted steel bridge from 1911, see Figure 1. Dynamic Amplification Factors, DAF, were estimated for different parts of the bridge, see Figure 2, by measurements of deflections and strains for a train passing over the bridge with different speeds. The measured dynamic amplification factors were considerably lower than the corresponding ones calculated according to available codes. Due to this it was possible to increase the axle load on the bridge from 22.5 to 25 tons. Additionally, a finite ele- ment model was established and updated with help of the results from the monitoring, see Figure 3. The influence of several parameters on the model’s accuracy are investigated and discussed.

Figure 1. The Keräsjokk Bridge on the Haparanda Line in northern Sweden seen from the north.

Figure 2. Measuring points at the Keräsjokk bridge. Global horizontal and vertical displacements were measured in point A and B (truss) with optical laser displacement sensors and local vertical displacements in point C and D (stringer and floor beam) with LVDTs i.e. relatively to light-weight trusses supported in its ends to the connection beams. Finally, local flexural and normal strains were measured in points C and D in the midsection of the stringer and the floor beam, respectively (only in August).

Figure 3. Three dimensional beam model, shown with deformed solids.

701 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Railway bridge loaded to failure test in Örnsköldsvik, Sweden – Strain measurement using Fiber Bragg Grating system incorporated in Carbon Fibre Reinforced Polymer

A. Kerrouche, J. Leighton & W.J.O. Boyle School of Engineering and Mathematical Sciences, City University, London, UK

1 INTRODUCTION

This paper describes a strain measurement experiment carried out on a concrete railway bridge in northern Sweden. The experiment formed part of the EU ‘Sustainable Bridges’ project, which seeks to improve the capacity of Europe’s railway bridges to take high-speed and heavy axel loads on trains. For this experiment, Fiber Bragg Grating (FBG) sensors were fixed and protected on the grooves positioned on the underside of the bridge. The sensors were linked to a compact multiple-channel system based on WDM architecture using a scanning Fabry-Perot filter. The bridge (already scheduled for destruction) was then loaded until failure. This paper describes the measurement system and describes and discussed the test procedure and results obtained. Health and safety monitoring of civil structures such as rail bridges presents a real engineering challenge because of increasing capacity forces due to longer faster trains. Fiber Bragg Grating based mentoring system has advantages optic sensing schemes represent an enabling technology which offers a number of advantages for real time structural health monitoring of engineering materials and structures. Key advantages of their use include the potential for a large number of sensors to be multiplexed along a single length of optical fiber, enabling single or multi- point measurements to be made. Physical measurands of general interest for civil structures include strain, temperature, vibration and acoustic emission as well as chemical measurements including pH, moisture ingress, oxygen, chlorides and the presence of a variety of molecules carried by the moisture into the structure itself. The data obtained from such monitoring may be used to validate engineering designs, optimize manufacturing processes and thus facilitate structural health determination.

2 THE MONITORING SYSTEM

The compact 8-channel system used in this work is based on multiplexed WDM (Wavelength Division Multiplexing) architecture which used a scanning tunable filter to interrogate the 10 protected FBG sensors fixed into the concrete bridge. The basic sensing principle of the system is measuring wavelength shifts of the spectral output of the system, which is a convolution of the spectral peak of the interferometer in transmission and the Gaussian reflection profile from the Bragg grating sensor elements in response to changes in a measurand of interest.

3 SENSORS PLACEMENT INTO THE BRIDGE

The Övick Bridge used in this test was situated in Örnsköldsvik, Sweden. Built of reinforced concrete with two spans, it was built in 1955 and was taken out of service due to relocation of a new high-speed railway, the Botnia Line. 702 A groove, ∼2 mm wide and ∼1 mm deep, was cut into the centre of a square cross section of 10 mm of carbon fiber rod in order to accommodate the optic fiber. The fibres were glued with low viscosity Cyano-Acrylate and then covered with epoxy. Carbon Fibre Reinforced Polymer (CFRP) rods were used like strength tendons placed into appropriate channels already cut into the underside of the bridge to bond the carbon fiber rods with FBG sensors as well as other strain gauge sensors.

4 LOAD TEST

The loading arrangement was completed within two weeks before the final test. A beam was placed in the centre of the span to transfer the load, attached by two cables which were anchored into the ground (about 10m below the bridge foundation slabs). Load on the span was increasing in stages, with time for relaxation and time for crack surveillance at major load intervals. According to preliminary estimation, the critical point load might reach up to 6 M Newton before failure. However, loading increased up to 12 MNewton when the failure occurred after more than one hour testing.

5 CONCLUSIONS

This paper demonstrates a new application test on a railway bridge loaded to failure in order to test its remaining ultimate load carrying capacity after a service period of 50 years. The system is proved to support measurement of high strain level of more than 5000 us. Also, this experience shows a very interesting opportunity to test how FBG systems deal with structures outside their normal working parameters, hopefully contributing to work on failure prediction in harsh engineering fields such as earthquake measurement, aerospace or maritime engineering.

703 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Single and multiple crack monitoring in concrete bridges

P.J.S. Cruz & A. Diaz de León ISISE, University of Minho, Department of Civil Engineering, Guimarães, Portugal C.K.Y. Leung Hong Kong University of Science & Technology, Clear Water Bay, Honk Kong, China

ABSTRACT: This work reports the applications of a distributed optical fibre crack sensor which does not require prior knowledge of the crack locations and is able to detect, locate and monitor multiple cracks with a single fibre. The paper gives an insight to the most important aspects of the sensor fabrication and presents guidelines for applying it to monitor flexural cracks with a low resolution OTDR. The field imple- mentation of the sensor in the Övik Bridge, in the north of Sweden, is described with details. The results obtained clearly demonstrate the sensor’s applicability to detecting and locating the formation of cracks in various locations of a structural member.

1 INTRODUCTION

Due to material inhomogeneities, the exact location of cracks in a concrete structure cannot be predicted. Conventional “point sensors” can easily miss the cracks. On the other hand, integrated sensors, which measure displacement between two points separated by a relatively large distance, are not able to distinguish between the harmless condition of many fine cracks and undesirable situation of one widely opened crack. A crack sensing concept which can detect and locate cracks without prior knowledge of crack locations has been studied in (Leung et al. 2000) and (Diaz de León et al. 2004). The operation principle of the sensor is based on the intensity variation of light power within an optical fibre that is inclined to the crack, when fiber microbending is induced by crack opening. With Optical Time Domain Reflectometry (OTDR), the location of the crack can be determined from the backscattered signal vs time (or distance) plot. If a calibration relationship is available the crack opening can be obtained from the magnitude of the power loss (Diaz de León 2007). The methodology in detecting and localizing the formation of flexural cracks in various locations and sensor’s capability in measuring a range of crack widths was demonstrated through testing of instrumented RC beams subjected to sustained to sustained and repeated loading (Diaz de León et al. 2006). The primary objectives of this paper are: 1) To examine the sensor’s applicability to detecting and locating the formation of cracks in a structural member; 2) To demonstrate the implementation of the sensor in monitoring flexural cracks on a real bridge. The results presented demonstrated the viability the sensor, to detect the formation and propagation of external cracks and measure its crack opening in practical applications.

704 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Structural assessment of concrete railway bridges

M. Plos, K. Gylltoft & K. Lundgren Chalmers University of Technology, Göteborg, Sweden L. Elfgren Luleå University of Technology, Luleå, Sweden J. Cervenkaˇ Cervenka Consulting, Praha; Czech republic A. Herwig & E. Brühwiler École Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland S. Thelandersson Lund University of Technology, Lund, Sweden E. Rosell Swedish Road Administration, Borlänge, Sweden

ABSTRACT: For a sustainable development of Europe, there is a need to at least double the railway transports in the coming 20 years. In order to reach this goal, the residual service lives of existing concrete railway bridges need to be extended, at the same time as they are subjected to higher axle loads, higher railway speeds and heavier traffic intensity. Today, many concrete bridges are replaced or strengthened because their reliability cannot be guaranteed based on the structural assessments made. The objective of the work presented here was to provide enhanced assessment methods that are able to prove higher load-carrying capacities and longer service lives for existing concrete railway bridges. The work accomplished was a part of the EU-project Sustainable Bridges. The results are implemented in the Guideline for Load and Resistance Assessment of existing European Railway Bridges, see SB-LRA (2007) and are reported in detail in a background document, see SB4.5 (2007). Improved methods for the determination of in-situ material properties in existing concrete bridges are presented. The methods cover the material properties needed for linear as well as non-linear analysis, for deterministic as well as for fully probabilistic assessments. Methods for structural analysis on different levels are presented. Recommendations were devel- oped for redistribution of sectional moments and forces from linear Finite Element (FE) analysis of slab bridges. It was shown that lateral redistribution of the linear moments is necessary to avoid under-estimation of the capacity when assessing existing bridges. Advanced methods for local resistance analysis are presented, e.g. regarding combined shear, torsion and bending interaction. One main objective was to facilitate the use of non-linear analysis for structural assessment. Non- linear analysis provides the greatest potential to discover any additional sources for load-carrying capacity, and gives a better understanding of the structural response, forming an improved basis for assessment decisions. Recommendations are given regarding methods and models to be used in non-linear FE analysis, as well as regarding when such methods are likely to reveal a higher capacity. Another main objective was to provide methods for assessing the remaining structural resistance of deteriorated concrete bridges. Recommendations are given for maintenance of bridges with reinforcement corrosion, and on the effect of the corrosion on the anchorage capacity. Furthermore,

705 a methodology is presented for improved assessment of the fatigue safety for existing concrete bridges. Here, the emphasis is on evaluation of the remaining fatigue life of short-span bridges and secondary elements. The proposed methodology includes (1) a study of the bridge with an evaluation of the reinforcement detailing, (2) an inspection of the bridge and its past performance and (3) a fatigue safety check. It is concluded that for bridges with concrete in good condition, the fatigue safety is in general determined by the steel reinforcement.

706 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Safety requirements in the capacity assessment of existing bridges

J.R. Casas Universitat Politècnica de Catalunya, Barcelona, Spain D.F. Wisniewski COWI A/S, Kongens Lyngby, Denmark

ABSTRACT: The paper deals with a summary description of the main safety criteria and require- ments adopted in the “Guideline for Load and ResistanceAssessment of Existing European Railway Bridges – advices on the use of advanced methods” (SB-LRA, 2007) developed within the European project Sustainable Bridges – Assessment for Future Traffic Demands and Longer Lives. Despite being developed for the case of railway bridges, the criteria and requirements are general for their use in other bridge types. A more complete and detailed description of the general basis and criteria can be found in the background document.

1 INTRODUCTION

One of the main issues related to the assessment of existing bridges is to decide on the safety formats that shall be used in the capacity evaluation and on the philosophy behind these formats. The safety issues to consider are not only those related to structural aspects, but also dealing with durability and service life design. The paper explains the criteria, boundary conditions and requirements adopted in the definition of the safety format as appear in the Guideline SB-LRA (2007) and defines the basis and criteria that can be used to set the required safety level when assessing existing railway bridges. The safety criteria are based on the adoption of a limit states format with different levels of complexity (partial safety factor method or full probabilistic analysis) depending on the importance of the structure and the results of previous assessments. The safety formats are divided according whether a member (element) or system (whole bridge) assessment is carried out. In the case of the formats for the system level assessment, the paper propose several assessment methods that allow to take into account the structural redundancy (longitudinal and/or transversal) characteristic for a bridge under consideration. Furthermore, emphasis is put on the reliability-based assessment using non-linear analysis. The available methods are presented and applied in a practical example. Due to the high complexity of the complete probabilistic non-linear analysis, two simplified methods are also presented and applied to same example to show their simplicity and accuracy. The safety level is worked out in the form of a target value of the reliability index for the Ultimate, Serviceability, Durability and Fatigue Limit States. Because bridge assessment is highly case-specific, the Guideline does not just propose a value to be adopted for the safety level, but gives information and guide how to fix this level for each case under study. The proposed target reliability levels proposed in different countries and by different international bodies (Eurocode, ISO, JCSS, etc.) are presented, jointly with the most significant assumptions. In this way, the engineer responsible for the assessment can choose the most suitable safety level for each specific case under consideration. The safety levels relevant for the assessment of a member and a whole system are presented.

707 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Guideline for load and resistance assessment of existing European railway bridges

Jens S. Jensen COWI A/S, Kongens Lyngby, Denmark Mario Plos Chalmers University of Technology, Göteborg, Sweden Joan R. Casas Universitat Politècnica de Catalunya, Barcelona, Spain Christian Cremona Laboratoire Central des Ponts et Chaussées, Paris, France Raid Karoumi Royal Institute of Technology, Stockholm, Sweden Clive Melbourne University of Salford, Salford, UK

Many of the European railway bridges are getting close to the end of their service life. At the same time the railway operators demand higher axle loads for freight trains and higher speeds for pas- senger trains. This requires new and better approaches for assessing both the railway loads and the resistance of railway bridges. The main objective of the “Guideline for Load and ResistanceAssess- ment of Existing European Railway Bridges – advices on the use of advanced methods”, developed within EU founded project “Sustainable Bridges – Assessment for Future Traffic Demands and Longer Lives”, is to provide bridge evaluators with the most advanced knowledge regarding meth- ods, models and tools that can be used in the assessment of existing railway bridges in order to get a realistic evaluation of their load carrying capacity and also more accurate evaluation of their remain- ing service life. This includes systematized step-level assessment methodology, advanced safety formats (e.g. probabilistic or simplified probabilistic) refined structural analysis (e.g. non-linear or plastic, dynamic considering train-bridge interaction), better models of loads and resistance param- eters (e.g. probabilistic and/or based on the results of measurements) and methods for incorporation of the results form monitoring and on-site testing (e.g. Bayesian updating). The SB-LRA Guideline define the “state-of-the-art” practice for assessing the load and resistance of existing railway bridges taking into account the measures of actual bridge condition, identi- fied according to “Guideline for the Inspection and Condition Assessment of Railway Bridges” (SB-ICA, 2007), and the results of monitoring, performed according to “Monitoring Guidelines for Railway Bridges” (SB-MON, 2007). The assessment performed using SB-LRA Guideline may give basis for the decision regarding repair or strengthening of a bridge which can be carried out according to the recommendations presented in “Repair and Strengthening of Railway Bridges – Guideline” (SB-STR, 2007). All three above mentioned Guidelines have also been developed within Sustainable Bridges Project. As presented in this paper, the SB-LRA Guideline provides a lot of data, models and tools specific for the assessment of existing railway bridges than can not be find in any of existing design and assessment codes. It also gives guidance on the comprehensive methods of the analysis and assessment that are normally not used in the every day design or assessment practice and therefore

708 are not very well known by the bridge engineering community. All these methods, models and tools might help in saving many existing railway bridges form unnecessary repair, strengthening or replacement. This paper gives a general overview of the whole Guideline. Nevertheless, the major focus is placed on the innovative elements proposed in the Guideline, which have been developed due to several research activities performed during the four years of the project duration. This includes general recommendations regarding: assessment procedures, safety formats and requirements, bridge loads and the dynamic effects; but also specific recommendations related to concrete, metal and masonry arch bridges.

709 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Improved assessment methods for static and fatigue resistance of old metal railway bridges

C. Cremona & A. Patron Laboratoire Central des Ponts et Chaussées, Paris, France S. Hoehler & B. Eichler RWTH Institute of Steel Construction, Aachen, Germany B. Johansson & T. Larsson Lulea University of Technology, Lulea, Sweden

A very important part of the bridges in the European railway networks are metallic bridges, and have been building during the last 75 years (some of them are much older). The increasing volume of traffic and axle weight of trains mean that for many structures the loads today are much higher than those envisaged when they were designed. This paper presents the recommendation and the advanced methods established to develop improved assessment methods for existing metallic bridges in the context of the Sustainable Bridges Project. The assessment of old metal bridge performance must include an overall conventional safety evaluation for all the joints and all the structural components versus the actual operating conditions. Fracture critical members represent the most sensitive parts in old metal bridges. The evaluation has the purpose to identify the risks to predict in terms of stability, strength and fatigue, and to localize the hot spots for which failure due to damages and undetected cracks could lead to bridge collapse. Fatigue phenomenon has puzzled researchers for over 200 years. The fatigue endurance is one of the major influencing factors concerning the service life for old metal bridges. Clamping force, corrosion, hole preparation and material properties largely influence fatigue performance. The analysis of many of the tests analyzed in the Sustainable Bridges project led to endurances lower than 2 predicted by detail category σc = 71 N/mm . Investigations concerning the constant amplitude limit and the cut off limit show that the level for no fatigue accumulation (cut off limit) can be raised from 28.7 MPa to 40 MPa. A constant stress range below 52.3 MPa does not provide cracking in components according to the evaluated tests. This is only valid providing that there is no severe corrosion or damage present on the structural components. For initial assessment, fatigue life is evaluated by using Miner cumulative damage law in conjunction with Wöhler curves. For intermediate and enhanced assessment, enhanced methods beyond the conventional design and assessment procedures as required in Eurocode 3 for the resistance assessment of old steel railway bridges must be introduced. Fracture mechanics models are applied for a more detailed assessment. Modern standards for design of steel structures like Eurocode 3 cover riveted structures but they do not give complete information. Old design standards on the other hand are quite incomplete concerning instability phenomena and they are covering elastic design only.This approach is appro- priate for initial assessment analysis at ULS and for assessing fatigue cycles. But for intermediate and enhanced assessment, if the resistance in ULS is insufficient it is proposed to allow plastic deformations. For advanced assessment it is recommended to perform a non linear analysis with FEM. In order to cover all possible failure modes it is needed to model the structure with shell elements and to apply local and global initial imperfections.

710 It is recommended to limit any testing as far as possible and to material parameters that may influence the assessment result in a great manner. Tests should only be performed to gain the most important values, such as chemical analysis, mechanical properties: yield strength and ultimate tensile strength and fracture toughness. For intermediate analysis, load field measurements can be performed. The provided data can be used for refining the structural analysis and the fatigue assessment as far as they are representative of the operating conditions over the bridge life (which is rarely the case, but it is the best estimation of load spectra of the actual traffic).

711 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Consideration of dynamic traffic action effects on existing bridges at ultimate limit state

E. Brühwiler & A. Herwig École Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland

1 INTRODUCTION AND BASIC APPROACH

The objective of the examination of existing bridges is to show that for actual traffic load action effects the requirements are fulfilled regarding the Ultimate Limit State, the Fatigue Limit State and the Serviceability Limit State. The basic approach of updating traffic action effects consists in a separate consideration of (static) loads Q and q (axle and line loads respectively), and forces due to dynamic traffic effects. Forces occurring in the bridge structure due to dynamic traffic action are often expressed by a dynamic amplification factor ϕi (amplifying the static load effect). This dynamic amplifica- tion factor depends on the limit state considered, e.g. ULS, SLS or FLS, and the corresponding characteristic structural behaviour. The dynamic behaviour of the bridge structure under traffic loads consists in absorption, storage, dissipation and release of energy that is stored in the structure due to dynamic traffic action. For elastic bridge behaviour, this energy stored in the bridge element consists in vibrations leading to increase of deflections and internal forces. Elastic bridge behaviour is considered for fatigue and service limit state. However, for the ultimate limit state, elastic-plastic structural behaviour must be accounted for, and formulas like those commonly given in design codes are then fundamentally wrong since they refer to elastic structural behaviour only. This paper presents a study of the dynamic action effects on the structural behaviour of “ductile” structural bridge elements showing significant deformations in the post-elastic regime. The results allow deriving dynamic amplifications factors valid for the structural safety verification at Ultimate Limit State (ULS).

2 DISSIPATION OF DYNAMIC EFFECTS AT ULS

At ultimate limit state (ULS), structural elements in reinforced and prestressed concrete and in steel provide significant plastic deformation due to yielding of the steel. In statically undetermined systems, the plastic deformation capacity of the structural elements is usually not fully consumed by internal redistribution of cross sectional forces. In this case, energy induced by dynamic action effects may also be dissipated by the structural element. However, the so-called “gravity effect” needs to be considered: Both the traffic loads and per- manent loads act in the same direction due to gravity, both leading to (external) work (energy) stored in the structural system (Fig. 1). This means that a considerable part of the total dissipation capacity of the structure is “consumed” by the static load effects. Only one part, i.e. roughly the non-linear domain, is available for dissipation of energy due to dynamic effects. By means of energy balance consideration and dynamic simulation using a simple analytical model, the present paper shows how the kinetic energy due to dynamic action effects is dissipated by the bridge structure shown in Fig. 1. Dynamic action effects consume a minor part of the whole energy dissipation capacity of the cross section at mid-span (Fig. 2). 712 Figure 1. “Gravity effect” and dissipation of energy in the structural response.

Figure 2. Work diagram of the plastic hinge with the result of the dynamic simulation.

3 CONCLUSIONS

In the context of updating of traffic action effects on existing bridges, dynamic amplification effects are investigated and dynamic amplification factors are derived for the deterministic verification of structural safety at Ultimate Limit State of bridges. In the case of significant plastic deformation of structural elements sufficient dissipation capacity is available. This means that the dynamic amplification factor may be set to 1.0. The present rational approach is simple and reasonably conservative. It most likely provides an important finding to demonstrate – in an efficient manner – that most existing bridges fulfil the requirements of structural safety when future traffic loads are increased.

713 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

High cycle fatigue strength of brick masonry. A probabilistic approach

J.R. Casas Universitat Politècnica de Catalunya, Barcelona, Spain

ABSTRACT: The paper presents the development of a model for the fatigue resistance of brick masonry under compression. The model takes into account the random nature of the fatigue strength phenomena and, as a consequence, proposes different model parameters as a function of the desired probability level. The model has been calibrated with the still limited experimental data on brick masonry under high cycle compression loading. Very good correlation is obtained. It is shown how the fatigue resistance of brick masonry is not only dependent on the magnitude of the stress cycles, but also on the magnitude of the minimum stress level.

1 CONCLUSIONS

The paper presents a probability-based fatigue model valid only for masonry under compression. The model allows the definition of the fatigue resistance of masonry with different confidence levels. For example, the proposed fatigue equation (S-N relation) for a survival probability of 95% for masonry under compression in any condition (dry, wet or submerged) has been obtained as:

S is the ratio of the maximum loading stress to the static strength and R is the ratio of the minimum stress to the maximum stress smin/smax. An endurance limit S = 0.5 was also obtained It was found that theWeibulldistribution fits with a good accuracy the experimental data available used to calibrate the model. For different ranges of the stress level S, the parameters α and u of the Weibull distribution were obtained as presented in table 1. In order to fully validate the present model for compression and develop similar models for masonry under shear (critical in many cases of multi-ring masonry arches) more experimental data is needed. It would be of interest to carry out a large number of triplet tests under shear and compression. In this way, SN curves for each failure mode could be obtained. Based on the fatigue model derived, a method is proposed for the calculation of the reliability index or probability of failure in a given reference time and/or the calculation of the remaining service life with a predefined probability level for existing masonry arch bridges.

Table 1. Parameters α and u of Weibull distribution for different values of stress level S. S α alnu u r2 r

0.9 0.8511 5.336 528 0.96 0.98 0.8 0.5353 4.3585 3436 0.98 0.99 0.75 1.0753 7.7298 1324 0.97 0.98 0.70 0.4604 4.8822 40306 0.88 0.94 0.65 0.2379 2.521 40010 0.50 0.71 0.60 0.4202 5.3679 353144 0.97 0.98 0.55 0.8785 9.0308 29138 0.88 0.94

714 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Probabilistic models for resistance of European concrete railway bridges

J.R. Casas Universitat Politècnica de Catalunya, Barcelona, Spain D.F. Wisniewski COWI A/S, Kongens Lyngby, Denmark

ABSTRACT: The paper presents the definition of a set of probabilistic models for the resistance of concrete bridge cross-sections that may be used in the assessment of the existing concrete railway bridges at a component/member level. The models are based on the calculation of the mean value and the Coefficient Of Variation (COV) of the sectional response (ultimate bending, ultimate shear). To obtain these values, the assessing engineer only has to calculate the nominal value of the section capacity, using traditional methods. The main results obtained are: Response to bending: For postensioned rectangular sections, the bias ratio of the ultimate bending moment is a function of the ratio of the area of prestressing steel to concrete (Ap/Ac). The COV is only dependant on the ratio Ap/Ac. For values lower than 0.4%, the influence of the COV of reinforcing steel and concrete on the COV of the ultimate bending resistance is negligible, and the values of 3% for COV of prestressing steel of 2%, and 5% for COV of prestressing steel of 5% can be used for the COV of ultimate bending resistance. For values between 0.4 and 0.7%, the COV of ultimate bending moment is only dependent on the COV of the concrete strength. For postensioned massive slabs, the bias ratio can be taken as equal to 1.11. For ratioAs/Ap < 1.2 (As = area of reinforcing steel, Ap = area of prestressing steel), the COV of the ultimate bending resistance can be taken as 5%, independent of variability of concrete, reinforcing and prestressing steel. For ratio As/Ap > 1.2, the COV of the ultimate bending resistance is only dependent on the COV of the reinforcing steel strength. For reinforced T-shape and rectangular elements, the bias ratio is equal to 1.15 and the COV of the ultimate bending resistance is equal to the COV of the yielding strength of the reinforcement. Due to the lack of influence of concrete properties in the results obtained, the conclusions drawn for rectangular and T-shape sections can be also extended to the case of reinforced concrete slabs. Response to shear: In all cases analyzed, the failure in shear was due to the rupture of the rein- forcing steel and never because of the excessive compression on the concrete diagonal strut. Also, when both longitudinal and transversal reinforcing is disposed the failure of the shear reinforce- ment occurs first. Therefore, the conclusions presented can not be applied to cross-sections failing in shear due to compression of the concrete or to the cases where the longitudinal reinforcement fails due to shear before the rupture of the reinforcing stirrups. For reinforced rectangular sections with shear reinforcement, the bias ratio is highly dependant on the assumptions made for the calculation of the nominal value of the shear response and the amount of longitudinal and transversal reinforcement. For reinforced rectangular sections without shear reinforcement, the mean value of the shear _ resistance is calculated as Vult = 1.71 kbd where b is the section width and d is the effective cross- section depth. k depends on the concrete strength. The coefficient of variation of the shear resistance is similar to the coefficient of variation of the concrete tensile strength (around 10–20%) For prestressed rectangular sections with shear reinforcement, the bias ratio is equal to 1.05 when both the contribution of concrete and steel are considered in the resistance of tension forces due to shear. The COV of the shear resistance is 7%. 715 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Probabilistic models of material properties for design and assessment of concrete bridges

D.F. Wisniewski COWI A/S, Kongens Lyngby, Denmark (formerly at the University of Minho, Guimaraes, Portugal) P.J.S. Cruz University of Minho, Guimaraes, Portugal A.A.R. Henriques University of Porto, Porto, Portugal R.A.D. Simões University of Coimbra, Coimbra, Portugal

ABSTRACT: In this paper the probabilistic models of basic mechanical properties of precast and cast-on-site concretes are analysed. The probabilistic models of mechanical properties of rein- forcing and prestressing steels are also presented and discussed. The extended review of existing models, available in the technical literature is made and some original probabilistic models devel- oped based on the huge amount of data collected by the authors are presented. The original models are proposed for concrete ultimate strength (separately for precast and cast-in-place concretes), for proportionality limit and ultimate strength of prestressing steel and for yield and ultimate strength of reinforcing steel.

1 INTRODUCTION

The theoretical models describing structural behaviour of reinforced or prestressed concrete struc- tures requires basic information about the structure geometry (dimensions of the cross-section, position of the reinforcement, eccentricities, etc.) and about mechanical properties of the materials (compression strength of concrete, yielding strength of reinforcing steel, proportionality limits of prestressing steel, etc.). The structure geometry as well as mechanical properties of materials composing the structure have a random nature and they should be treated as random variables. Con- sequently, in order to describe accurately the structural behaviour of the reinforced or prestressed concrete structure the complete probabilistic models (probability distribution function and basic statistics) of those variables are indispensable. This paper presents the probabilistic models of basic mechanical properties of concretes (pre- cast and cast-on-site) and steels (reinforcing and prestressing). Quite extensive review of existing models, available in the technical literature is made. However, the major focus is placed on the original probabilistic models developed based on the huge amount of data collected by the authors. The probabilistic models of structure and structural member geometry due to limited space are not presented in this paper.

2 CONCLUSIONS

Extensive review of the existing probabilistic models of the reinforcing and prestressing steels properties, and plant-cast and site-cast concrete and its comparison with the models developed 716 by the authors, allow to define reliable models that can be used in the probability based design or assessment of concrete bridges. The performed analysis shows that the variability of strength properties of steel and concrete (described by standard deviation or by COV) reduce significantly in last decades and that some probabilistic models available in literature for this properties may not be appropriate for modern structures and construction materials.

717 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Safety assessment of railway bridges by non-linear and probabilistic methods

J. Cervenka,ˇ V. Cervenka,ˇ Z. Janda & R. Pukl Cervenka Consulting, Prague, Czech Republic

ABSTRACT: The paper discusses safety formats suitable for assessment of railway bridges using non-linear analysis. Engineers often use non-linear analysis while making assessment of old existing structures or when designing new ones. This evolution is supported by rapid increase of computational power as well as by new capabilities of the available software tools for numerical simulation of structural performance. The code provisions on the other hand provide very little guidance how to use the results of a non-linear analysis for structural assessment or design. The safety formats and rules that are usually employed in the codes are tailored for classical assessment procedures based on beam models, hand calculation or linear analysis and local section checks. On the other hand, non-linear analysis is by its nature always a global type of assessment, in which all-structural parts, or sections, interact. Until recently the codes did not allow applying the method of partial safety factors for non-linear analysis, and therefore, a new safety format was expected to be formulated. Certain national or international codes have already introduced new safety formats based on overall/global safety factors to address this issue. Such codes are, for instance, German standard DIN 1045-1 (1998) or Eurocode 2 EN 1992–2, (2005). This paper compares several possible safety formats suitable for non-linear analysis: partial factor method, global format based on EN 1992–2, (2005) and fully probabilistic method. A new alternative safety format is also proposed by the authors (EVC), which is based on a semi-probabilistic estimate of the coefficient of variation of resistance. The discussed safety formats are tested on four examples. They include ductile as well as brit- tle modes of failure and second order effect (of large deformation). For the investigated range of problems, all the methods provide quite reliable and consistent results. The reliability of the numerical model was validated using experimental results. For this purpose the example of a rail- way bridge from the Sustainable bridges project was used. The numerical predictions show a very good agreement with the experimental load-displacement curves as well as the peak load carrying capacity. Based on the limited set of examples the following conclusions are drawn: The proposed EVC method gives consistent results compare to other approaches. The PSF method, which uses input parameters with partial safety factors appears to be sufficiently reliable and it is a natural exten- sion of the classical approach to the modern design methods based on non-linear analysis. Fully probabilistic analysis is sensitive to the type of random distribution assumed for input variables or resistance. It can provide additional load-carrying capacity if statistical properties of the analyzed system are known or can be accurately estimated. The methods are currently subjected to further validation by authors for other types of structures and failure modes. The research presented in this paper was in part resulting from: Grant no. 103/08/1527 of the Czech Grant Agency and European project Sustainable bridges TIP3-CT-2003-001653. The financial support is greatly appreciated.

718 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Complex multi-tool inspection of a masonry arch bridge using non-destructive testing

R. Helmerich, E. Niederleithinger & C. Trela BAM, Federal Institute for Materials Research and Testing, Berlin, Germany J. Bien Wrocław University of Technology, Wroclaw, Poland G. Bernardini IDS Ingeneria Dei Sistemi S.p.A., Pisa, Italy

ABSTRACT: The purpose of the paper is to present a case study about a complex special inspection of a masonry arch bridge in Poland according to the Guideline on Inspection and Condition assessment. The mentioned guideline summarizes the current state of the art of inspec- tion procedures for the step by step investigation of railway bridges applying enhanced methods developed during the EU-funded project Sustainable Bridges. In complex investigations, differ- ent non-destructive methods for the investigation of the same bridge parameters were applied to increase the quality of the final information of the bridge. Furthermore, the case study used the results from inspection as input in numerical calculations. In this case study, besides the arch barrel, the ballast quality and the inner structure of the backfill behind the brick arch barrel have been investigated with appropriate non-destructive methods. During the advanced inspection and survey we found irregularities of the structure not expected and unknown to the bridge owners, such as hidden chambers in the arch barrel and a load distributing concrete layer above the barrel. Adequate methods for their detection were e.g. impulse radar echo with different antennas and several antenna set-ups, spectral induced polarisation and the minimal- invasive testing of core samples from the arch. Furthermore, deformation measurements were performed using three different measurement systems as laser vibrometer, LTVD and microwave radar. The displacement measurement delivered data for input into the numerical calculation (Bien, Kaminski 2008).

Figure 1. Manual measurement at the arch barrel (bold line in track direction) with 500 and 900 MHz Antennas, right: a strong reflection zone visualizes a concrete drainage plate above the arch barrel. The concrete layer was confirmed by drilling from track level.

719 Besides the concrete drainage plates above the arch in Figure 1, the non-destructive testing method as ground penetrating radar documented hollows to serve as explosion chambers in WWII and humid areas in deeper parts of the masonry. The humid areas were confirmed by spectral induced polarisation. A prototype for a new rail based equipment with an optimised set up for a radar array can be used for the appraisal of deteriorating track bed. Finally, all information gathered from refined special inspection and displacement measurements were integrated in the refined assessment.

720 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Evaluation of Corrosion Situation on Reinforced Concrete By Portable Electrochemical Technique

Ralph Bäßler & Andreas Burkert Federal Institute for Materials Research and Testing – BAM, Germany Thomas Frølund COWI, Denmark

ABSTRACT: Within several EU-Projects different monitoring systems have been evaluated regarding their suitability for reduction of the inspection and maintenance costs as well as the traffic impairments. One part dealt with portable techniques for assessment of reinforcement corrosion. During this work potential field mapping and a portable equipment based on the Galvanostatic Pulse Method (GPM) were tested and compared in different situations at laboratory and on-site conditions. This paper deals with the results and analysis of the GPM measurements performed at laboratory and on-site conditions in comparison to results of potential mapping. Additionally results of average corrosion rates determined by weight loss and galvanostatic pulse technique were compared. Special attention was paid to the comparability of instrument readings to real behavior. The limitation and applicability of the technique on real structures have been evaluated on a bridge. Finally, the necessary precautions, which need to be taken when the on site data are used for service life prediction of structures, are discussed.

721 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Assessment of a railway concrete arch bridge by numerical modelling and measurements

G. He & Z. Zou Department of Civil, Environmental and Mining Engineering, Luleå University of Technology, Luleå, Sweden Civil Engineering, and Architecture College, Central South Forestry University, Changsha, Hunan, P.R. China O. Enochsson, A. Bennitz & L. Elfgren Department of Civil, Environmental and Mining Engineering, Luleå University of Technology, Luleå, Sweden A. Kronborg, B. Töyrä & B. Paulsson Banverket, Borlänge & Luleå, Sweden

ABSTRACT: The Swedish Railways have a history of about 150 years and several of the bridges along the lines have been in service more than 50 years, and some even more than 100 years. The owner, Banverket, wanted to increase the maximum allowed axle load from 225 kN to 250 kN along two of the main railway lines in northern Sweden. Two of the bridges are quite long and slender concrete arch bridges.At visual site inspections of one of the bridges, theVindel Bridge, it quite large movements in the transverse direction were observed from passing trains. Before the higher axle load could be allowed it was necessary to check the maximum deflection and the general dynamic behaviour of the two bridges. The Vindel River Railway Bridge was constructed in 1952 and has a height of 22 m and a span of 110 m. The other bridge with a similar design, the Långforsen Bridge, was built in 1960 and has a span of 90 m and a height of 13.7 m. In this paper, several finite element models of the Vindel River Railway Bridge are discussed. Field tests are carried out under service condition and with ambient vibrations. They were used to update and validate the FE models. At last, the refined models are used to check the possibility to increase the axle load. The calculated dynamic behaviour agrees well with the measured behaviour and the estimated displacements in vertical and horizontal directions shows that the Vindel River Bridge is capable to carry the higher axle load.

Figure 1. Elevation of the Vindel Railway Bridge and cross section of the deck and typical hollow section of the arch, each with locations points for the measurement sensors A to I. 722 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Field test – strengthening and monitoring of the Frövi Bridge

A. Kerrouche, W.J.O Boyle, Y. Gebremichael, L. Alwis & K.T.V.Grattan School of Engineering & Mathematical Sciences City University London, UK B. Täljsten & A. Bernnitz Luleå university of technology, Luleå, Sweden Denmark Technical University, Lyngby, Denmark

1 INTRODUCTION

This paper describes a strain measurement experiment carried out on a concrete railway bridge in northern Sweden. The experiment formed part of the EU ‘Sustainable Bridges’ project, which seeks to improve the capacity of Europe’s railway bridges to take high-speed and heavy axel loads on trains. For this experiment, Fiber Bragg Grating (FBG) sensors were fixed and protected on the grooves positioned on the underside of the bridge. The sensors were linked to a compact multiple-channel system based on WDM architecture using a scanning Fabry-Perot filter. The bridge (already scheduled for destruction) was then loaded until failure. This paper describes the measurement system and describes and discussed the test procedure and results obtained. Health and safety monitoring of civil structures such as rail bridges presents a real engineering challenge because of increasing capacity forces due to longer faster trains. Fiber Bragg Grating based mentoring system has advantages optic sensing schemes represent an enabling technology which offers a number of advantages for real time structural health monitor- ing of engineering materials and structures. Key advantages of their use include the potential for a large number of sensors to be multiplexed along a single length of optical fiber, enabling single or multi-point measurements to be made. Physical measurands of general interest for civil structures include strain, temperature, vibration and acoustic emission as well as chemical measurements including pH, moisture ingress, oxygen, chlorides and the presence of a variety of molecules car- ried by the moisture into the structure itself. The data obtained from such monitoring may be used to validate engineering designs, optimize manufacturing processes and thus facilitate structural health determination.

2 THE MONITORING SYSTEM

The compact 8-channel system used in this work is based on multiplexed WDM (Wavelength Division Multiplexing) architecture which used a scanning tunable filter to interrogate the 10 protected FBG sensors fixed into the concrete bridge. The basic sensing principle of the system is measuring wavelength shifts of the spectral output of the system, which is a convolution of the spectral peak of the interferometer in transmission and the Gaussian reflection profile from the Bragg grating sensor elements in response to changes in a measurand of interest.

3 SENSORS PLACEMENT INTO THE BRIDGE

The Övick Bridge used in this test was situated in Örnsköldsvik, Sweden. Built of reinforced concrete with two spans, it was built in 1955 and was taken out of service due to relocation of a new high-speed railway, the Botnia Line. 723 A groove, ∼2 mm wide and ∼1 mm deep, was cut into the centre of a square cross section of 10 mm of carbon fiber rod in order to accommodate the optic fiber. The fibres were glued with low viscosity Cyano-Acrylate and then covered with epoxy. Carbon Fibre Reinforced Polymer (CFRP) rods were used like strength tendons placed into appropriate channels already cut into the underside of the bridge to bond the carbon fiber rods with FBG sensors as well as other strain gauge sensors.

4 LOAD TEST

The loading arrangement was completed within two weeks before the final test. A beam was placed in the centre of the span to transfer the load, attached by two cables which were anchored into the ground (about 10 m below the bridge foundation slabs). Load on the span was increasing in stages, with time for relaxation and time for crack surveillance at major load intervals. According to preliminary estimation, the critical point load might reach up to 6 M Newton before failure. However, loading increased up to 12 M Newton when the failure occurred after more than one hour testing.

5 CONCLUSIONS

This paper demonstrates a new application test on a railway bridge loaded to failure in order to test its remaining ultimate load carrying capacity after a service period of 50 years. The system is proved to support measurement of high strain level of more than 5000 us. Also, this experience shows a very interesting opportunity to test how FBG systems deal with structures outside their normal working parameters, hopefully contributing to work on failure prediction in harsh engineering fields such as earthquake measurement, aerospace or maritime engineering.

724 Bridge Maintenance, Safety, Management, Health Monitoring and Informatics – Koh & Frangopol (eds) © 2008 Taylor & Francis Group, London, ISBN 978-0-415-46844-2

Author Index

Abdessemed, M. 562 Bernnitz, A. 723 Cha, S.-W. 371 Aboutaha, R.S. 228, 467, Berthellemy, J. 167 Chang, K.-C. 368, 512, 627 555 Betti, R. 614 Chang, K.C. 130, 194, 428 Abrahams, M.J. 636 Bian, L. 311 Chang, S.P. 63 Adey, B.T. 524 Bien, J. 206, 719 Chateauneuf, A. 292 Agano, Y. 561 Bie´n,J. 478 Chatelain, J.-L. 562 Agrawal, A.K. 644 Biondini, F. 268, 445, 452, 689 Chellini, G. 97, 179 Ahn, S.-s. 251 Bizindavyi, L. 559 Chen, C.-C. 139 Akbari, R. 117, 556 Bjerrum, J. 380 Chen, C.C. 194 Aktan, A.E. 387, 389, 399 Bleiziffer, J. 163 Chen, G. 161 Aktan, E. 397 Bloodworth, A. 472 Chen, S. 135 Aktan, H.M. 239 Bontempi, F. 267, 269, 448, Chen, S.S. 603 Alampalli, S. 351 687, 688 Chen, W.I. 130 Alwis, L. 723 Boro´nczyk-Plaska,G. 557 Chen, Z. 138 An, H.J. 352 Bortot, F. 170 Chen, Z.J. 135, 311, 316 Andersen, H. 462 Boscato, G. 299 Cheon, J.-H. 475 Andersen, J.E. 684 Boyle, W.J.O. 702, 723 Cheon, J.H. 191 Ang, A. H-S. 439 Branco, F.A. 83 Chin, W.J. 531 Arangio, S. 336 Brehm, M. 180, 183 Cho, B.-C. 665 Arizumi, Y. 294 Bressolette, Ph. 292 Cho, C.B. 197, 629 Asada, N. 617 Brown, C.J. 412, 413 Cho, E.Y. 120 Ashayeri, H. 556 Browne, E.H. 232 Cho, H.-N. 236, 681 Aso, T. 349 Brownjohn, J.M.W. 401 Cho, H.N. 539 Attanayake, U.B. 239 Brühwiler, E. 522, 705, 712 Cho, J.-Y. 544 Azevedo, A. 181 Buchanan, B. 397 Cho, J.R. 107, 109, 531 Burkert, A. 721 Cho, K. 107, 109, 275 Bae, B.S. 190 Burns, D. 153 Cho, K.-S. 460 Bae, D. 253 Byuoun, J. 569 Cho, K.I. 369 Bae, D.B. 131 Cho, S.W. 659 Bae, E.-H. 229 Cabral-Fonseca, S. 83 Cho, T. 540 Bae, I.H. 350 Calçada, R. 181, 182 Cho, Y. 326, 328 Bae, S.H. 246, 436 Cammarata, M. 652 Cho, Y.-R. 544 Bae, Y.-G. 463 Caner, A. 636 Choi, B.H. 122, 125 Bali, A. 562 Cantieni, R. 183 Choi, C.-Y. 103 Banakiewicz, A. 206 Cao, S.H. 135 Choi, D.-H. 129, 211 Bang, D.J. 126 Cardone, D. 601, 615 Choi, E.S. 531 Banks, J. 472 Carter, M. 227 Choi, H. 513 Barbier, V. 314 Casas, J.R. 473, 592, 707, Choi, H.R. 275 Bastidas-Arteaga, E. 292 708, 714, 715 Choi, J.-H. 124, 475 Bäßler, R. 721 Caspeele, R. 252 Choi, J.-S. 655 Beben, D. 256, 483, 628 Castelli, E.A. 165 Choi, J.H. 85 Bennitz, A. 722 Catbas, F.N. 399, 585, 587, Choi, K.-Y. 371, 665 Bergmeister, K. 495, 496, 593, 597 Choi, K.Y. 154 504, 505 Cavaco, J.A. 286 Choi, W. 121 Bergström, M. 698 Cervenka,ˇ J. 705, 718 Choi, Y. 276, 277 Bernardini, G. 719 Cervenka,ˇ V. 505, 718 Choi, Y.-S. 459 725 Choo, C.C. 625 Duffy, L. 547 Fu, G. 657 Choo, J.F. 419, 680, 693 Dutta, D. 649, 652 Fujimoto, T. 221 Chou, C.C. 194 Duwadi, S. 88 Fujino, Y. 49, 331, 346 Chromá, M. 507 Fukunaga, S. 119 Chryssanthopoulos, M.K. 284 Eden, R. 563 Fumoto, K. 313 Chun, P.-J. 657 Eichler, B. 710 Furlanetto, G. 447 Chun, S.-B. 455 El-Azazy, S. 513, 633 Furuta, H. 398, 415, 423 Chung, K.-Y. 464 Elfgren, L. 697, 698, 699, 701, Chung, M. 112 705, 722 Gabaldón, F. 184 Chung, W. 529 Ellis, R.M. 159 Gagnon, R. 159 Chung, Y.-S. 237, 612 Elnashai, A.S. 511 Galmarini, A. 684 Ciesla, J. 136 Emoto, H. 342 Garavaglia, E. 452 Collin, P. 90 Endo, K. 119 Ge, Y.J. 53 Correia, J.A.F.O. 290 Enevoldsen, I. 41, 380, 382 Gebremichael, Y. 723 Correia, J.R. 83 Engelund, S. 377, 378 Geier, R. 493 Cremona, C. 192, 314, 708, Enochsson, O. 697, 698, 699, Ghasemi, H. 389, 397 710 701, 722 Ghosn, M. 501 Cremona, C.F. 144, 167 Erhan, S. 199 Giacosa, L.M. 317 Crocombe, A.D. 550 Ettouney, M. 351 Gieseking, A. 287 Cruz, P.J.S. 700, 704, 716 Euler, M. 370 Giuliani, L. 687, 688 Cuadrado, M. 184 Eusébio, M.I. 83 Giuliano, F. 448 Cui, J. 153, 241, 542 Giussani, A. 263 Faddoul, R. 248 Gliši´c,B. 499 Da Lozzo, E. 433 Fallis, G.J. 563 Goicolea, J.M. 184 Dan, D. 327, 333, 410 Fan, C.P. 573 Gonzalez, P. 184 Daniel, R.A. 300 Fang, N. 623 González, A. 624 Dasgupta, S. 348 Farinholt, K. 348 Goto, S. 349 de Boer, A. 437 Farrar, C. 348 Grattan, K.T.V. 723 Deix, S. 491, 494 Farrell, A. 547 Green, M.F. 559, 563 Del Grosso, A. 329 Fassó, A. 264 Grendene, M. 266 Delgado, R. 182 Feng, J. 657 Guillier, B. 562 De Pauw, P. 87 Feng, L.Y. 316 Gul, M. 585 Dere, Y. 593 Feng, Y. 651 Gupta, N. 653 De Roeck, G. 97 Fernandes, A.A. 290 Gupta, R. 348 DesRoches, R. 545, 634 Ferreira, J.G. 83 Gupta, R.K. 142, 143, 431 De Salvo, V. 608 Ferretti Torricelli, L. 447 Gurian, P. 387 de Smet, C.A.M. 564 Fidler, P.R.A. 310 Gustavsson, L. 382 De Stefano, A. 317 Figueiras, J. 592 Gwon, Y.-S. 463 Dias, R. 184 Figueiredo, H. 182 Gylltoft, K. 605, 705 Diaz de León, A. 704 Figueiredo, M.A.V. 290 Györgyi, J. 480 Dicleli, M. 199, 638 Fisher, J.W. 3 Gómez, J.D. 473 Ding, H.S. 625 Flynn, E. 348 Dinitz, A.M. 560 Ford, J. 559 Ha, D.-H. 567, 610 Diniz, S.M.C. 446 Franchetti, P. 266, 433 Ha, G.H. 350 Djorai, B.(M.H.) 437 Frangopol, D.M. 319, 391, Haberland, M. 691 Dodson, A.H. 414 398, 399, 445, 452, 500, Hahm, D. 442 Doerrer, R.R. 201 501, 588, 590, 597 Hajdin, R. 524 Dogan, E. 636 Frantík, P. 504 Ham, H.J. 350 Doi, K. 230 Frederiksen, P.D. 684 Hammarbäck, J. 382 Dolce, M. 601 Frenz, M. 427 Han, D.-J. 487 Domaingo, A. 613 Friedl, H. 494 Han, G.-M. 463 Dong, X. 138 Frizzarin, M. 266 Han, S.-H. 439 Du, J. 487 Frølund, T. 721 Hanjari, K.Z. 481 726 Hara, T. 331 Huang, Y. 411 Johansson, B. 710 Harik, I.E. 625 Hussain, N. 227 Johnson, D. 605 Harries, K.A. 652 Huston, D. 153 Johnson, E.C. 88 Hartnett, M. 176 Hwang, D.-W. 665 Johnson, N. 514 Hassanain, H. 397 Hwang, E.-S. 477, 485 Jonsson, F. 605 Hattori, H. 195, 398 Hwang, I. 273, 419 Jung, D.S. 215 Haugeto, W.L. 235 Hwang, J.-K. 124 Jung, H.-g. 148 Hayashida, M. 523 Hwang, J.S. 130 Jung, H.J. 659 Hayrapetova, A. 474 Hwang, K.-J. 303 Jung, H.Y. 241 He, G. 722 Hwang, W.S. 146 Jung, K. 569, 667 He, X. 620 Hwang, Y.-K. 108 Jung, S. 603 Hechler, O. 90, 287 Jung, W.T. 111 Helmerich, R. 719 Iemura, H. 515 Jurado, C. 127, 365 Hennecke, M. 321 Igarashi, A. 515 Henriques, A.A.R. 716 Ihm, Y.-R. 236 Kagayama, T. 523 Henriques, J.F. 178 Im, D.K. 669 Kallinikidou, E. 594 Herwig, A. 705, 712 Im, K.-H. 274 Kami´nski,T. 478 Higgins, M.S. 141 Imam, B.M. 284 Kamiharako, A. 423 Hill, P.G. 310 Inaba, N. 294 Kamya, B.M. 169 Hindi, R.A. 201, 635 Inaudi, D. 499, 658 Kaneuji, M. 393, 394 Hino, Y. 543 Inman, D.J. 533 Kang, D. 529 Hirabayashi, Y. 294 Inoue, M. 230 Kang, D.-O. 371 Hirohata, M. 558 Irube, T. 570 Kang, D.O. 154, 436 Hoashi, H. 307 Ishida, J. 450, 520, 521 Kang, D.W. 122 Hodgson, I. 482 Ito, H. 331 Kang, J. 216 Hoehler, S. 710 Iwaki, I. 423 Kang, M.-S. 275 Hoffmann, S. 495, 496 Iyama, J. 289 Kang, S.-C. 293 Hoffmeister, B. 183, 287 Kang, S.C. 245 Hokayem, J. 248 Jalinoos, F. 153, 389, 397, Kang, S.G. 422 Holub, C.J. 511 596 Kang, S.T. 110 Homma, A. 525 Janda, Z. 718 Kang, W.-H. 546 Hong, A.L. 614 Jang, D.D. 659 Kano, M. 195 Hong, D.S. 532, 656 Jang, H.-S. 459 Karoumi, R. 708 Hong, H.-K. 612 Jang, J.-H. 320 Kasperski, M. 98 Hong, J.Y. 646 Jang, W.-S. 457 Kaszynska, M. 89, 91 Hong, K.J. 301, 302 Jang, Y.-I. 459 Kawamura, H. 393, 394 Hong, K.S. 234 Janjic, D. 613 Kawamura, K. 342, 344, 520, Hong, N.K. 190 Jee, H. 367 521 Hong, S.G. 190 Jensen, F.M. 377 Kawamura, S. 332 Hong, S.H. 273 Jensen, J.S. 379, 708 Kawaragi, H. 393 Hong, S.N. 566 Jeon, B.G. 177 Kawatani, M. 100, 195, 219, Hong, Y.H. 150 Jeon, S.-M. 238 221, 415, 606, 620 Hoorpah, W. 175 Jeong, J. 108 Kawatoh, C. 119 Hopf, S. 229 Jeong, S. 361 Kelly, J. 547 Hori, M. 400 Jeong, S.-W. 461 Kenai, S. 562 Hosseini, M. 611 Jeong, W. 137 Kerrouche, A. 702, 723 Hoult, N.A. 310 Jesus, A.M.P. 290 Kettil, P. 481 Hsu, C.C. 428 Jiang, C. 161 Khavari, S.R. 611 Hsu, D. 348 Jin, S. 330 Kibboua, A. 562 Huang, H. 653 Jo, B.-W. 675 Kim, B. 650 Huang, M.G. 434 Jo, B.W. 678 Kim, B.-G. 339, 519 Huang, S.X. 403 Jo, J.B. 123, 569 Kim, B.G. 434 Huang, W. 406 Jo, S.I. 550 Kim, B.H. 534, 679 727 Kim, B.S. 107, 109, 531 Kim, S.J. 511 Kwon, K. 590 Kim, C.-S. 460 Kim, S.T. 107, 109 Kwon, O.S. 511 Kim, C.-W. 100, 221, 606 Kim, S.W. 110 Kwon, S.-D. 456 Kim, C.H. 669, 684 Kim, S.Y. 107 Kwon, Y.B. 122 Kim, C.W. 219 Kim, T.-h. 203 Kyung, Y.-S. 249 Kim, C.Y. 215 Kim, T.H. 233 Kim, D. 241, 417, 542 Kim, W.-J. 460 Lagerqvist, O. 285 Kim, D.-H. 275 Kim, Y.-C. 312, 558 Lagoda, M. 136 Kim, D.-S. 103, 211, 551 Kim, Y.-H. 320 Lai, M.C. 428 Kim, D.H. 542 Kim, Y.-J. 203 Lambert, P. 232 Kim, D.J. 573 Kim, Y.-S. 129, 655 Lanata, F. 329 Kim, D.Y. 684 Kim, Y.C. 561 Larsen, J. 435 Kim, E.K. 468 Kim, Y.H. 217 Larsson, T. 285, 710 Kim, G.O. 243 Kim, Y.J. 197, 559, 563, 629 Lauridsen, J. 377 Kim, H. 137, 693 Kim, Y.M. 684 Le Cam, V. 314 Kim, H.-J. 339, 352 Kim, Y.P. 146 Leconte, R. 314 Kim, H.-K. 320 Kim, Y.T. 131 Lee, A.-Y. 276 Kim, H.J. 255, 567 Kim, Z.-C. 670 Lee, B. 348, 420 Kim, H.K. 680 Kitada, T. 617 Lee, B.-J. 475 Kim, H.S. 663, 669 Kitaura, R. 606 Lee, B.-j. 148 Kim, H.T. 671 Kiviluoma, R. 340 Lee, C. 468 Kim, H.Y. 418 Klatter, L.(H.E.) 392 Lee, C.-G. 234 Kim, I.H. 415 Kleywegt, H.S. 188 Lee, C.M. 275 Kim, J. 326, 328, 420, 569, Knight, T. 472 Lee, C.S. 234 648 Knippers, J. 303 Lee, D.-H. 612 Kim, J.-H. 455, 460, 670 Ko, J.-S. 359 Lee, D.-S. 129 Kim, J.-M. 464 Kobayashi, K. 162 Lee, D.H. 540 Kim, J.H. 246, 258, 479, 645, Kobayashi, Y. 591 Lee, E. 258 671 Koh, H.-M. 103, 190, 245, Lee, G.C. 368, 512 Kim, J.K. 479 255, 293, 442, 551, 567, Lee, G.H. 422, 471 Kim, J.-o. 148 693 Lee, H. 247, 320 Kim, J.S. 246 Koh, K.T. 110 Lee, H.-E. 118, 544 Kim, J.T. 218, 532, 656 Konaka, S. 606 Lee, H.-G. 670 Kim, J.W. 123 Kondo, T. 432 Lee, H.-M. 203 Kim, K. 488 Kong, J.-S. 247, 421, 449 Lee, H.J. 367, 659 Kim, K.-S. 203 Koo, I.M. 275 Lee, H.M. 85 Kim, K.-T. 457, 459 Koo, J.H. 659 Lee, H.S. 150, 205, 216, 217, Kim, K.-Y. 610 Koo, K.Y. 241, 646 273, 667 Kim, K.H. 234 Kozikowski, M. 435 Lee, I.K. 422, 678, 679 Kim, K.T. 680 Kpotufe, S. 348 Lee, I.-K. 681 Kim, M. 357, 363 Krecak, A. 476 Lee, J. 112, 208, 278, 675, 676, Kim, M.-S. 208 Kronborg, A. 701, 722 693 Kim, M.-Y. 249, 475 Kubota, K. 331 Lee, J.-H. 124 Kim, N.S. 177, 530 Kudo, K. 393 Lee, J.-J. 485 Kim, R.-G. 229 Kudo, M. 230 Lee, J.-S. 236, 249 Kim, S. 277, 364, 419, 500, Kuhlmann, U. 370 Lee, J.-W. 534 555, 603, 675 Kunz, J. 564 Lee, J.H. 276 Kim, S.-J. 293, 339 Kusuhara, S. 119, 313 Lee, J.J. 477 Kim, S.-W. 203 Kwahk, I.J. 197, 567, 629 Lee, J.S. 273, 277, 419, 529, Kim, S.-Y. 229 Kwak, K. 112 589 Kim, S.B. 154 Kwark, J.W. 531 Lee, K.-J. 421 Kim, S.H. 369 Kwon, H.-C. 462 Lee, K.M. 246, 436 Kim, S.I. 177, 530 Kwon, J.B. 422 Lee, M.-J. 455, 456, 462 728 Lee, P.-G. 238 Majka, M. 176 Mufti, A.A. 33 Lee, S. 251 Malerba, C. 263 Muscolino, G. 608 Lee, S.-H. 339, 352, 456, 457, Malerba, P.G. 263, 268, 445 Musiani, D. 348 519 Mangat, P.S. 232 Lee, S.-L. 463 Manko, Z. 256, 373, 483, 549, Na, H.-S. 211 Lee, S.-Y. 449 579, 628 Na, U.J. 433 Lee, S.C. 191 Marchiondelli, A. 447 Na, W.B. 656 Lee, S.H. 434, 561, 671, 680 Marcotte, C. 192 Nagai, M. 324 Lee, S.J. 154 Marvel, L.A. 635 Nagata, K. 583 Lee, S.L. 471 Masala-Buhin, M. 476 Nagayama, T. 575 Lee, S.W. 302 Mascarenas, D. 218, 348 Naitou, Y. 332 Lee, S.Y. 255, 258, 567, 575 Masri, S.F. 594 Nakajima, H. 523 Lee, W.-S. 665 Mattsson, H.-Å. 429 Nakamura, H. 570 Lee, W.-t. 148 Matsuki, Y. 332 Nam, S.-S. 274 Lee, W.S. 171 Matsumura, M. 617 Namatame, N. 195 Lee, Y. 364 Matsuno, H. 432 Nardini, L. 97, 179 Lee, Y.-H. 108 Matsuo, K. 332 Nasarre, J. 184 Lee, Y.-K. 670 Mautner, M. 496 Nayeri, R.D. 594 Lee, Y.-S. 322 McCarten, P.S. 187, 308 Neves, S. 181 Lee, Y.H. 154 McGinnis, M.J. 147 Newhook, J.P. 33 Lee, Z.K. 194 Medani, T.O. 204 Niederleithinger, E. 719 Lehký, D. 504 Melbourne, C. 708 Nishiyama, S. 620 Leighton, J. 702 Meng, X. 401, 406, 407, 408, Nomura, Y. 100 Leung, C.K.Y. 704 410, 412, 413, 414 Nordin, H. 698 Li, H. 189, 409 Mercalli, A. 329 Nothnagel, M. 348 Li, Q. 164, 200 Mertol, H.C. 121 Novák, D. 498, 504, 505, 507 Li, X. 138, 643 Messervey, T.B. 319, 391 Nowak, A.S. 89, 435, 540 Li, X.-X. 430 Middleton, C.R. 310 Li, X.Y. 316 Min, M.K. 367 OBrien, E.J. 474, 624 Li, Y.B. 626 Min, Z. 327 Odent, N. 167 Liao, H.-K 166 Min, Z.H. 641 Ogata, N. 525 Lim, H. 357, 363 Mirmiran, A. 121 Oguni, K. 400 Lim, K.-Y. 279 Mitsunari, K. 396 Oh, H. 278 Limonta, P. 268 Miyamoto, A. 342, 344, 349, Oh, J.-K. 276 Lin, J. 139 353, 450, 520, 521 Oh, M.-S. 668 Lin, K. 348 Miyamoto, N. 331 Oh, M.S. 85, 191 Lin, Z.-Y. 627 Miyashita, T. 324 Oh, S.-B. 457 Liu, J. 623 Mizuno, Y. 346 Oh, S.-T. 603 Liu, K. 97 Modena, C. 266, 283, 433 Oh, S.H. 350 Liu, K.Y. 130, 428 Moerman, W. 87 Oh, S.M. 276 Liu, X. 204, 404, 651 Mohammadkhani-Shali, S. Ohama, T. 398 Livingston, R.A. 330 192 Ohdo, K. 358, 360 Loh, C.-H. 309, 644 Montalto, F. 387 Ohshiro, T. 525, 568 Loh, K.J. 648 Moon, C. 539 Ok, S.-Y. 103, 442 Lu, W. 588 Moon, D.-J. 274 Okamoto, T. 307, 331 Lundgren, K. 481, 705 Moon, F. 387, 397 Okuno, S. 432 Luo, Y. 404, 405, 406, 411 Moon, F.L. 389 Olaszek, P. 136 Lutomirski, T. 435 Moon, J. 118 Olofsson, J. 697, 699 Lynch, J.P. 345, 644, 645, 648 Moon, Y.S. 276, 277 Olson, L. 155 Lyu, K.W. 205 Morcous, G. 89 Ono, S. 294 Mordak, A.G. 549 Onoufriou, T. 550 Maalek, S. 117, 556 Moreu, F. 575 Orcesi, A.D. 144, 167 Maeda, K. 570 Morikita, K. 568 Oshima, Y. 591 729 Otani, Y. 396 Park, M.S. 678, 679, 680 Ralbovsky, M. 491, 494 Otsubo, Y. 100 Park, S. 228, 275, 533, 560 Ranasinghe, A.P. 235 Ou, Y.-C. 368, 512 Park, S.-H. 421 Raphael, W. 248, 292 Ozkaya, C. 636 Park, S.-K. 205 Rauert, T. 183, 287 O’Connor, A. 380, 382, 547 Park, S.-Y. 655 Reiterer, M. 492 O’Flaherty, F.J. 232 Park, S.D. 680 Restelli, S. 689 Park, S.H. 468 Reynders, E. 97 Paczkowski, P. 89 Park, S.K. 275, 566 Rezai, A. 596 Padgett, J.E. 545, 634 Park, S.Y. 109 Ribeiro, A.S. 290 Pae, H.-J. 551 Park, T. 575 Richard, B. 192 Paik, I. 417 Park, W. 255, 293, 442, 551, Richard, G. 159 Paik, I.R. 485 680 Ricles, J.M. 289 Paik, I.Y. 126 Park, Y. 693 Righiniotis, T.D. 284 Paik, J.-G. 359, 464 Park, Y.C. 295 Rizkalla, S. 121 Palmeri, A. 608 Park, Y.H. 111, 434 Rizzo, P. 652 Paoletti, I. 244 Park, Y.S. 125, 253 Ro, S.-K. 320 Pardi, L. 329, 601, 615 Patron, A. 710 Roberts, G.W. 412, 413, 414 Park, C. 420 Patsch, A. 229 Rodrigues, M.P. 83 Park, C.-H. 421, 676 Paulsson, B. 697, 699, 701, Roh, J.S. 120 Park, C.-Y. 612 722 Roh, Y. 650 Park, C.M. 678 Pedersen, C. 380, 382 Rosell, E. 705 Park, D. 420 Peeters, B. 97 Rosing, T. 348 Park, D.-U. 303 Peil, U. 427 Roy, S. 3, 295 Park, D.-j. 148 Pellegrino, C. 283 Ruan, X. 430 Park, G. 348, 533 Peng, J.X. 440 Ruiz, M.E. 165 Park, H. 223, 258 Penka, E. 321 Rus, G. 575 Park, H.J. 646 Perrone, G. 601, 615 Russo, S. 299 Park, H.W. 150, 217 Pessiki, S. 147, 482 Ryo, G.S. 110 Park, J. 112, 301 Petcherdchoo, A. 577 Ryu, J. 326, 328 Park, J.-C. 681 Petrini, F. 448 Ryu, Y.S. 218 Park, J.-G. 203 Petschacher, M. 508 Park, J.-H. 251, 462 Pezzetti, G. 263, 264 Saeki, M. 400 Park, J.-N. 519 Phares, B.M. 322 Sahnaci, C. 98 Park, J.C. 678, 679, 680 Pimentel, M. 592 Saiidi, M. 13, 514, 633 Park, J.G. 191, 475 Pipinato, A. 283 Saiidi, M.S. 513 Park, J.H. 218, 233, 237, 243, Plos, M. 481, 605, 705, 708 Sakagami, T. 312 488, 532, 656 Podroužek, J. 507 Sakai, S. 525 Park, J.I. 436 Popa, V. 470 Sakai, Y. 162 Park, J.J. 110 Pozzi, M. 658 Sakaida, M. 617 Park, J.M. 566 Prader, J. 397 Sakino, Y. 312 Park, J.S. 111 Prato, T.A. 165 Salgado, R. 700 Park, K. 459 Praxmarer, L. 492 Salvatore, W. 97, 179 Park, K.-H. 247, 421, 449 Proença, J.M. 178 Sanders, D. 514 Park, K.-L. 668 Psimoulis, P.A. 402 Santos, J. 592 Park, K.-S. 103, 442, 551 Pukl, R. 498, 505, 718 Sause, R. 295 Park, K.-T. 108 Pulido, M.D.G. 304 Scarpas, A. 204 Park, K.-W. 681 Puurula, A. 697, 699 Scheller, J. 99 Park, K.H. 258 Puz, G. 163 Schendel, I. 427 Park, K.T. 171 Pytharouli, S. 402 Schlune, H. 605 Park, M.-S. 519, 675, 676, Póvoas, A.A. 362 Schäfer, Th. 321 681 Seo, I. 364 Park, M.-Y. 681 Radic, J. 163 Seo, J. 137, 357, 363 Park, M.G. 258 Radomski, W. 557 Seo, J.W. 364 730 Seong, D.-J. 203 Spencer, B.F. 511 Tozser, O. 141 Seong, D.J. 85 Starossek, U. 99, 690, 691, Trela, C. 719 Sesar, P. 476 692 Trela, Ch. 478 Sgambi, L. 210 Steinhauser, W. 496 Tsai, M.-H. 627 Shan, D. 164, 200 Stewart, M.G. 440 Tsai, M.-S. 368, 512 Shao, X.D. 440, 526 Stiros, S. 402 Tsuji, T. 195 She, Y. 408 Strauss, A. 495, 500, 504, Tsuruta, H. 423 Sheng, L.-H. 594 507 Turer, A. 149, 152, 595 Shi, B. 318 Stucchi, R. 268 Täljsten, B. 697, 698, 699, 723 Shi, C. 559 Stütz, R. 491 Töyrä, B. 722 Shi, X.-F. 430 Su, C.K. 428 Shim, C.-S. 612 Sugiura, K. 583, 591 Shim, C.S. 237, 238 Sumitro, S. 307 Uang, C.M. 288 Shim, H.-B. 577 Sun, J.W. 539 Ubertini, F. 269 Shim, I.S. 663 Sun, L. 327, 333, 410 Uejima, H. 456 Shim, J.H. 367 Sun, L.M. 647 Uetsuka, H. 162 Shim, Y.-W. 603 Sun, R.J. 647 Ulku, A.E. 239 Shimozato, T. 294 Sun, Z. 641, 647 Ulstrup, O.B. 379 Shin, D.K. 120 Sundquist, H. 429 Umemoto, S. 331 Shin, H. 326, 328 Sung, Y.C. 428 Shin, H.-M. 191, 203 Suzuki, H. 570 Van den Buverie, N. 87 Shin, H.-Y. 665 Suzuki, M. 423 Veit-Egerer, R. 502 Shin, H.M. 85, 475 Svensson, H. 17 Vosooghi, A. 633 Shin, H.Y. 436, 663, 671 Swartz, R.A. 345 Voˇrechovský, M. 498 Shin, J. 420 Szabó, G. 480 Shin, K.J. 246 Sánchez-Silva, M. 292 Wada, Y. 525, 568 Shin, S. 223, 417 Wahbeh, M. 594 Shin, S.-H. 455, 461, 670 Taerwe, L. 87, 252 Walser, P. 229 Shin, Y.S. 233, 243 Takahashi, H. 358, 360 Walsh, C. 474 Shinagawa, K. 620 Takahashi, T. 568 Wan, Z. 414 Shinozuka, M. 433 Takahashi, Y. 396 Wang, C. 409, 651 Sikorsky, C. 288 Takanashi, S. 358, 360 Wang, J.-C. 512 Sim, H.B. 288 Takeuchi, A. 568 Wang, M. 645 Sim, J.G. 278 Takizawa, Y. 230, 461 Wang, M.L. 307 Simões, R.A.D. 716 Tamura, I. 583 Wang, P.-H. 368, 512 Siqueira, C.H. 140 Tanaka, N. 195 Wang, R. 406 Siringoringo, D.M. 49 Tang, M.-C. 25 Wang, W. 135 Sloth, M. 378 Tani, K. 195 Watanabe, E. 394, 537, 583 Sohn, H. 645, 649, 652 Tapan, M. 467 Wegian, F.M. 657 Soma, M. 393, 394, 415, 423 Tasbihgoo, F. 594 Weidner, J. 397 Somerville, P. 513 Teplý, B. 505, 507 Wendner, R. 495, 496 Son, H.-S. 124, 668 Tervo, M. 340 Weng, J.-H. 309 Son, W.-H. 275 Tessier, C. 314 Wenzel, H. 95, 502 Son, Y. 361 Thelandersson, S. 705 West, J.C. 635 Song, G.-Y. 371 Thompson, P.D. 159 Whang, J.W. 217 Song, G.Y. 154 Thun, H. 697, 699 Whang, S.-H. 275 Song, H. 405 Tinkey, Y. 155 Wight, R.G. 559, 563 Song, H.-W. 577 Tisalvi, M. 97 Willberg, U. 321 Song, J. 363, 546 Todd, M.D. 218, 348 Wipf, T.J. 322 Song, J.-Y. 665 Tonnoir, B. 192, 314 Wisniewski, D.F. 379, 707, Song, M.-K. 464, 540 Toriel, M. 167 715, 716 Sonoda, J. 344 Toyama, N. 313 Wolfe, R.W. 594 Sorrentino, G. 97 Toyooka, A. 515 Wolff, M. 690 731 Won, J.H. 369 Yao, P. 407 Yu, P.-J. 461 Wu, A.-L. 309 Yasumori, H. 568 Yu, X. 651 Wu, H.Y. 658 Yau, N.-J. 166 Yue, D. 409 Wu, W.-H. 139 Yazdani, N. 88 Yun, C.-B. 655 Wunderlich, Th. 321 Yen, B.T. 573 Yun, C.B. 533, 645, 646 Yen, C.-I. 166 Yun, H.-B. 594 Xiang, H.F. 53 Yeo, I. 326, 328, 529 Yun, J.H. 539 Xie, D. 333, 410 Yi, B.-J. 276 Yun, M.G. 663 Xie, Y. 408 Yi, J.-H. 352 Xu, L. 414 Yi, J.H. 532 Zabel, V. 180, 183 Xue, T. 311 Ying, T.-Y. 430 Zandonini, R. 170 Yoo, D. 361 Zatar, W. 625 Yabuki, T. 294 Yoo, D.-H. 359 Zaurin, R. 587 Yamada, I. 313 Yoo, H. 129, 211 Zeman, J. 575 Yamagami,T. 523 Yoo, J.G. 669 Zhang, N.N. 316 Yamaguchi, E. 332 Yoo, M.-S. 610 Zhang, Q. 623 Yamaguchi,T. 415, 591 Yoon, H.J. 197 Zhang, Q.W. 626 Yamasaki,Y. 460 Yoon, J.-Y. 237 Zhang, W. 161, 318 Yamazaki,A. 432 Yoon, J.G. 589 Zhang, X.W. 311 Yan, B.F. 526 Yoon, J.H. 369 Zhang, Y. 643, 645 Yan, W. 200 Yoon, K.-Y. 118 Zhang, Y.F. 318 Yang Son, J.H. 663 Yoon,T. 361 Zhou, W. 189 Yang, B.C. 403 Yoon,T.-Y. 534 Zhou, Z.F. 641 Yang, D.S. 566 Yoon,T.Y. 122, 125 Zhu, Y.Q. 318 Yang, H.W. 276 Yoshida, I. 219 Zhuang, Y. 657 Yang, J.-H. 668 Yoshida, K. 398 Zi, G. 278 Yang, J.N. 644 Yoshida, M. 617 Zia, P. 121 Yang, K.-T. 274 Yosuhisa, H. 610 Zilch, K. 321 Yang, Z. 333 Youn, S.G. 131, 253, 468 Zimmerman, A.T. 345 Yao, B. 161 Young, J. 472 Zonta, D. 170, 658 Yao, L. 406, 407, 408 Yu, K.G. 436 Zou, Z. 722

732