PB2004101677 ·11111111111111111111111111111111111111111111111111111111111111111
Research Progress and Accomplishments 2000 - 2001
Multidisciplinary Center for Earthquake Engineering Research University at Buffalo, State University ofNew York
May 2001
MCEER-Ol-SPOl Red Jacket Quadrangle, Buffalo, New York 14261 Tel: 016) 645-3391; Fax: 016) 645-3399; Email: [email protected] World Wide Web: http://mceer.bujJalo.edu Page Intentionally Left Blank Report Doumentation Page 11. Report No. 3. Recipient's Accession No. 50272-101 MCEER-01-SP01 4. Title and Subtitle 5. Report Date Research Progress and Accomplishments: 2000-2001 6/4/01
7. Authors 8. Perfoming Organization Report No. Multidisciplinary Center for Earthquake Engineering Research (MCEER) 10. Project! Task I Work Unit No
None 9. Performing Organization Name and Address 11.Contract (C) or Grant (G) No. (C)
(G) EEC-91 01471
12. Sponsoring Organization Name and Address 13. Type of Report / Period Covered Multidisciplinary Center for Earthquake Engineering Research Technical Report State University of New York at Buffalo Red Jacket Quadrangle 14. Buffalo, NY 14261
15. Supplementary Notes Funded principally by NSF, NYS, and the FHWA, the Center derives additional support from FEMA, other state governments, academic institutions, foreign governments and private industry.
16. Abstract (Iimit200 Words) This report offers a selection of papers chronicling technical achievements of the Multidisciplinary Center for Earthquake Engineering Research in the year 2000-2001. Papers included in this volume: 1) Earthquake Motion Input and Its Dissemination Via the Internet; 2) Addressing Political, Institutional and Behavioral Problems in Earthquake Hazard Mitigation Policies; 3)Large Scale Experiments of Permanent Ground Deformation Effects on Steel Pipelines; 4) Experimental and Analytical Study of Base-Isolation for Electric Power Equipment; 5) Recommended Changes to the AASHTO Specifications for the Seismic Design of Highway Bridges 6) Literature Review: Performance of Seismically Isolated Bridges; 7) Risk Based Assessment of Seismic Performance of Highway Systems; 8) Collapse Limit States of Frames; 9) Centrifuge-Based Evaluation of Pile Foundation Response to Lateral Spreading; 10) Buildings with Added Energy Dissipation Systems; 11) Cost Benefit Analysis and Evaluation for Mitigation of Lifelines; 12) Retrofit of Hospitals in the Eastern United States; 13) Passive Site Remediation for Liquefaction Mitigation; 14) Advanced GIS for Loss Estimation and Rapid Post-Earthquake Assessment of Building Damage; 15) Retrofit of the Ataturk International Airport Terminal Building; 16) Estimating Earthquake Losses for the Greater New York City Area.
17. Document Analysis a. Descriptors Earthquake Engineering. Ground Motion. Public Policy. Political Behavior. Institutional Behavior. Permanent Ground Deformation. Steel Pipelines. Base Isolation. Electric Power Equipment. Risk-Based Assessment. Seismic Performance. Highway Systems. Centrifuge Testing. Foundation Response. Lateral Spreading. Added Energy Dissipation Systems. Cost Benefit Analysis. Lifelines. Retrofitting. Hospitals. Eastern United States. Passive Site Remediation. Liquefaction Mitigation. Advanced GIS. Loss Estimation. Rapid Post-Earthquake Assessment. Building Damage. Ataturk International Airport Terminal Building. Earthquake Loss Estimates. New York City.
b. Identifiers/Open-Ended Terms c. COSATI Field/Group
18. Availability Statement 19. Security Class (This Report) 21. No. of Pages Unclassified 194 Release Unlimited. 20. Security Class (This Page) 22. Price Unclassified Page Intentionally Left Blank Foreword
by George C. Lee, Director, Multidisciptinmy CenterJor Earthquake Engineering Research
The vision of the Multidisciplinary Center for Earthquake Engineering Research (MCEER) is to help foster activities that will lead to more earthquake resilient communities.. Its mission is to discover, nurture, develop, promote, help implement, and in some instances pilot test, innovative measures and advanced and emerging technologies to reduce losses in future earthquakes in a cost-effective man ner. Research findings show that relatively new buildings and infrastructure that are designed and constructed to the current state-of-practice in earthquake engineering perform significantly better than older ones during earthquakes, and that the largest threat to society lies in the seiSmically vulner able infrastructure designed and constructed at a time when earthquake-resistant design had not yet matured. With that knowledge in mind, it is MCEER's view that the best way of achieving the stated vision of earthquake resilient communities in the short term is to invest in two focused system-integrated en deavors: the rehabilitation of critical infrastructure facilities that society will need and expect to be operational following an earthquake, more specifically hospitals and lifelines; and the improvement of emergency response and crisis management capabilities to ensure efficient response and appropriate recovery strategies following earthquakes. MCEER works with the entire earthquake loss reduction community, which consists of practicing engineers and other design professionals, policy makers, regulators and code officials, facility and building owners, governmental entities, and other stakeholders who have responsibility for loss re duction decision making, to ensure that research results are implemented to improve safety and ad vance earthquake loss reduction for government, private industry; and the public-at-large.
Earthquake Resilient Communities Through Applications of Advanced Technologies
Infrastructures that Must be Available f Operational following an Earthquake
Hospitals
More Cost Water, Gas Pipelines Earthquake Effective Retrofit Electric Power Infrastructure Strategies Networ'j( System
Bridges and Highways
The flowchart above schematically shows how the Center's research interests contribute collec tively to achieve the vision of more earthquake resilient communities. The papers in this volume
iii highlight efforts in intelligent response and recovery,hospitals, water and gas pipelines, electric power networks, and bridges and highways. These studies involve many different disciplines, whose work will cohesively join together toward successful implementation of design and retrofit techniques to protect urban infrastructures from earthquake damage. This report is the third in our annual compilation of research progress and accomplishments. It is available in both printed and electronic form (on our web site in PDF format at bttp:// mceer.buffalo.edu/publications/dejault.asp, under Special Publications). If you would like more information on any of the studies presented herein, or on other MCEER research or educational activities, you are encouraged to contact us by telephone at (716) 645-3391, facsimile (716) 645-3399, or email at [email protected].
iv Contents
Earthquake Motion Input and Its Dissemination Via the Internet 1 Apostolos S. Papageorgiou (Principal Author), Benedikt Halldorsson and Gang Dong
Overcomillg Obstacles to Implementation: Addressing Political, Institutional and 9 Behavioral Problems in Earthquake Hazard Mitigation Policies Daniel J Alesch and William J Petak
Large Scale Experiments of Permanent Ground Deformation Effects on Steel Pipelines 21 Koji Yoshizaki, nJomas D. O'Rourke, Timothy Bond, James Mason and Masanon Hamada
Experimental and Analytical Study of Ease-Isolation for Electric Power Equipments 29 M. Ala Saadeghvaziri and Maria Q. Feng
Recommended Changes to the AASHTO Specifications for the Seismic Design of 41 Highway Bridges (NCHRP PToject 12-49) Ian M. Friedland, Ronald L. Mayes and Michel Bruneau
Literature Review of the Observed Performance of Seismically Isolated Bridges 51 George C Lee, Yasuo Kitane and Ian G. Buckle . A Risk-Based Methodology for Assessing the Seismic Performance of Highway Systems 63 Stuart D. Werner
Analysis, Testing and Initial Recommendations on Collapse Limit States of Frames 73 Darren Vian, Mettupalayam Sivaselvan, Michel Bruneau and Andrei Iv!. Reinhom
Centrifuge-Eased Evaluation of Pile Foundation Response to Lateral Spreading 87 and Mitigation Strategies Ricardo Dobry, 'Ihrek H. Abdoun and Thomas D. O'Rourke
Analysis and Design of Buildings with Added Energy Dissipation Systems 103 Michael C Constantinou, Gary R Dm'gush, George C Lee (Coordinating Authory, Andrei lvL Reinhom andAndrew S. Whittaker
Using Cost-Benefit Analysis to Evaluate .Mitigation for Lifeline Systems 125 Howard Kumeuther (Coordinating Authory, Chris C)'1; Patricia Grossi and Wendy Tao
Retrofit Strategies For Hospitals in the Eastern United States 141 George C Lee, Mai Tong and Yasuhide Okuyama
Passive Site Remediation for Mitigation of Liquefaction Risk 149 Patricia M. Gallagher andJames K Mitchell
v Advanced GIS for Loss Estimation and Rapid Post-Earthquake Assessment of 157 Building Damage Thomas D. O'Rourke, Sang-Soo Jeon, Ronald T Eguchi and Charles K Huyck
Seismic Evaluation and Retrofit of the Ataturk International Airport 165 Terminal Building Michael C. Constantinou, Andrew S. Whittaker and Emmanuel Velivasakis
Estimating Earthquake Losses for the Greater New York City Area 173 Andrea S. Dargush (CoordinatingAuthor), MichaelAugustyniak, George Deodatis, Klaus H. Jacob, Geo1'ge M.vlonakis, Laura McGin(v, Guy JF. Nordenson, Daniel O'Brien, Scott Stanford, Bruce Swiren, Michael W Tantala and Sam Wear Contributors
Tarek H. Abdoun, ResearchAssistant Professor, Civil and Environmental Engineering Department, Rensselaer Polytechnic Institute Daniel J. Alesch, Professor, Department of Public and Environmental Affairs, University of Wisconsin - Green Bay Michael Augustyniak, Supervising Planner, New Jersey State Police Office of Emergency Management Timothy Bond, Laboratory Manager, Department of Civil and Environmental Engineering, Cornell University Michel Bruneau, Deputy Director, Multidisciplinary Center for Earthquake ·Engineering Research and Professor, Department of Civil, Structural and Environmental Engineering, University at Buffalo Ian G. Buckle, Professor, Department of Civil Engineering, University of Nevada, Reno Sungbin Cho, Research Associate, School of Policy, Planning and Development, University of Southern California Michael C Constantinou, Professor and Chair, Department of Civil, Structural and Environmental Engineering, University at Buffalo Chris Cyr, Undergraduate Student, Wharton Risk Management and Decision Processes Center, The Wharton School, University of Pennsylvania Andrea S. Da~ush,Associate Director for Education and ResearchAdministration, Multidisciplinary Center for Earthquake Engineering Research Gary F. Dargush,Associate Professor, Department of Civil, Structural and Environmental Engineering, University at Buffalo George Deodatis, Professor, Civil Engineering & Operations Research Department, Princeton University Ricardo Dobry, Professor, Civil and Environmental Engineering Department, Rensselaer Polytechnic Institute Gang Dong, Post-Doctoral Research Associate, Department of Civil, Structural and Environmental Engineering, University at Buffalo Ronald T. Eguchi, President, ImageCat, Inc. Selahattin Ersoy, Graduate Student, Department of Civil and Environmental Engineering, New Jersey Institute of Technology Maria Q. Feng, Associate Professor, Civil and Environmental Engineering Department, University of California, Irvine Ian. M. Friedland,Associate Director for Development,Applied Technology Council Patricia M. Gallagher,Assistant Professor, Department of Civil and Architectural Engineering, Drexel University Siang-Huat Goh, Research Assistant, Department of Civil and Environmental Engineering, Cornell University
vii Juan Gomez, Graduate Student, Department of Civil, Structural and Environmental Engineering, University at Buffalo Patricia Grossi, Post-Doctoral Researcher, Wharton Risk Management and Decision Processes Center, The Wharton School, University of Pennsylvania Benedikt Halldorsson, Graduate Student, Department of Civil, Structural and Environmental Engineering, University at Buffalo Masanori Hamada, Professor, Department of Civil Engineering, Waseda University,Japan Wilhelm Hammel, Senior Consultant, KPMG, Germany Charles K. Huyck, Vice President, ImageCat, Inc. Klaus H. Jacob, Senior Reserch Scientist, Seismology Department, Geology and Tectonics, Lamont-Doherty Earth Observatory Sang-Soo Jeon, Graduate Research Assistant, Department of Civil and Environmental Engineering, Cornell University Yasuo Kitane, Graduate Student, Department of Civil, Structural and Environmental Engineering, University at Buffalo Howard Kunreuther, Co-director, Wharton Risk Management and Decision Processes Center, The Wharton School, University of Pennsylvania George C Lee, Director, Multidisciplinary Center for Earthquake Engineering Research and Samuel P. Capen Professor of Engineering, Department of Civil, Structural and Environmental Engineering, University at Buffalo Li Lin, Associate Professor, Department of Industrial Engineering, University at Buffalo Wei Liu, Ph.D. Candidate, Department of Civil, Structural and Environmental Engineering, University at Buffalo James Mason, Graduate Research Assistant, Department of Civil and Environmental Engineering, Cornell University Ronald L Mayes, Independent Consultant, Moraga, California Laura McGinty, GIS Specialist, Westchester County Geographic Information Systems, Department of Information Technology James K. MitcheU, University Distinguished Professor Emeritus, Department of Civil Engineering, Virginia Polytechnic Institute and State University James E. Moore, III,Associate Professor, Departments of Civil Engineering and Public Policy and Management, University of Southern California Nobuo Murota, Bridgestone Company, Japan George Mylonakis, Assistant Professor, Department of Civil Engineering, City University of New York Guy J.P. Nordenson,Associate Professor, Civil Engineering & Operations Research Department, Princeton University Daniel O'Brien, Earthquake Program Manager, New York State Emergency Management Office Yasuhide Okuyama,Assistant Professor, Department of Architecture and Planning, University at Buffalo Thomas D. O'Rourke, Thomas R. Briggs Professor of Engineering, Department of Civil and Environmental Engineering, Cornell University Apostolos Papageorgiou, Professor, Department of Civil, Structural and Environmental Engineering, University at Buffalo
viii William J. Petak, Professor, School of Policy, Planning and Development, University of Southern California Rajesh Radhakrishnan, Graduate Student, Department of Civil, Structural and Environmental Engineering, University at Buffalo Oscar Ramirez, Graduate Student, Department of Civil, Structural and Environmental Engineering, University at Buffalo Ricardo Ramos, Graduate Student, Civil and Environmental Engineering Department, Rensselaer Polytechnic Institute Andrei Reirihorn, Professor, Department of Civil, Structural and Environmental Engineering, University at Buffalo M. Ala Saadeghvaziri,Associate Professor, Department of Civil and Environmental Engineering, New Jersey Institute of Technology Ramesh San!, Ph.D. Candidate, Department of Civil, Structural and Environmental Engineering, University at Buffalo Masanobu Shinozuka, Fred Champion Chair in Civil Engineering, Department of Civil Engineering, University of Southern California Ani N. Sigaher, Ph.D. Candidate, Department of Civil, Structural and Environmental Engineering, University at Buffalo Mettupalayam Sivaselvan, Ph.D. Candidate, Department of Civil, Structural and Environmental Engineering, University at Buffalo Scott Stanford, Geologist, New Jersey Geological Survey Ernest Sternberg,Associate Professor, Department of Architecture and Planning, University at Buffalo Bruce Swiren, Natural Hazards Program SpeCialist, Federal Emergency Management Agency, Region II Michael W. Tantala, Ph.D. Candidate, Civil Engineering & Operations Research Department, Princeton University Wendy Tao, Undergraduate Student, Wharton Risk Management and Decision Processes Center, The Wharton School, University of Pennsylvania Craig E. Taylor, President, Natural Hazards Management, Inc. Mai Tong, Senior Research Scientist, MultidiSciplinary Center for Earthquake Engineering Research and the Department of Civil, Structural and Environmental Engineering, University at Buffalo Panos Tsopelas,Assistant Professor, Catholic University of America Emmanuel Velivasakis, Senior Vice President, LZA Technology Darren Vian, Ph.D. Candidate, Department of Civil, Structural and Environmental Engineering, University at Buffalo Jon S. Walton, Vice PreSident, MacFarland Walton Enterprises Yinjuang Wang, Graduate Student, Civil and Environmental Engineering Department, Rensselaer Polytechnic Institute Sam Wear, GIS Manager, Westchester County Geographic Information Systems, Department of Information Technology Stuart D. Werner, Principal, Seismic Systems and Engineering Consultants
ix Andrew S. Whittaker, Associate Professor, Department of Civil, Structural and Environmental Engineering, University at Buffalo Katrin Winkelmann, Visiting Research Associate, Department of Civil, Structural and Environmental Engineering, University at Buffalo, from the University of Darmstadt Eric Wolff, Graduate Research Assistant, Department of Civil, Structural and Environmental Engineering, University at Buffalo Yihui Wu, Research Scientist, Multidisciplinary Center for Earthquake Engineering Research and the Department of Civil, Structural and Environmental Engineering, University at Buffalo Tony Yang, Undergraduate Student, Department of Civil, Structural and Environmental Engineering, University at Buffalo Koji Yoshizaki, Pipeline Engineer, Fundamental Technology Laboratory,Tokyo Gas Company, Ltd., Japan
x Earthquake Motion Input and Its Dissemination Via the Internet
by Apostolos S. Papageorgiou(PrincipaIAuth01j, Benedikt Halldorsson and Gang Dong
Research Objectives The objectives of this task are to conduct research on seismic hazards, National Science Foundation, and prm1de relevant input on the expected levels of these hazards to Earthquake Engineering other tasks. Other tasks requiring this input include those dealing with Research Centers Program inventory, fragility curves, rehabilitation strategies and demonstration projects.The corresponding input is provided in various formats depend ing on the intended use: as peak ground motion parameters and/or re sponse spectral values for a given magnitude, epicentral distance and site Apostolos S. Papageorgiou, conditions; or as time histories for scenario earthquakes that are selected Professor, Gang Dong, based on the disaggregated seismic hazard mapped by the U.S. Geological Post -Doctoral Research Associate, and Benedikt Survey and used in the NEHRP Recommended Provisions Jar tbe Devel Halldorsson, Graduate opment of Seismic Regulations for New Buildings (BSSC, 1998). Stuknt, Department of Civil, Stntctural and Environmental rediction of the seismic hazard in tectonic regions of moderate-to Engineering, University at Plow seismicity is a difficult task because of the paucity of data that Blif.falo are available regarding such regions. Specifically, the problem of earth quake ground motion prediction in Eastern NorthAmerica (ENA) is hin dered by two factors: (1) the causative structures of seiSmicity in ENA are largely unknown, and (2) the recorded strong motion database is very limited (if at all existent), especially for moderate to large magnitude CMw ~ 6) events virtually for all epicentral distances. For these reasons, predic tion of strong ground motion in ENA makes the use ofwell-founded physi cal models imperative.These models should, among other things, provide the means to make extrapolations to large magnitudes and/or short dis tances with confidence. (This is contrast to the western U.S. (WUS), where the abundance of recorded strong motion data makes prediction of ground motion by empirical methods a viable procedure.)
Strong Motion Synthesis Techniques Our task is the synthesis of strong ground motion input over the entire frequency range of engineering interest. There are two approaches for modeling earthquake strong motion:
1 (l)The Stocbastic (Engineering) forms of this approach depending ~tocurrent Approacb, according to which, on whether the slip function (i.e., Researcb earthquake motion (acceleration) is the function that describes the evo Program 1: Seismic Evaluation modeled as Gaussian noise with a lution of slip on the fault plane) and Retrofit ofLifeline spectrum that is either empirical, and/or the Green functions are syn Systems or a spectrum that is based on a thetic or empirical.The Kinematic physical model (such as the SPecifiC Modeling Approacb involves the Program 2: Seismic Retrofit of Hospitals Barrier Model) of the earthquake prediction of motions from a fault source.This approach is expedient that has specific dimensions and and therefore cost-effective, and has orientation in a specified geologic been extensively used in the past setting As such, this approach more by engineers (using empirical spec accurately reflects the various wave tra) and recently by Seismologists propagation phenomena and is use (using spectra derived from physi fiu for site-specific simulations. cal models of the source). The in tent of this approach to strong Stochastic (Engineering) motion simulation is to capture the essential characteristics ofhigh-fre Approach quency motion at an average site Recognizing that the Stocbastic from an average earthquake of Approacb for synthesizing earth specified size. Phrasing this differ quake strong motion time histories ently, the accelerograms artificially is the most expedient method, generated using the Engineering ground motion synthesis efforts Approacb do not represent any spe were initiated with this approach. cific earthquake but embody cer The following computer codes tain average properties of past have been developed, and are avail earthquakes of a given magnitude. able to any interested user via the (2) The Kinematic Modeling Ap Internet: proacb was developed by seismolo 1. SGMP_ENA -Stocbastic Ground gists. In this approach, the rupture Motion Prediction for Eastern process is modeled by postulating Nortb America: Random Vibra a slip function on a fault plane and tion Tbeory (RV1): is used to es then using the Elastodynamic Rep timate the mean/expected resentation Theorem to compute values of peak ground motion the motion (e.g.,Aki and Richards, parameters (i.e., peak accelera 1980). There are several variant tion, peak velocity and peak
Thepserco111~~ty for thisresearch~s bQttf~~~J~~i¢fe sear:cijers an~ pr~Cii)g ehgineers whoritayti~.e'w,~'s.~~~c, .•.• ; input generated b'i.the. synthesis techniq~e~i 'C'~;" ~~~e.!~iif1 oped ~ 2 displacement), and spectral re tions ofthe 1997 NEHRP Provisions sponse amplitudes (e.g., Rice, (Bsse, 1998).) s#es 1944, 1945; Cartwright and Of all the source models, the Spe Engineering Seismology Longuet-Higgins, 1956; Shino cific Barrier Model is favored be Lalmratory: zuka andYang, 1971;Soongand cause it provides the most http://ciztileng.bzif.{alo.edu/ Grigoriu, 1993). complete, yet parsimonious, self engseislab/ 2. SGMS_ENA - Stochastic Ground consistent description of the fault Motion Simulation for Eastern ing processes that are responsible USGS National Seismic Hazards Project oftbe North America: Synthetic for the generation of the high fre Earthquake Hazards ground motions are generated quencies, and at the same time pro Program: using the Stochastic (Engineer vides a clear and unambiguous way http://geohazards.cr.usgs.gov/ ing) Approach briefly described of how to distribute the seismic eq/ above (e.g., Shinozuka and Jan, moment on the fault plane.The lat 1972; Shinozu~ and Deodatis, ter requirement is necessary for the 1991; Boore, 1983; Grigoriu, implementation of the Kinematic 1995). Modeling Approach described 3. RSCTHS - Response Spectrum above. We calibrated the SPecific Compatible Time Histories: Syn Barrier Model using the available thesis of ground motion time his strong motion data base of ENA tories that are compatible with earthquakes, as well as inferred prescribed response spectra us source spectra for the 25 Novem ing the Spectral Representation ber 1988 SaguenayearthquakeCMw Method (Deodatis, 1996). 5.8) and 19 October 1990 Mont Various earthquake source mod Laurier earthquake (Mw 4.5) els that have been proposed in the Quebec, Canada. The estimates of published literature [such as the the global and local stress drops Specific Barrier Model (Papa that resulted from the calibration georgiou and Aki, 1983a,b; 1985; are believed to be representative of 1988) and the ro2-model (Brune, earthquake sources in ENA.Thus, a 1970; Frankel et aI., 1996)), have "scaling law" of the source spectra been implemented (and are pro ofENA earthquakes has been estab vided as options) in computer lished. Such a "scaling law" allows codes SGMP_ENA and SGMS_ENA. the prediction/synthesis of ground [Both computer codes predict/syn motion at magnitude-distance thesize ground motions that are ranges that currently are not cov- compatible with the site classifica- • Table 1. Scenario Earthquake Events for Six Oscillator Periods for New York, NY Mw 4.8 4.8 5.7 5.1 6.2 6.2 R (km) 12 12 33 33 34 34 (factor) 0.712 1.253 1.458 1.692 1.170 1.999 Notes: Mw = Moment Magnitude R (km) = Epicentral Distance (factor) = scaling factor used to adjust/scale the synthetic ground motions of the modal event so that they are compatible with the Uniform Hazard Spectrum Eartbquake l}fOtiOll Input and Its Dissemination "Via tbe Internet 3 are linked with estimates of the Scaled Pseudo Spectral Accelerations for New York, NY (site B-C. SBM) -- NEHRP (B--C) design spectra seismic hazard produced by the US _ NEHRP (B-C) MCE spectra Geological Survey (USGS), the Na _t_ ~:~~~o~~~r~.8:m _ M =4.80, R=12.0 km w tional Seismic Hazards Project of ..... Mw=5.70, R=33.0 kIn _ M =5.70, R=33.0km w the Earthquake Hazards Program • _ •• Mw=6.20, R=34.0 km _ Mw =6.20, R=34.0 km (http://geohazards. cr. usgs.gov/ eq). Specifically, the disaggregated seismic hazard (Frankel, 1995; Frankel et aI., 1996; Harmsen et aI., 1999; Harmsen and Frankel, 2001) for five cities in the eastern U.S. (Boston, MA; Buffalo, NY; Charles 0.2 0.4 0.6 0.8 1 t.2 1.4 1.6 1.8 Period (s) ton, SC; Memphis, TN; and New York, NY) was obtained from the above USGS website. For each city/ • Figure 1. Thi~ figure displays: (1) The PseudO-Spectral Acceleration site, the disaggregated seismic haz (PSA) spectra of the scenario events tabulated in Table 1; (2) the ard is expressed in the form of a Uniform Hazard SPectrum, which is obtained from the PSHA; (3) the "modal event" for each one of six j,faximum Considered Earthquake (MCE) spectrum as specified by selected oscillator periods (0.1,0,2, the 1997 NEHRP Provisions on the basis of the "Uniform Hazard Spectrum"; (4) the Design Basis Eatthquake spectrum, which is 0.33, 0.5, 1.0 and 2.0 seconds) obtained by multiplying the amplitudes of the MCE spectrum by the ["modal event" is the seismic event factor (213). that makes the most significant contribution to the seismic hazard at a site for a specified oscillator New York. NY. M·=6.2. R'=34.0 km. Ty'= 1.00 sec .. Scale=1.170, Soil=BC (SBM) period; each "modal event" is char acterized/specified by a moment magnitude Mw and a distance R]. The spectral amplitude of the -~0~----7---~~---7----~--~1~0--~'2 "modal event" at the correspond ing oscillator period usually does not agree with the spectral ampli tude of the Uniform Hazard Spec i~~::1o 2 4 6 8 10 12 tra that are produced by the Probabilistic Seismic Hazard Analysis (pSHA) and are incorpo rated in building codes. [Usually, for (~.l6 10 12 Tlme(s) eastern North America, the ampli tudes of the modal events are • Figure 2, Scaled synthetic ground accelerations of the scenario events higher than the Uniform Hazard corresponding to oscillator periods 1.0 and 2.0 seconds, The soil Spectra, while in the western U.S. conditions correspond to the boundary B-C of site classes of the 1997 the reverse is true]. Current prac NEHRP Provisions (BSSC, 1998). tice recommends (Shome et al., 1998) to adjust/scale the entire ered by the existing strong motion spectrum of the "modal event" so database of ENA. that its spectral amplitude matches In order to make it more conve that of the Uniform Hazard Spec nient for the practicing engineer/ trum at the oscillator period to designer, the ground motion syn which it corresponds. thesis capabilities described above 4 For each one of the selected five eastern North America cities, fol lowing the procedure briefly de scribed above, we generated and posted on the web (http:// civil. eng. buffalo.edu/engseislab/) a suite of ground motion realiza tions (scenario earthquake mo tions). [The site class used for these simulations is the B-C NEHRP site class (BSSC, 1998), which is the ref erence site condition for the USGS seismic hazard maps.] Further more, the ground motion synthe sis computer codes discussed above are accessible and may be iii Figure 3. The PSA spe{;tra of ten realizations of a scenario event are used to generate additional time compared with the predktions of Random Vibration Theory histories, possibly for different soil conditions, ifthe user deems it nec oscillator period has the same PSA essary. value as the uniform hazard spec As an example, consider New trum at that period. York, NY. Table 1 shows the six The scaled synthetic ground ac "scenario earthquake events" that celerations (two horizontal and have been selected (based on the one vertical components) of the disaggregated seismic hazard pro scenario events corresponding to vided by USGS) for New York, cor oscillator periods 1.0 and 2.0 sec responding to the six oscillator onds are displayed in Figure 2.The periods that were mentioned soil conditions correspond to the above and essentially cover the fre boundary B-C of site classes of the quency range of interest for most 1997 NEHRP Provisions (BSSC, practical applications. 1998). Figure 1 displays the following: The consistency between the (1) the Pseudo-Spectral Accelera predictions of RVT with the syn tion (PSA) spectra of the thetic ground motions are demon abovementioned scenario events strated, as seen in Figure 3. The (the SPecific Barrier Model was agreement, as anticipated, is very used as the source model); (2) the satisfactory. Uniform Hazard Spectrum, which is obtained from the PSHA; (3) the Development of Kinematic Maximum Considered Earth quake (MCE) spectrum as speci Modeling Approach fied by the 1997 NEHRP Provisions Having completed the work re (BSSe, 1998) on the basis of the lated to the implementation of the Unifonn Hazard Spectrum; ( 4) the Stochastic (Engineering) Ap Design Basis Earthquake spec proach of strong motion synthesis, trum, which is obtained by multi efforts are now focused on the de plying the amplitudes of the MCE velopment of the Kinematic Mod spectrum by the factor (2/3). Note eling Approach. As discussed in that a scenario event for a specific Eart!Jquake Lllotion Input and Its Dissemination Via the Internet 5 previous research progress reports small events that have been re JIii;;;'b;~7i;:: (papageorgiou, 2000), two impor corded in ENA (US and Canada) to propose tant elements are necessary for this by various networks (including simple method of ground motion synthe broadband, short period and anillytical sis: (1) Sub-event models which are strong motion networks) with functions to used to synthesize composite earth the intention of performing quake events; (2) Scattering effects analysis that will provide values describe near of the lithosphere which are very of important parameters (e.g., source pulses. important for stable tectonic envi quality factor, Q, and scattering These ronments such as that of Eastern coefficient g) of the scattering descriptions NorthAmerica.We have made sub characteristics of the lithosphere will be readily stantial progress regarding the first in ENA. The objective is to syn useful to element. Specifically, we have de thesize results developed in the engineers lind rived closed form solutions for the field of Stochastic Seismology mllybe seismic radiation of a new class of with techniques for the genera included in kinematic models (asymmetrical tion of evolutionary stochastic future 'Versions circular and elliptical crack models) processes developed in the field of the building that will be used to simulate sub of Probability Theory [such as code." events in the synthesis. of strong the Spectral Representation ground motion generated by large (Shinozuka and Deodatis, 1991, earthquake events. These results see also Applied Non-Gaussian have been recently presented in a Processes (Grigoriu, 1995)]. forum consisting primarily of seis • Variable size sub-events: We mologists ("International Work have collected all the necessary shop on the Quantitative literature related to Poisson Ran Prediction of Strong-Motion and dom Pulse Processes and are the Physics of Earthquake well pOSitioned to investigate Sources" held in Tsukuba, Japan, the consequences of using vari Oct. 23 - 25, 2000) and were very able size sub-events on synthetic well received. These results have motions. been presented in papers recently • Sub-event Models:We have made submitted to the Bulletin of the substantial progress in develop Seismological Sodety. ing closed-form mathematical expressions of the far-field radia tion of new kinematic models to Conclusions - Future represent the sub-events. We Research want to extent this line of re search to other models of sub The efforts regarding the devel events such as asperity models. opment and implementation of the We thus aim at creating a "li Kinematic Modeling Approach to brary" of models of sub-events strong motion synthesis focus on adequate to simulate the various the following aspects of the modes of rupture of real faults. method: • Validation: Any simulation • Scattering effects of the lithos method should be validated by phere: Scattering effects are a comparing the synthetic seismo very important consideration in grams against the recorded ones the synthesis of ground motion for as many earthquake events in ENA. We have initiated the as possible. We have initiated collection of seismograms of 6 such validation comparisons us objective of such an investiga ing the 1988 Saguenay earth tion is to propose simple analyti quake event as a case study. cal functions to describe near • We intend to address also the source pulses. Such simple de issue of near source ground mo scriptions will be readily useful tions.In particular, there are two to engineers and may be in competing physical effects that cluded in future versions of the would affect near-source ground building code. . motions in ENA: Earthquake • Finally, in response to the needs sources in ENA are characterized of various MCEER investigators, by higher stress drops and we intend to provide ground shorter rise times as compared motion synthesis capabilities for to corresponding motions in sites in the western U.S. (e.g., California.Thus ENA near-source California). For this, we need to "killer pulses" are expected to calibrate the SPecific Barrier be stronger and of higher fre ~Model using the extensive quency (i.e., shorter duration). strong motion database that ex On the other hand, ENA earth ists for the above tectonic re quake sources appear to occur gion.For this calibration, we plan at greater depths and thus the to exploit some recent develop source-to-station distance (and ments regarding the improved consequently geometric attenu mathematical description of the ation) is greater. It is of great geometric attenuation ofground practical significance to investi motion in the vicinity of an ex gate which one of the above two tended earthquake source. effects dominates. The ultimate References------ Aki, K. and Richards, P., (1980), Quantitative Seismology: Theory and Methods, Volume I, II, w: H. Freeman and Company, San Francisco, 948 pp. Boore, D. M., (1983), "Stochastic Simulation of High-Frequency Ground Motions Based on Seismological Models of the Radiated Spectra; Bull Seism. Soc. Am., Vol. 73, pp. 1865-1894. Brune, J. N., (1970), "Tectonic Stress and the Spectra of Seismic Shear Waves from Earthquakes; J Geophys. Res., Vol. 75, pp. 4997-5009. Building Seismic Safety Council (BSSC) (1998), 1997 NEHRP Recommended Provisions for the Development of Seismic Regulation for New Buildings, Federal Emergen<-y Management Agency, Washington, D.C., FEi\1A 302. Cartwright, D. E. and Longuet·Higgins, M. S., (1956), "The Statistical Distribution of the Maxima of a Random Function; Proceedings of the Royal Statistical Society, Vol. 237, pp. 212-232. Deodatis, G., (1996), "Non-stationary Stochastic Vector Processes: Seismic Ground Motion Applications," Prob. Eng. Mech., Vol. 11, pp. 149-167. Fraukel,A., (1995), "Mapping Seismic Hazard in the Central and Eastern United States; Scism. Res. Lett., Vol. 66, pp. 8-21. Eartbquake i.vIotioll Input and Its Dissemination Via the Internet 7 References (Con't)------ Frankel, A., Mueller, C., Barnhard, T., Perkins, D., leyendecker, E.Y., Dickman, N., Hanson, S. and Hopper, M., (1996), National Seismic-Hazard Maps: Documentation, U.S. Geological Survey, Open-File Report 96-532, 110 pp. Grigoriu, M., (1995), Applied Non-Gaussian Processes, Prentice Hall, Englewood Cliffs, NJ, 442 pp. Harmsen, S. and Frankel, A., (2001), "Geographic Deaggregation of Seismic Hazard in the United States;' Bull. Seism. Soc. Am., Vol. 91, pp. 13-26. Harmsen, S., Perkins, D. and Frankel, A., (1999), "De aggregation of Probabilistic Ground Motions in the Central and Eastern United States;' Bull. Seism. Soc. Am., Vol. 89, pp. 1-13. Papageorgiou, A. S., (2000), "Ground Motion Prediction Methodologies fur Eastern North America," Research Progress and Accomplishments: 1999-2000, Multidisciplinary Center for Earthquake Engineering Researdl, University at Buffalo, pp. 63-69. Papageorgiou,A. S., (1988), "On Two Characteristic Frequencies of Acceleration Spectra: Patch Comer Frequency and fmax," Bull. Seism. Soc. Am., Vol. 78, pp. 509-529. Papageorgiou, A. S. and Aki, K., (1983a), "A Specific Barrier Model for the Quantitative Description of Inhomogeneous Faulting and the Prediction of Strong Ground Motion. I. Description of the Model," Bull. Seism. Soc. Am., Vol. 73, pp. 693-722. Papageorgiou, A. S. and Aki, K., (1983b), "A Specific Barrier Model for the Quantitative Description of Inhomogeneous Faulting and the Prediction of Strong Ground Motion. Part II. Applications of the Model," Bull. Seism. Soc. Am., Vol. 73, pp. 953-978. Papageorgiou, A. S. and Aki, K., (1985), "Scaling law of Far-field Spectra Based on Observed Parameters of the Specific Barrier Model;' Pure Appl. Geophys., Vol. 123, pp. 354-374. Rice, S. 0., (1944), "Mathematical Analysis of Random Noise;' Bell Syst. Tech. J., Vol. 23, pp. 282-332. Rice, S. 0., (1945), "Mathematical Analysis of Random Noise (concluded)," Bell Syst. Tech. J, Vol. 24, pp. 46-156. Shinozuka, M. and Deodatis, G., (1991), "Simulation of Stochastic Processes by Spectral Representation;' Appl. Mech. Rev., ASME, Vol. 44, pp. 191-204. Shinozuka, M. and Jan, c., (1972), "Digital Simulation of Random Processes and Its Applications," J Sound. Vib., Vol. 25, pp. 111-128. Shinozuka, M. and Yang, ].-N., (1971), "Peak Structural Response to Non-stationary Random Excitations," J Sound Vib., Vol. 16, pp. 505-517. Shome, N., Cornell, A., Bazzurro, P. and Carballo,]. E., (1998), "Earthquakes, Records and Nonlinear Response," Earthquake Spectra, Vol. 14, pp. 469-500. Soong, T. T. and Grigoriu, M., (1993), Random Vibration of Mechanical and Structural Systems, Prentice-Hall, Englewood Cliffs, NJ, 402 pp. 8 Overcoming Obstacles to Implementation: Addressing Political, Institutional and Behavioral Problems in Earthquake Hazard Mitigation Policies by Daniel] Alesch and William] Petak Research Objectives TIlis project is aimed at bridging the three planes, from basic research, National Science Foundation, through enabling processes, to engineered systems. At the basic re Earthquake Engineering search plane, we have been working to improve our collective under Research Centers Program standing about obstacles to implementing mitigation practices, owner decision processes (in connection with other MCEER projects), and public policy processes.At the level of enabling processes, we have been seeking to develop an understanding of how obstacles to greater William]. Petak, Professor, mitigation can be overcome by improved policy design and processes. Schoolo/Policy, Planning, At the engineered systems plane, our work is intended to result in and Development, Uniz;ersity o/Southern practical guidelines for devising policies and programs with appropri California ate motivation and incentives for inlplementing policies and programs Daniel]. Alescb, Professor, once adopted. Department o/Public and This phase of the research has been aimed, first, at a thorough, Environmental Affairs, multidisciplinary review of the literatUl"e concerning obstacles to imple University o/Wisconsin mentation. Second, the research has focused on advancing the state of GreenBny the art by developing means for integrating the insights offered by diverse perspectives on the implementation process from the several social, behaVioral, and decision sciences.The research establishes a basis for testing our understanding of these processes in the case of hospi tal retrofit decisions. A s. (development continues to concentrate in high risk earthquake areas, ..lithe probability increases that disastrous earthquakes will occur. Pub lic officials face the prospect of dealing with earthquake crises that could have been reduced Significantly with the application of known technologies. Although earthquakes are an uncontrollable force of na ture, unnecessary losses in life and property and social disruptions are generally the result of not having implemented precautions that we know could have mitigated the losses. The primary research emphasis in earthquake hazard mitigation has been on developing increased knowledge about the earthquake phe nomena, increasing understanding of structural performance under earthquake conditions, developing advanced design methods and stan dards, and improving building codes. Improved knowledge about the 9 physical and technical aspects of earthquake hazard mitigation? the problem has led to the adop That is, how can we determine tion of public policies intended to whether a policy has been • The lnternationailnstitute reduce the probability of loss, but implemented appropriately or for Applied Systems they must be implemented appro successfully? Anal;'Si~; Vienna, Austria, priately to actually reduce risk. To • Second, what are the key vari provided Professor Petak day, there remains an inadequate ables thOUght to affect the suc witb extended lise ofits understanding of the barriers, im cess of implementation? Which research links and materiaL~ pediments, disincentives, and issues of those variables can be con during his recent sabbatic associated with implementing ap trolled to help ensure success leave so tbat he was able to initiate and develop the propriate earthquake hazard miti ful implementation? literature review. gation technologies and strategies, • Third, how can mitigation advo much less with overcoming those cates help to ensure that public • Representatives from the barriers. This research was under policies adopted in an attempt San Francisco Department taken because no complete con to reduce the probability of ofBuilding and Safety, ceptual model or empirically losses to life and property from structural engineers validated model exists to explain earthquakes are implemented (Degenkolb Engineering, mitigation adoption processes for successfully? How can we help Comartin andReis,. Daniel Shapiro), the Applied the rehabilitation of existing struc ensure that the resources de Technology Council, tures, either across or within spe voted to hazard reduction in California Seismic Safety cific stakeholder groups. pursuit of these policies are used Commission, and urban effectively? planning consultants Our flfSt paper is based on a re reviewed the draft Background view of the implementation litera document. Our broadly-based assessment of ture that has developed over the • Social Science scholars barriers to policy implementation past three decades.A rich body of from MAE, PEER, and resulted in a background working research on implementation exists MCEER reviewed early draft paper entitled Barriers to Success within the social and behavioral documents. ful Implementation ofEarthquake sciences, but very little of it ad Hazard Mitigation Policies. In it, dresses natural hazards risk reduc we address three questions that are tion. Consequently, we've drawn on central to implementation: a broad base of literature to draw • First, what constitutes appropri inferences about implementing ate, successful implementation earthquake hazard mitigation. The of public policies concerning 75 page draft report cuts across the The' primat"faud!enc~ i6'((}~!c~ep~~f0l11prlse~)t#() . 'groups.Otie groupfuc~Ude$~S~i~;~e~~~rs~dev~lo~ ers, b~ildersf buildi1lg0W'tters~~:an:d.. c;t~~~i0nS ..•whose actions of' irul.ctions ar~the .. tatg~~;~fii?9· ., '::irit!!rldedto et them to reduce riskstothetnSeft'es:mf .."efriiefu:lkrs g, " :' '-'/ ','~, ',""" '.',,',' ':. "\",' --, '\,0,:,': '" " ;'-\'<,. ·.\'''",>",<,,,~~'~?,:::;ffij}-1f'C'>'f~'W<''~4\>,·:;,:;,''::''"<' ,... " ,-"~'\' of the. comnlUnity from ..natQrltl: I1a~ar~\iEY"""<~"~'nte~ecorid '~oup>c~nsis1s of policYandP1"O~'~~:';;'.~~ppbli~ and private program managers wiil1P<' ... ,,' ~!1i!ies in earthquake hazard risk reduction atld:m1tii~ii()n.:.;~'/;·· '< \,:;:~::}.~~: :~~;::\~'~. '. '... ~- 10 social and behavioral sciences, intended as a conceptual break drawing on more than three dozen through; it is simply the most con 'oC~ Significant studies, and resulting in venient, useful way we know to Resean:b 37 basic propositions concerning organize this diverse, complex body • Our eiforts to seek greater impediments to effective imple of research. Our model also con llnderstandingofhotll to mentation. The propositions are tains non-organizational variables ensure increased application ofmethods Jor organized around a model we de that give it substance: the problem earlhquakehazanirisk veloped that attempts to identify giving rise to the policy, the policy reduction by public and key inter-institutional nexuses in itself, the characteristics of the sys private organizations cuts the implementation process. tem, and characteristics of the across the program Jocuses system's environment. Each of in lifelines, hospitals, and these is thought to affect implemen response and recot-ery, The Implementation tation. Finally, the model incorpo including community Web rates dynamic elements. These are sustainability and community resilience. characteristics of the system that We view organizations as open affect in the modeL Four points systems, comprising elements re • Hospital retrofit cases are about the model are particularly being used as a test bedJor lated to one another in ways such important. First, the entire process filrthering our knowledge that perturbations of one element is dynamic and, typically, iterative; andJor developing have ramifications for the others. policies are often revisited· after practical gUides to Organizations exist within environ having been enacted. Second, improving mitigation ments of varying complexity with practice, coonii7Ulting our policy gets defined and redermed which they inexorably interact. lIJork closely with that of at each step in the implementation Organizational systems, we believe, von WinterJelt, Tierney, process as it is interpreted and re are inherently unstable, requiring and otbers. Since the ality-checked by the participants in ltfCEER agenda is aimed at continual resources from the envi that organizational node.Third; ob the application ofnetIJ ronment to survive and requiring stacles to implementation can arise technolOgies to the continual adjustment simply to at each link in the implementation reduction ofrisk, maintain their relative position. Our implementation process. They can also arise at the model enables us to examine mul necessarily cuts across all points at which organizations and tiple organizations embracing all programs and is relevant processes are joined with one an the levels of government within the toeacb. other. Fourth, the nature of the en systems framework. tire process itself may engender In our work, we examined orga obstacles to implementation, par nizational variables, including ticularly if the process is long and policy makers, the program design complex, involving lots of actors ers, the program implementing and transactions. agency, staff assigned to implement To guide our work, we conceptu the program, and the target popu alized a simple model to embrace lation. We were unable to find a the processes that extend from model in the literature that em policy adoption through implemen braces the contributions ofthe vari tation by operating agencies (see ous social and behavioral science figures 1 and 2). We describe this concepts, constructs, and analysis. as a complex web of expectations We chose, therefore, to create a and actions. The model extends to model based on general systems include organizations with imple theory to try to embrace the mentation roles as well as the char breadth of implementation re acteristics of the implementation search.The original model was not Ol1e·rcoming Obstacles to Implementation 11 velopers, builders, building owners, and organizations whose actions or <---j--, Characteristics and perceptions of the problem inactions are the targets of policies intended to get them to reduce risks to themselves and other mem bers of the community from natu ral hazard events. Market Intermediary Organizations are de Environmental characteristics fined as private organizations and Characteristics of the public agencies that provide mort system as a whole gage monies, mortgage insurance, or insurance against losses from • Figure 1. Implementation Within a Single Jurisdiction natural hazard events including, es pecially, those whose policies and practices affect the behavior ofPri mary Target Organizations. Characteristics and perceptions of the We have defined Front Line Imple problem menting Organizations as agencies, typically public agencies, such as building and safety and code en forcement agencies, that are charged with program implemen tation.That is, they are organizations that are expected to allocate re sources, including time and person nel, to bring about the desired effects in the target organizations. • Figure 2. Implementation in a Multi-Level Governmental Setting We include the individuals within those organizations assigned to take process itself. Our focus is on action. Indirect Implementing Orga sets of organizations and the nizations, on the other hand, are systemic setting within which those organizations charged with they play their implementation designing implementation pro roles. We call these different sets grams which mandate others, like of organizations primary target local departments of building and organizations, market interme safety, to take action, inflict sanc diaries, front line implementors, tions for not taking action, or pro indirect implementors, nongov vide incentives to take action. ernmental policymaking partici Typically, these are viewed as fed pants, and public policymaking eral or state agencies responsible organizations. for ensuring that municipalities ad minister mitigation programs. Nongovermnental Policymaking Par Primary Actors in the ticipants are an important element Implementation in the implementation web. They Process comprise private organizations that participate in policy development, We define Primary Target Organi such as professional associations zations as architects, engineers, de- (SEAOC, BSSC), and private interest 12 groups that seek to influence policy evolved from one ofviewing imple (lCBO, trade organizations). Some mentation as simply the process of participants are fully engaged in carrying out policy directives to public policy formation to the ex where implementation "is now in tent that they are almost indistin tegral to the field of policy interven guishable from authorized public tion, including recognizing its policymakers in their sphere of in influence on policy formulation" fluence. (Calista, 1994, p. 117). Evidence Policy Making O..-ganizations are continues to mount demonstrating defined for our purposes as public that, often, policies are not imple legislative, executive (or occasion mented in accord with the policy ally judicial) entities that adopt and makers' intent. Indeed, it may be authoritatively state a policy in that successful implementation is tended to reduce risk to life and the exception r.ather than the rule. property from natural hazard Calista reports that the most preva events lent finding in implementation re search is that outcomes are either disappointing or unwitting (Calista, What Constitutes 1994, citing Derthick, 1990). Oth Successful Policy ers suggest that policy implemen Implementation? tation is "the continuation of politics with other means" (Majone At the simplest level, "implemen and Wildavsky; 1978). tation represents the faithful fulfill We have concluded that success ment ofpolicy intentions by public ful implementation is not the same servants"(Calista, 1994, p. 117). as solving the problem. It is entirely Newcomers to business and gov possible that a program could be ernment often assume that a policy, implemented exactly as intended once adopted, will be implemented and that the program is ineffective in accord with the policy makers' because the problem transformed intent. An increasingly rich body during the implementation period of research confirms what old or because the program was flawed hands know - that is just not the conceptually. Ifa local government case. Practitioners and scholars provides incentives for action by have come to understand that private actors, some will choose to policy adoption is simply one mile participate and others will not. It is stone in a continuing process of possible that an overwhelming addreSSing an issue. Researchers majority act in such a way as to have concluded, however, that convince even the most jaded skep implementation is a critical part tic that the policy has been imple of policy making process. Policy mented. Suppose, however, that is adopted and adapted and drifts, only 10 percent of those targeted morphs, and mutates through the by the policy and eligible to partici implementation process. The ex pate actually volunteer to partici tent of drift and mutation depends pate in the program or implement on a myriad of variables, only some the policy.Is policy implementation of which can be controlled by successful? Presumably not, be policy makers. Calista's assessment cause such a small proportion of of the field of study is that it has the target was reached. Careful di- Overcoming Obstacles to Implementation 13 agnosis of the implementation pro • Did the policy have the nomi rich body of cess might focus attention on a spe nally intended effect on the in research on cific aspect of the program tended target population? implementation designed to implement the policy, • To what extent were there un exists within such as providing greater incentive intended side effects and were the social and or engaging in more effective cam those side effects adverse? behavioral paigns to make members of the tar • To what extent did the various sciences, but get audience more aware of the elements of the implementation very little of it program. If each of the organiza network comply with policy di tions in the implementation net rectives? addresses does precisely what is called for in • What proportion of the target natural an agency's program plan, but, still, was reached? hazards risk the private citizens or organizations • Did implementation take place reduction." targeted for action fail to take the within a reasonable time frame? steps that bring about clear intent That is, if one expected imple of the public policy, is implementa mentation of a retrofit to be com tion successful or has it failed? pleted within five years and it We concluded, too, that success has been fifteen and the job is ful implementation is a matter of not yet completed, how effective degree. Consider the fanning out of has the implementation been? responsibility for implementation. • Were the costs of implementa A federal agency looks to fIfty states, tion acceptable and reasonable? each of which looks to a hundred or more municipalities, each of which looks to several employees, Barriers to each of whom tries to affect the Implementation behavior of a dozen or more indi We developed 37 propositions viduals or ftrms. What proportion concerning obstacles to implemen of the several hundred thousand tation, including obstacles through potential "implementations" in this out the policy development and example has to "take" for imple implementation process. For Year 4, mentation to be judged successful? we focused our attention on exam Successful implementation is ining reasons private organizations clearly, then, a relative concept. We might not implement risk reduction have to think of it in terms of the measures. We summarized those extent to which it has occurred impediments into four barriers. rather than whether it has occurred. Success, in the case of implemen tation, is not a matter of absolutes. Barrier 1. The organization does not perceive itself at risk. Criteria for Evaluating The first barrier to successful Implementation implementation of earthquake risk reduction is that the organization We have identified six criteria for facing the risk does not perceive evaluating the extent to which itself at risk. Most practitioners in implementation has been success the earthquake hazard field may ful: find it difficult to comprehend there are organizations that are not 14 fully aware of their exposure and gation does not always win those vulnerability; many are not, even in comparisons. Second, any risk re seismically active areas. In such duction measures must be congru cases, appropriate means for over ent with organizational culture, coming the ignorance barrier by goals, and priorities; if they are not, efforts to communicate the hazard then the investment is directed to along with improving the target ward other goals. Similarly, the pro organization's perceptions of expo posed mitigations must be sure, vulnerability; and probable ef congruent with organizational mo fects. tivation. Barrier 2. The organization Barrier 4. The individual may perceive itself to be at organization may be aware of risk, but does not perceive that the problem and risk it can do something to reduce reduction measures and be the risk. eager to reduce its risks, but Organizations may see them find that it is impossible to selves as having both exposure and take action at this time. vulnerability; but not know what to Any of several reasons can keep do to reduce the probability of loss an organization from acting when when the event occurs.This condi it knows the risk, understands there tion can exist when there is a small are ways to reduce risk, and is will inventory of acceptable risk reduc ing to move ahead. There may not ing actions or when no means have be space on the organizational yet been devised to mitigate the agenda right now. The organiza potential loss. The organizational tion may not have the capacity at dedsion makers may have a fatalis present in terms of financial tic mind set that dictates against strength or available skills. Or, it attempts at risk reduction. Or, the may be that the organizational problem may be viewed as intrac environment presents sufficient table by organizational decision ambiguity to cause the organization makers. to defer action. Banier 3. The organization Overcoming Barriers sees the risk and possible actions, but doesn't see taking to Implementation risk-reducing actions as in In Year 4, we also worked to de being in its best interest. velop a document useful to the de cision making community and to This barrier is not at all unlikely. integrate our work with other Decision makers must weigh the MCEER tasks. Our second working sure costs of risk reduction against draft, entitled Overcoming Barri the possible benefits and, then, ers To Implementing Earthquake compare those costs and returns Hazard Mitigations: A Practical against those of other possible in Guide, is an attempt to develop vestments. Earthquake hazard miti- practical means for overcoming Overcoming Obstacles to I1nplementation 15 barriers that occur throughout the the problem.To increase the prob ebave implementation network. The ability of implementation, design concluded tbat guidelines on which the draft is profeSSionals and public officials successful based are summarized below. must ensure that decision makers implementation Guideline 1: If a basic obstacle to see the link between the risk and is not tbe same taking precautions is an inaccurate the solution. as solving tbe assessment of risks by the target or Guideline 6: The prudent problem." ganization, then the hazards profes implementor does not assume that sional, if he or she expects to have the Primary Target Organization un any impact, must provide the orga derstands the availability of a range nization with a clear, compelling of solutions and the possibility of statement of the risks to the orga modifying solutions to match the nization. specific needs of the individual or Guideline 2: Risk reduction mea ganization. sures are more likely when the Pri Guideline 7: The more a problem mary Target Organizations see a clear is viewed intractable, the less likely link between the potential hazard it is that implementation will be ous event and adverse effects on successful. Hazard mitigators must them and their businesses. Prudent work to reduce ignorance about hazard mitigators work to make the the phenomenon, reduce diversity linkage clear to those organizations. in specific target populations, work Guideline 3:The probability of suc with reasonably sized target popu cessful implementation of a policy lations, and become more skilled in (in either a multi-organizational set understanding means for changing ting or in a single organization) in perceptions and behavior in the tar creases when both those involved get population. in policy making and implementa Guideline 8: Successful implemen tion have a similar perception of the tation of new policies and ap problem. It is sensible, then, to help proaches is more likely to occur ensure similar perceptions by ac promptly in organizations that are tive communication among the two traditionally amendable to change parties. or have a culture that embraces in Guideline 4: Problems often shift novation. out from under solutions, rendering Guideline 9: : The probability of policies obsolete, ineffective, or dys successful implementation in functional. Prudent implementors creases to the extent that actors in work to ensure that policies are the implementation process per modified as necessary to ensure ceive congruence between means they are both effective and appro and ends; that is, they will work priate, so that efforts at implement harder to ensure implementation if ing the policy will not diminish. they perceive that the policy and Guideline 5: For an organization to the programs designed to imple take steps to reduce its exposure ment the policy are appropriate, or vulnerability to earthquakes, key given their perception of the prob decision makers in the organization lem. must believe that practical steps Guideline 10: The probability of exist to reduce the risks associated successful implementation in either with the event or condition and a multi-organizational setting or in that those steps are congruent with a single organization increases to 16 the extent that various actors in the objectives, and that achieving the organization have similar goals with program objectives is consistent respect to risk reduction and buy with and supportive of the into the means selected for risk re organization's primary objectives. duction. Guideline 16: Private organizations Guideline 11: Risk reduction mea are more likely to implement risk sures are more likely to be faithfully reduction practices when they see implemented when the orgaruza- . that the risk poses a clear and tionalleadership's support is unam present danger to their enterprise. biguous, the order is widely Coupling risk reduction with rou publicized, the people charged tine business concerns, such as with implementation have every property and casualty insurance thing needed to implement the and related risk management con measures, and those charged with cerns, helps bring it to the atten implementation have no doubt of tion of the organizational decision the authority of the leadership to makers. issue the decision. Guideline 17: Public policies in Guideline 12: Unless the interests tended to induce private parties to of the various stakeholders, espe reduce natural hazard risks to the cially those of the PrlmaryTarget Or organization and to the public at ganizations, are accommodated at large are more likely to be imple some minimally acceptable level,it mented when the rmancial con is likely that mitigation policies and cerns of the private parties are programs will face guerilla action, acknowledged explicitly in the be subject to subsequent watering policy and provisions are made to down, and face court challenges. alleviate rmanciaI burdens associ Gnidelme 13: "Hazard mitigation ated with implementation. is not a technical exercise; it is in Guideline 18: Other things being herently and often intensely politi equal, successful implementation cal because mitigation usually depends on entrusting implemen involves placing cost burdens on tation to organizations with suffi some stakeholders,and may involve cient capacity to administer the a redistribution of resources. Haz program. Iflocal government agen ard mitigators must, therefore, de cies are called upon to implement velop political as well as technical risk reduction programs, they solutions" (Alesch and Petak, 1986). should be provided with the re Guideline 14: Policies are more sources necessary to do the job. likely to be implemented success Guideline 19: Implementation pro fully when they are entrusted for ceeds more effectively when "the implementation to organizations leaders of the implementing agen that embrace the same goals and cies possess substantial managerial values as those implicit or explicit and political skill and are commit in the policy. ted to statutory objectives (Sabatier Guideline 15: Organizations will and Mazmanian, 1979). work toward achieving successful Guideline 20: Smaller organiza implementation to the extent that tions may need technical assistance, they believe they can implement in the form of consultants or self the policy; that implementing the help instructional materials, to aug policywi11 achieve desired program ment their staffs so they are capable Overcoming Obstacles to Implementation 17 ffi!:7:"":~~7~"'!E«'>-"<~"~'/_N/__ ";',,Z;::J:-::~ Itt of making pmdent choices con integrate their critique into the Ii'<;ase studies cerning risk reduction for buildings document. involving and structures. We have also distributed our first seismic retrofit Guideline 21: In complex organi draft - the synthesis of the prior of hospitals are zational environments character research on implementation - to planned to ized by instability and change, it a half dozen scholars drawn from evaluate and may be useful to test implement the social and behavioral sciences. extend the public risk reduction programs Each of these scholars is associated propositions aimed at private organizations in with one of the three engineering pilot projects in a variety of settings. research centers (MCEER, MAE, or concerning This will help avoid implementa PEER) and is actively engaged in barriers to tion pitfalls that could come from earthquake hazard research. The implementation immediate, widespread implemen group of scholars collectively de and a means to tation. cided it would be appropriate to overcome them." Guideline 22: If the purpose of a meet in a central location to review public program is to induce private their work. The group will have as organizations to implement risk re sembled by the time this report is duction policies and practices,gov published and will have provided ernmental organizations should a substantive critique of our draft work to make it easy for the private report. Along with the critiques by organizations to understand the re the practitioners, this review will quirements and to facilitate imple guide our development of final mentation by the private documents. organizations. Several tasks dominate our cur rent research activities. The first is completing our review of the state Conclusions and of the art in understanding ob Further Research stacles to implementation and means for overcoming them. We In Febmary 2001, we distributed have essentially completed our as both drafts to a select group of four sessment of policy, intergovernmen teen practicing structural engi tal, interorganizational, political, and neers,public officials,and members process variables.We still have work of the earthquake hazard commu to complete on obstacles to imple nity who agreed to read both docu mentation associated with organi ments and to participate in a one zational behavior and decision day session in San Francisco to cri making. Second, we will revise our tique our work and our products. drafts based on critique from the The session, held in March, was scholars and practitioners who extremely successful. Because the have reviewed our work.Third, we reviewers were practitioners, the will begin case studies involving focus was primarily on the sec seismic retrofits of hospitals as a ond working draft. The partici means for evaluating and extending pants pointed out what was our set of propositions concerning useful and what was not, helped barriers to implementation and us clarify our target audience, and means for overcoming them. provided guidance on how we Fourth, we will continue the devel could make the draft more useful. opment of our conceptual model At this writing, we are working to 18 concerning the implementation Detlof von Winterfelt (University process. Development of the con Southern California) and Kathleen ceptual models has been guided by Tierney (Disaster Research Center, the existing research literature, but University of Delaware). Third, we refining and specifying the model will complete the development of is dependent on the data elicited our conceptual model of the imple in our stakeholder studies. mentation process.Finally,we expect Beyond this year, we hope to ex to complete two monographs for pand our hospital retrofit case stud publication by MCEER focusing on ies. Second, we will continue our means for overcoming barriers to current efforts to integrate ourwork implementation both in the public with other MCEER researchers, and private sectors. References------ Alesch, Daniel]., (1998), "Adopting And Implementing Performance-Based Seismic Design Standards,n paper prepared for the Earthquake Engineering Research Institute, Oakland, CA. Alesch, D.]., and Petak, w.J., (1986), The Politics and Economics of Earthquake Hazard Mitigation, Natural Hazard Research and Application Centet; University of Colorado, Boulder, CO. Bardach, Eugene, (1977), The Implementation Game: W'bat Happens After a Bill Becomes Law, Cambridge, Mass.: The MIT Press. Calista, Donald, (1994), ~Policy Implementation," Encyclopedia of Policy Studies (ed. Stuart Nagel), Marcel Dekker; New York, NY., pp 117-155. . Calista, Donald, (1986), Linking Policy Intention and Policy Implementation: The Role of the Organization in the Integration of Human Resources, Administration and Society, Vol. 18, pp. 263-286. Drabeck, Thomas E., Mushkatel, Alvin H. and Kilijanek, Thomas S., (1983), Earthquake Mitigation Policy: The Experience of Two States, University of Colorado, Boulder, Institute of Behavioral Sdence, Monograph 37. Earthquake Engineering Research Institute, (1988), Incentives and Impediments to Improving the Seismic Peiformance of Buildings, Oakland, CA. GodschaIk, David R., Beadey, Timothy, Belke, Philip, Brower, David J., Kaiset; Edward ]., BohI, Charles c., and Goebel, R. Matthew, (1999), Natural Hazard Mitigation: Recasting Disaster Policy and Planning, Island Press, Washington, D. C. Upsky; M., (1971), "Street Level Bureaucracy and the Analysis of Urban Reform;" Urban Affairs Quarterly, Vol. 6, pp. 391-409. Mazmanian, D. A. and Sabatier, EA., (1989), Effective Policy Implementation, Lexmgton Books, Lexington, MA. Majone, G. and Wildavsky, A., (1978), "Implementation as Evolution," Policy Studies Annual Review, Vol. 2. H. E. Freeman (ed.), Sage, Beverly Hills. CA May, Peter J. and 'Villiams, Walter, (1986), Disaster Policy Implementation: Managing Programs Under Shared GOvernance, Plenum Press, New York. Overcoming Obstacles to Implementation 19 References (Con't) Pressman and Wildavsky, (1984), 3rd edition, Implementation, University of California Press, Berkeley, California Sabatier, P. and Mazmanian, D., (1979), "The Conditions of Effective Implementation; a Guide to Accomplishing Policy Objectives," Policy Ana(Jlsis, No.5, pp. 481-504. Sabatier, P. and Mazmanian, D., (1981), "The Implementation of Public Policy: A Framework of Analysis," in Effective Policy Implementation, (eds. D. A. Mazmanian and P. A. Sabatier), Heath, Lexington, MA. Sabatier, P. A., (1986), "Top-Down and Bottom-Up Approaches to Implementation Research: A Critical Analysis and Suggested Synthesis;' Journal of Public Policy, Vol. 6, pp. 21-48. Taylor, Craig, Mittler, Elliot, and Lund, LeVal, (1998), Overcoming Barriers: Lifeline Seismic Improvement Programs, Monograph No. 13, American Society of Civil Engineers, Reston, VA. 20 Large Scale Experiments of Permanent Ground Deformation Effects on Steel Pipelines by Koji Yosbizaki, Thomas D. O'Rourke, Timothy Bond, James Mason and Masanorf Hamada Research Objectives The objectives of the research are to: 1) simulate in the laboratory full National Science FOllndation, Earthquake Engineering scale permanent ground deformation (pGD) effects on steel pipelines ~ith Research Centers Program elbows, 2) develop an extensive and detailed experimental database on Tokyo Gas Company, Ltd. pipeline and soil reactions triggered by earthquake-induced PGD, and 3) National Science Foundation, refine and validate analytical models so that complex soil-pipeline interac U.S.-Japan Cooperatit'e tions can be numerically simulated with the precision and reliability nec Research Jor ~Jitigation 0/ essary for planning and design. To accomplish these objectives, an Urban Earthquake Disasters international partnership was organized, involving the Tokyo Gas Com pany, Ltd., MCEER, and NSF through its program for USlJapan Cooperative Research in Urban Earthquake Disaster Mitigation. The project combines e.,"{perimental and analytical research performed atTok'Yo Gas facilities ~ith e.,"{perimental and analytical work undertaken at Cornell University. The Koji Yosbizalzi, Pipeline Engineer, Fundamental e.,"{periments at Cornell represent the largest simulations of PGD effects Technology Laboratory, on pipelines ever performed in the laboratory. Tokyo Gas Company, Ltd. Thomas D. O'RDurke, Thomas R. Briggs Professor uring earthquakes, permanent ground deformation (pGD) can dam o/Engineering, Timothy D age buried pipelines. Earthquake-induced PGD can occur as sur Bond, Laboratory Manager, face fault deformation, liquefaction-induced soil movements, and land andJames Mason, slides. There is substantial evidence from previous earthquakes, such Graduate Research as the 1983 Nihonkai-chubu (Hamada and O'Rourke, 1992), the 1994 Assistant, Northridge (O'Rourke and Palmer, 1996), and the 1995 Hyogoken Department o/Civil and Environmental Nanbu (Oka, 1996) earthquakes, of gas and water supply pipeline dam Engineering, Cornea age caused by earthquake-ifiduced PGD. More recent earthquakes, University including the 1999 Kocaeli and Duzce earthquakes in Turkey; and the Masanori Hamada, 1999 Chi-chi earthquake in Taiwan, have provided additional evidence Professor, Department oj of the importance of liquefaction, fault rupture, and landslides through Civil Engineering, Waseda their effects on a variety of highway; electrical, gas, and water supply University lifelines. Gas and other types of pipelines must often be constructed to change direction rapidly. In such cases, the pipeline is installed with an elbow that can be fabricated for a change in direction from 90 to a few de grees. Because elbows are locations where flexural and axial pipeline 21 deformations are restrained, con performed as part of MCEER Pro ~e centrated strains can easily accu gram 1 on the Seismic Retrofit Ptumers mulate at elbows in response to and Rehabilitation of Lifelines and • Cambridge University PGD. the NSF program for U.S./Japan The response of a pipeline el Cooperative Research in Urban • Kyoto University bow, deformed by adjacent Earthquake Disaster Mitigation. ground rupture and subject to the MCEER has a long history of • Rensselaer Poo'technic Institute constraining effects of surround productive collaboration with the ing soil, is a complex interaction Japanese earthquake engineering • University o/Southern problem. A comprehensive and research community. Seven U.S./ California reliable solution to this problem Japan workshops on the earth requires laboratory experiments quake performance of lifeline fa • University o/Tokyo on elbows to characterize their cilities and countermeasures three-dimensional response to against liquefaction have been co • Yamaguchi University axial and flexural loading, an ana sponsored by MCEER, and the lytical model that embodies soil proceedings of these workshops structure interaction combined have been published and distrib with three-dimensional elbow re uted by MCEER.The latest in this sponse, and full-scale experimen series of workshops (O'Rourke,et tal calibration and validation of al., 1999) was held in Seattle,WA the analytical model. in conjunction with the 5 th US. To resolve this problem, an in Conference on Lifeline Earth ternational team was organized. quake Engineering. The next The principal participants are To workshop is planned during the kyo Gas Company, Cornell and forthcoming year in Tokyo,Japan. Waseda Universities.The research U.S. participants in the work also involves the University of shops include MAE, MCEER, and Cambridge, UK, Rensselaer Poly PEER researchers. technic Institute, and the Univer sity of Southern California. Waseda University leads a consor Experimental and tium of Japanese university par Analytical Models ticipants that include Kyoto and One of the deformation condi Yamaguchi Universities and the tions of interest is illustrated in University ofTokyo. The work was Figure la that shows a pipeline Titb;~~rSOf'this research inc1ude:lmbllc and private util- . ityc9,mpani~,induding gas distribu!i0n companies, such as ... r0ky0(iasan.dMemphls Gas,. Light and Water; and water ·~tributipn, companies,such as the Ws Atigeles Departtnent ()fW-ater'itId Power (LADWP), and the East Bay Municipal /UtiliiY D~ttict (EBMuD). The research is also of interest to ... en~eering design and consulting companies. The experi tnen.tai4ata and analytical modeling procedures developed for thiS project are of direct relevance for underground gas, w~ter,Petroleum and electrical conduits. 22 with an elbow subjected to PGD associated with lateral spread, consistent with lateral spread landslides, and fault crossings, and 'oCu~ and/or landslides.Although lateral therefore applies to many differ Res~tJTt:b spreads and landslides involve ent geotechnical scenarios. In ad Program 1: Seismic complex patterns of soil move dition, the experimental data and Evaluafwn and Retrofit of ment, the most severe deforma analytical modeling products are Lifeline Networks tion associated with these of direct relevance for under Program 2: Seismic Retrofit of phenomena occurs at the elbows ground gas, water, petroleum, and Hospitals, Task 2.7a, Cost and near the margins between the electrical conduits. Benefit Studies of displaced soil mass and adjacent, A modeling technique, named Rehabilitation Using more stable ground.The deforma HYBRID MODEL, was developed Advanced TechnokJgies tion along this boundary can be for simulating large-scale pipeline simplified as abrupt, planar soil and elbow response to PGD displacement. Pipelines that can (Yoshizaki et al., 1999 and 2001). be sited and designed for abrupt The model uses shell elements for lateral displacement will be able the elbow where large, localized to accommodate complex pat strains occur. Shell elements are terns of deformation that fre located over a distance of 20 quently involve a more gradual times the pipe diameter from the distribution of movement across center pOint of the elbow. The the pipeline. Abrupt soil displace shell elements are linked to beam ment also represents the principal elements that extend beyond this mode of deformation at fault crossings. Figure 1b illustrates the concept of the large-scale experiments. A steel pipeline with an elbow is installed under the actual soil, using fabrication and compac tion procedures encoun tered in practice, and then subjected to abrupt lateral soil dis placement.The scale of the experimental facil ity is chosen so that large soil movements are generated, inducing soil-pipeline interaction unaffected by the boundaries of the test facility in which the pipeline is buried. The Ii Figure 1. E,,\llerimental Concept for PGD Effects on Buried Pipelines with Elbows ground deformation simulated by the experi- ment represents defor- mation conditions Large Scafe 'E.;(perim£llts ofPermattellt tjroum[ 'Deformation 'Effects on Steel Pipefines 23 even when the opposing ends of the elbow were deformed into contact with each other and strains as high as 70% were mea G sured.In contrast,leakage was ob served in the opening mode for all cases except those in which the deformation was restricted because of the experimental load ing device. The HYBRID MODEL was used to simulate the deforma A tion behavior of the elbows using linear shell elements. Very good agreement was achieved between the analytical and experimental results for all levels of plastic de formation. Movable box Large Scale Teflon sheet Fixed box Experiments 1 OO~mm-diameter pipe 2m 90-degree elbow A Figure 2 shows a plan view of Pulley system the experimental setup that con Sand storage bin Data aquisition system sisted of five main components, including a test compartment (A and C in the figure), pulley load • Figure 2. Plan View of Experimental Setup ing system (F), sand storage bin (G), sand container hoisted from storage bin to test compartment (not shown), and data acquisition distance. Soil-pipeline interaction system (H). The test compart under PGD is characterized by p ment was composed of a movable Y and t-z curves, linking soil box (A) and fixed box (C) within stresses on the pipe to the rela which the instrumented pipeline tive displacement between them. was installed and backfilled. The Lateral soil-pipeline interaction is L-shaped movable box had inside characterized on the basis of labo dimensions of 4.2 m by 6m by 1.5 ratory experiments originally per m deep. It was constructed on a formed at Cornell (Trautmann base of steel I-beams positioned and O'Rourke, 1985) and dupli overTeflon sheets that were fixed cated at Tokyo Gas experimental to the floor. The Teflon sheets facilities. provided a low-friction surface on In-plane bending experiments which the movable box was dis were conducted by Tokyo Gas placed by a pulley loading system. (yoshizaki et aI., 1999 and 2001) The fixed box, which was an on full-scale specimens of steel el chored to the floor, was designed bows in both the closing and to simulate stable ground adja- opening modes. No leakage was observed in the closing mode, 24 cent to a zone of PGD similar to that illustrated in Figure 1. A 100-mm-diameter pipeline with 4.l-mm wall thickness was used in the tests. It was composed of two straight pipes welded to a 90-degree elbow CE). The short section of straight pIpe CD) was 5.4 m long, whereas the longest section was 9.3 m. Both ends of the pipeline were bolted to reac tion walls.The elbows were com posed of STPT 370 steel (Japanese Industrial Standard, JIS-G3456) with a specified minimum yield stress of 215 MFa and a minimum ultimate tensile strength of 370 MPa. The straight pipe was com posed of SGP steel (JIS-G3452) with a minimum ultimate tensile strength of 294 MPa. About 150 strain gauges were installed on the pipe to measure strain during the tests. Extensometers, load cells, and soil pressure meters were also deployed throughout II Figure 3. Overhead View of Test Compartment Before (top) and the test setup. After (bottom) an E.,\.1Jeriment The pipeline was installed at a 0.9-m depth to top of pipe in each of four experiments. In each ex periment, soil was placed at a dif ferent water content and in situ container that was hoisted with density.All experiments were con the overhead conveyor. The sand ducted to induce opening-mode was placed and compacted in deformation of the elbow. 150-mm lifts with strict controls The experimental facility was on water content and in situ den- designed with the assistance of sity. One of the most significant the HYBRID MODEL that was challenges during the testing was used to simulate various testing the movement and controlled configurations and compartment placement of such large volumes dimensions. Significant character- of sand with water content that istics of the experimentalfacility was intentionally changed for are its size and volume. The stor- each experiment. age bin for the sand was over The movable box was pulled by three stories tall, with a capacity an overhead crane with an 8 to 1 for 75 tons. Approximately 60 mechanical advantage obtained tons of sand were moved from the through the pulley system shown storage bin into the test compart- in the figure. The maximum ca- ment for each experiment with a pacity of the loading system was Large Scare 'E;(periments ofPerman.ent qrouna'Deformation 'Effects on Steel Pipefines 25 simulations discussed in the next section. Analytical and Experimental Results Finite element analyses were conducted with the HYBRD MODEL to check the ability of the analytical simulations to capture key aspects of the pipeline and elbow response to abrupt lateral displacement. Figure 5a compares • Figure 4. Overhead View of Deformed Experimental Pipeline the deformed pipeline shape of the analytical model with mea sured deformation of the experi mental pipeline. There is 1 m of lateral displacement and excellent agreement between the 784 kN.The rate of displacement two, and there is obvious agree of the movable box was approxi ment between the analytical de mately 16mm/s. formation and the overhead view Figure 3 shows the ground sur of the deformed pipeline in Fig face of the test compartment be ure 4. Figure 5b shows the mea fore and after an experiment. sured and predicted strains under Surficial cracks can be seen in the maximum ground deformation on area near the pipeline elbow and both the tensile (extrados) and the abrupt displacement plane compressive (intrados) surfaces between the movable and fixed of flexure along the pipeline. Fig boxes after the test. In all cases , ures 5c and d show the measured planes of soil slip and cracking and analytical strains around the reached the ground surface, but pipe circumference in which the did not intersect the walls of the angular distance is measured from test compartment to any appre extrados to intrados of pipe, cor ciable degree. responding to 0 and 180°, respec Figure 4 shows an overhead tively. In Figure 5d, the data point view of the test compartment af with an upward arrow indicates ter soil excavation to the pipeline the maximum strain measured following one of the experiments. when the gauge was discon Because each experiment was run nected during the experiment.Be until a total displacement of about cause the disconnection occurred 1m, the analytical models can be before maximum deformation of tested and calibrated through a the elbow, it is likely that the ac broad range of deformation and tual strain was larger than the strain in the elbow. The deformed value plotted. Overall, there is shape of the pipeline can be seen good agreement for both the mag clearly in the figure. Its shape is nitude and distribution of mea remarkably consistent with the sured and analytical strains. shape shown by the finite element 26 The large-scale experiments c Experiment (extrados) have provided a ~ Experiment (intrados) 40 Bbow _ _ FEA (extrados) rich and compre FEA (intrados) hensive database m ~ r-r======,------, 10 for understand 8 -4 (;~ 20 £g '" ~ ing pipeline re "ecO ·2 .",.. " gi (j) - ;!! sponse to large 0 - "7;..- ~:a t: £ 0 [J PGD, particularly .l!!~ t~~~~~~~$> under conditions is'" 2 L.J...... I--L.....l..--,--L...I-...I.-.l...-L..J.....J , , ·10 -8 ~ -4 ·2 0 2 ·20 o· 5 10 15 where local con Distance from the comer to south (m) Distance from the east edge, Lp (m) straint at an el (a) Deformation of the pipeline after the test (b) Distribution of axial strain in the longitudinal bow leads to direction complex three- .A. Experiment (Longitudinal) A Experiment (Longitudinal) Experiment (Circumferential) • Experiment (Circumferential) . dimensional re • FEA (LongITudinal) FEA (LongITudinal) sponse and high FEA (Circumferential) FEA (Circumferential) 20 "40 ~=;:====:;=;:==::::;::=~~ strain concentra tions. Although 30 ...... r. s·'" the interpreta 10 ~onA-A 20 Large Scafe 'El(perittletlts ofPe11tUlnent grOund'Defonnation 'Effects on Steel Pipefines 27 ground deformation effects on planned in the forthcoming year pipelines ever simulated in the to investigate further the p-y char lab.The tests allow for calibration acterization for partially saturated of a sophisticated soil-pipeline in sand as a function a water con teraction analytical program de tent and compaction effort. Work veloped in conjunction with the also is planned with collaborating experimental work. The experi universities for developing the mental data and analytical mod next generation of analytical eling products are of direct model that will represent the soil relevance for underground gas, as a continuum with specific con water, petroleum, and electrical stitutive relationships capable of conduits. simulating large ground deforma Additional work at Cornell tion and its interaction with bur sponsored by Tokyo Gas is ied pipelines. References------ Hamada, M. and O'Rourke, T.D., Eds., (1992), Case Studies of Liquefaction and Lifeline Peiformance During Past Earthquakes, Vol. 1, NCEER-92-0001, National Center for Earthquake Engineering Research, University at Buffalo, April. Oka, S. (1996), "Damage of Gas Facilities by Great Hanshin Earthquake and Restoration Process;' Proceedings, 6th japan-US. Workshop on Earthquake Resistant Design of Lifeline Facilities and Countermeasures Against Soil Liquefaction, NCEER-96-0012, MCEER, University at Buffalo, pp.ll 1-124. O'Rourke, T. D. and Palmer, M. C. (1996), "Earthquake Performance of Gas Transmission Pipelines," Earthquake Spectra, Vol. 20, No.3, pp.493-527. O'Rourke, T.D., Bardet, J.P., and Hamada, M., (1999), Proceedings, 7"> US-japan Workshop on Earthquake Resistant Design of Lifeline Facilities and Countenneasures Against Soil Liquefaction, MCEER-99-0019, MCEER, University at Buffalo, Nov. Trautmann, C.H. and O'Rourke, T.D, (1985), "Lateral Force-Displacement Response of Buried Pipe;' journal of Geotechnical Engineering, ASCE, Vol.111, No.9, pp.1077- 1092. Yoshizaki, K., Hosokawa, N., Ando, H., Oguchi, N., Sogabe, K. and Hamada, M., (1999), "Deformation Behavior of Buried Pipelines with Elbows Subjected to Large Ground Displacement; journal of Structural Mechanics and Earthquake Engineering, No.626/I-48, pp.173-184 (in Japanese). Yoshizaki, K., O'Rourke, T. D. and Hamada, M., (2001), "Large Deformation Behavior of Buried Pipelines with Low-angle Elbows Subjected to Permanent Ground Deformation," journal of Structural Mechanics and Earthquake Engineering, I-2130, pp. 41-52. 28 Experimental and Analytical Study of Base-Isolation for Electric Power Equipments by M. Ala Saadeghvazir-i and Jl:laria Q. Feng Research Objectiyes The goals of this project are to develop rehabilitation strategies and en National Science Foundation, Earthquake Engineering hance seismic design guidelines for key equipments in power substations, Research Centers Program which are one of the most critical facilities in a power system. Further more, it will provide the knowledge base to be integrated into the overall loss estimation model for the entire power network.To achieve these goals, key substation equipments are identified and effectiveness of base-isola tion to increase their seismic resilience are assessed using a comprehen sive experimental and analytical study. AI. Ala Saailegbvaziri, Associate Professor and Selahattin Ersoy, Graduate Stllden~ ritical power system facilities, such as substations, sustained signifi Department ofCivil and Ccant damage in California and Japan earthquakes and more recently Environmental . during the 1999 Chi-Chi,Taiwan, and the Izmit,Thrkey earthquakes. Func Engineering, New Jersey tionality of electric power systems, especially in the age of information Institute ofTechnology technology, is vital to maintaining the welfare of the general public, sus Maria Q. Feng, Associate taining economic activities and assisting recovery, restoration, and recon Professor, Chil and struction of the seismically damaged built environment. Furthermore, Environmental Engineering Department, enhanced seismic design and rehabilitation will ensure long-term reliabil University ofCa/ijomia, ity and longevity of critical equipments. In order to be able to assess the [nine reliability of power systems and to develop seismic resistant mitigation Nobuo Murota, Bridgestone strategies, critical components must be identified and their seismic per Company,Japan formance evaluated.Transformers and their bushings are among the most critical components in a complex power system and their seismic perfor mance during past earthquakes has not been satisfactory. Figure 1 shows damage to a transformer in Izmit-2 substation during the Izmit, Turkey earthquake ofAugust 17, 1999 (EERI Newsletter, 1999). Generally, there are several modes of substation transformer failure dur ing an earthquake, namely movement and turn over of unanchored trans formers, anchorage failure that can cause ripping of the transformer case and oil leakage, foundation failure causing rocking and tilting, and failure of the gasket and oil leakage due to the interaction between the trans former and the busrung. Some transformers are supported on rails for ease of installation and because such a set up allows for air circulation to provide additional cooling, thus enhancing corrosion resistance.Another 29 form of damage in an unanchored transformer is large movement that • National Center for can cause damage to other Research in Earthquake components connected to Engineering (NCREE), the lmit, such as control Taipei, Taiwan cables and bushings. Dam age due to the latter fac • Earthquake Protection System, Inc., Richmond, tor (i.e., interaction California between transformer bushing and other equip • Bridgestone Corporation, ments) can be critical and Kanagawa-ken,Japan can cause significant dam- - Figure 1. Transformer Failure During the August 17, age and delays in service. 1999, Izmit, Turkey Earthquake (EERI, Oct. 99) There are many cases of bolt or It is more common to anchor weld failure during past earth transformers where considerations quakes (ASCE, 1999). However, are given to seismic loads.This can with better attention to the details be done by either bolting the trans in the design of supports, their per former to its footing slab or by formance during an earthquake can welding it to steel embedded in the be improved. Implementing well slab (ASCE, 1999). Designing the designed anchorage for retrofit of anchorage at the supports requires existing transformers,nevertheless consideration to large forces not can be difficult and costly. Furthe: only due to gravity and horizontal more, in many situations, for both seismic forces but also from over new and existing transformers, a turning moments in both direc well-designed anchorage may only tions. Furthermore, other change the mode of failure to the appendages such as radiators and foundation (i.e., the next weak link reservoirs are attached to a typical in the system). The failure shown transformer.These appendages can in Figure 1 appears to be due to cause significant torsional forces foundation settlement. that must also be considered in the Boundary gaps due to back and design of the supports. In addition forth motion of transformers and to strength, the anchorage system rocking of transformers and their must have adequate stiffness and footings due to soiI-stmcture inter tight tolerances to prevent initia action have been observed during tion of impact forces that can dam past earthquakes (ASCE, 1999). age internal elements or excite Therefore, in many cases, the use higher modes that can damage of base-isolation for transformers brittle porcelain members. "Pri1I1aryfise~ oft~sre~earcIi includeutUity companies and o:wner!iptelectrical substation eqD:ipments, manufac turers of electiical ~quipments, and manufacturers ofbase isolation dertc~s. The research results can further be used by structural and electrical engineers for design and ret rofit of electrical power equipments•. 30 may be the only suitable remedy to sponses of the individual systems 1):\ • alleviate these problems, especially involved as well as their interac Jj~~IJns study for existing transformers in high tions. represents the seismic regions. Base-isolation will Within this study, two base-isola first efforl in also reduce the input acceleration tion systems are being investigated testing hnse into the bushing and will lessen the under a collaborative effort among isolated l;a?ge interaction between the trans several institutions and industrial sc&ile trnns former and the bushing, which has partners. The following sections former-bushing been the cause of many bushing discuss the experimental program damages during past earthquakes. along with some of the findings. systems using Furthermore, by reducing the iner The two systems considered are a Rn enrlbquake tia forces, base-isolation can also friction pendulum system and a simVJlgto7.." prevent the possibility of internal hybrid system consisting of sliding damage. and rubber bearings. It represc:nts The after effect of an earthquake the first effort in testing base-iso on reliability and longevity of a lated large-scale transformer-bush transformer is directly related to ing systems using an earthquake the level of shaking of internal ele simulator. ments. High levels of uncontrolled shaking may very well reduce the life expectancy and reliability of Experimental Study internal elements. Internal damage An extensive series of tests is normally difficult to observe and were conducted on a transformer document because of limitations model supporting a bushing. The with post-earthquake inspections primary objective was to com of the transformer internal system. pare the response of a fIXed based However, there are reports of cases transformer-bushing system to of internal damage to transformers that of the system when isolated. (ASCE, 1999). The testing was conducted on the Satisfying the mobility require earthquake simulator at the Na ment for maintenance purposes is tional Center for Research on another advantage of base-isolation Earthquake Engineering (NCREE) over a tightly designed anchorage in Taiwan in collaboration with system. An issue with the use of the manufacturers. Uniaxial, bi base-isolation that demands careful axial, and triaxial excitations were consideration is the possible ad conducted employing several verse effect of relatively large dis earthquake records with PGAs in placements on the response of the range of 0.125g to O.5g. The inter-connecting equipments, espe testing schedule also included cially bushings. Among possible white noise tests to identify dy remedies to be considered are a namic characteristics of the bush balanced approach to the design of ings and the transformer model, the isolation system (displacement resulting in more than 200 tests. vs. inertia reduction), appropriate Considering the payload capac design of conductor slacks, and use ity ofthe earthquake simulator, the offlexible conductor connections. transformer model was designed to Therefore, a successful application weigh 235.5 kN. The model is a of this technology requires in four-layer steel frame structure depth understanding of the re- 'E:;q;erimenta[ ana.:.7l.nafyticaf Stucfg of'Base- Isofation for 'EiectTtC Power 'Equipments 31 quantifying the important interaction between the two components. Figure 2 shows isometric and el evation views of the entire model.As can be seen from this figure, the bush ing projects inside the transformer model as is the case in an actual situ ation. Two types of bushings, 69 kV and 161 kV, were used in the experi ment; Several earthquake records were used for the testing, among • Figure 2. 1\vo Views of them were the 1940 El-Centro, 1994 Transformer-Bushing Model Sylmar (Northridge), and 1995 Takatori (Kobe) records.Transformer frequency was measured to be 12.5 Hz in both x and y directions. The frequencies for the 69 kV and 161 kV bushings were 27.0 Hz and 12.5 Hz in both x and y directions, respec tively. The equivalent damping ratio for each component was arOlmd 2%. Friction Pendulum System Results One of the most recent base iso with lead blocks loaded to repre lation systems to improve the earth sent the internal core and coil of a quake resistance of structures is transformer. Friction Pendulum System (FPS), Consistent with the practice em shown in Figure 3. Detailed descrip ployed byTaiPower, the bushing was tion of the system along with in attached to the transformer model depth discussion of experimental top plate at a right angle. The con and analytical results can be found nection represents some of the im elsewhere (Ersoy, et. al 2001; portant structural aspects of the Saadeghvaziri and Ersoy, 200 O. For support of the bushing, which has structures isolated by this system, been shown to be very important in the period of vibration (similar to a pendulum) only depends on ge ometry (in this case radius of cur vature) and the gravitational constant, and it does not depend on the mass.Therefore,FPS bear ings are one of the most suitable isolation devices for a system SECTION VIEW with relatively small weight com pared to a high-rise building (such as substation equipments). • Figure 3. Photograph and Cross-Section View of an FPS Isolator Furthermore, through friction, 32 1\ Table 1. FL'{ed and FPS Isolated Results for Several Cases Using the Sylmar (Northridge) Record \9 PGAT."" 0.2500 0.1250 I - 0.2500 0.1250 - PGAReJ.! 0.2313 0.1093 - 0.2637 0.1462 - A, 0.2513 0.1082 - 0.1221 0.1158 - A, 0.4282 0.1155 - 0.1883 0.2008 - As, 10.4819 0.4807 0.3981 0.4484 4 - - As! 10.4267 0.2835 - 0.2722 0.3728 - As. 10.9657 0.6177 - 0.2255 0.2739 - All-! 11.793? 1.3097 - I 0.4328 0.6308 - O, 1 1.5519 1.6678 - I 21.1447 21.5628 - 0, 1 2.8855 I 3.3082 - I 22.50;4 21.1414 - Note: * For the base isolated easel the displacement limit of the FPS bearing is reached, causing impact. the system can provide a high level and ~4 is the top. Note that the of damping. Four 18.64" radius FPS bushing projects inside the trans bearings at the four corners were former modeL That is, it is con used to support the model for the nected to the transformer model isolated case. somewhere between points ~2' The test results for Syhnar record and ~3' Thus, the bottom of the are tabulated inTable 1 for isolated bushing (i.e., point~l) can have a and non-isolated cases. In these response as large as the top of the tables, ~ and ~ show the accel bushing (~,). Dz' and D3 (in rom) eration values (in g) at the bottom are the relative displacement values and the top of the transformer of the transformer model at the model, respectively. ~1 '~Z'~~~4 bottom and top, respectively. represent the accelerations at dif In comparing these numbers a ferent locations along the bushing. point should be noted with regard ~1 is the bottom of the bushing to the base-isolated results for Case 'E;q;erimenta[ anc£.'Jl.nafyticafStUffy of'Base- /so[ation for 'Electric Power 'Equip7t"£nts 33 5. Inspection of the displacement level of acceleration reduction de results for this case indicates that pends on the type of earthquake the displacement capacity of the record used. That is, in addition to bearing has been reached causing acceleration level and nature of the impact.Thus, the results are not use input (e.g., 1-0 vs. 2-D), the ground ful for comparison to the ftxed base motion characteristics affect the situation. However, they do high level of acceleration reductions. light the fact that vertical motion The experimental results can be has a noticeable effect on the re summarized as follow: sponse of FPS isolated structures. • Inertia reductions depend on In this case, vertical motion has peak ground acceleration (PGA) caused reduction in the horizontal and bearing radius. The FPS sys displacements. Thus, for the 3-D tem is more effective in reduc case (Case 7 in Table 6), unlike the ing inertia forces for higher 2-D case, the displacements are PGAs. Furthermore, both inertia within the bearing capacity. reductions and maximum dis The response acceleration maps placements are affected by the for one case is shown in Figure 4. earthquake record used. In this figure, x-axis shows the ac Records with dominant period celeration values normalized with in the vicinity of the isolator respect to PGA. The y-axiS shows period reduce the isolator effec different locations along the height tiveness. of the test specimen ranging from • FPS bearings can provide, on the the top of the shake table to the average, 60% acceleration reduc top of the bushing.As one can see tions within their displacement from this figure, acceleration re limits. This number is with re sponse of the transformer model is spect to the isolation level. For reduced significantly at different a flexible system (as seen from levels throughout its height. The the acceleration maps) accelera tions are different at various lev els, and the effectiveness of the base-isolation is more apparent when one considers the entire Bushing lop picture. For example, there is a Bushing mid. , Trans. top significantly greater reduction _".: fps0250gx0125gy0125gz ..... fps0375gx0250gy0250gz in the bushing acceleration than ... non0250gxOl25gyOl25gz .-,- ...... + ... -\1-.+------+ ... -. - non0375gx0250gy0250gz that of the transformer. , ---"- _.. _.*, .. - _.. _.. _._-- • Coupling of responses in two ! horizontal directions does exist, which is due to dependency of frictional characteristics on total Trans. Bottom velocity. However, the effect tends to diminish for higher PGAs, since Shake Table 0 at higher velocities, frictional con stants are less sensitive to the magnitude ofvelocity. • Figure 4. Acceleration Maps: Triaxial Simulation, FPS Bearings, Sylmar • The vertical component of (Northridge) Record ground motion has an effect on the response of the FPS bear- 34 i:\~~";:;.;:::·,:::-:::;_....,;-,-_.:.. :::. == figs. This effect is expected to former model and the bushing, i.~::':\ be more pronounced for near while the rubber bearings provide iJ!il!!ese~1I"cb field earthquakes (higher PGAs) a horizontal restoring force with efforts ove7 the and for sites where mtering of out sustaining any vertical load. p~st seve7al the motion due to local soil con Detailed information on the design yearsb~ve ditions is possible. of the hybrid-isolation system and reve~ded that experimental results can be found -unde7standing Hybrid Base-Isolation in (Murota arid Feng, 2001). the seismic Peak response accelerations inte?.Rction Results along with acceleration maps over Rmongkey the height of the transformer It is difficult to use normal rub equipments of R model and bushing, with and with ber bearings for seismic isolation s-ubstGtion is out the base isolation, are shown of lightweight structures, such as critical to as a function of PGA in Figures 6 transformers, due to their limited and 7 for the El Centro ground P'1l"Dpc7ly ability to elongate the natural pe motion. Under other ground mo Rssessing tbeir riod of the entire isolated system tion records, the peak responses seismic without buckling. This difficulty show a similar trend. like the FPS pcrfoTr.J1,ance." was alleviated in this study by com bearings, the hybrid system is very bining sliding bearings (to support effective in reducing the accelera the entire weight) with rubber tions, especially in terms of the re bearings (to provide restoring sponse of the bushing top .Without forces). The hybrid system was base isolation, the peak accelera tested under a transformer-bushing tion at the top of the bushing model, as discussed, and the results reached 3.66 g,resulting in an am were compared to the :f!Xed base plification factor of 10.80. situation to investigate its effective On the other hand, for the base ness. isolated case, the peak response ac Four sliding bearings were in celeration at the top of bushing was stalled at the corners of the trans 0.354(g) with an amplification fac former model and two rubber tor equal to 1.05. Note that in these bearings were placed at the middle discussions, including those on the on two opposite sides, as shown in FPS bearings, the PGA referred to Figures 5. The sliding bearings and shown on the corresponding carry the entire weight of the trans- figures is the target PGA.The actual SLB: Sliding Bearing RB: Rubber Bearing 2200 II Figure 5. Layout and Experimental Set-up of the Hybrid Sliding-Rubber Bearings 'E;r:perimenta[ ana ::tnafgtica! Stuig of'Base- Iso[ation for 'E!ectric Power 'Equipments :; 5 tern. It is observed that for the base isolated system, the transformer response is not sensitive to the ~::::: 'b '/"k'«'~~':::>"" grolmd motion. For the fIxed-based case, the interaction between the , : , : transformer (12.5 Hz) and the bush 4 t : r ~ ing was observed. As a result, the ,:,: .: : 3 .: response of the 161kV bushing, ,. : . : -e- Is: xO.25g : with 12.5 Hz frequency, became 2 •: ..... Is: xO.375g , -e- Is: xO.5g larger than that of the 69kV bush Trans. bottom 1.1'" -IC-· Fb: xO.25g ,: -i:J,.. Fb: xO.375g ing, which has a dynamic frequency Platform lL of27 Hz. 00 - 2 4 6 8 10 12 Acc.! PG\ Under triaxial shaking, the trans former-bushing system showed sig nificant difference in response • FJgUfC 6. Acceleration Maps: Uniaxial Simulation, Hybrid Bearings, EI from those under uniaxial and bi Centro Record. axial shaking. The response accel erations at the tops of both 161kV and 69kV bushings for the isolated EI Centro 1940 NS cases were ampHfted, and in some cases (especially for the 69kV bush § 4 ing), the response exceeded that of U 3.5 the fixed-based system. Figure 8 ~ III shows the acceleration response 1/1 3 C 0 maps (ampHftcation factors) for the Co 2.5 1/1 ... Isolated:Transformer worst case. III Top 0:: 2 l!. Fixed Base:Transformer Detailed inspection of the results E 1.5 Top :J indicates that this is due to the ef E ...... 0 Fixed Base:Bushing Top 'x fect of vertical motion on the fric III 6. :e 0.5 -.Do . tion force acting on the sliding 0 I. bearings. Vertical records are gen 0 0.1 0.2 0.3 0.4 0.5 0.6 erally rich in frequency content, PGA(g) reSUlting in high frequency fluctua tions in the frictional force (20-30 Hz), which in turn causes excita • Figure 7. Peak Response Accelerations Under Uniaxial Shaking tion of high frequency modes.Thus, the reason for the 69 kV bushing, or real input acceleration may have which has a fundamental fre been different due to difficulty in quency around 27 Hz, to be more exactly matching the intended affected by triaxial excitation.This PGA. Of course, response param is valuable information that has not eters (such inertia reduction) are be observed and/or investigated calculated with respect to actual before. It also has signifIcant impli acceleration, not the target accel cations for other structures (e.g., eration. response of secondary systems in The hybrid base isolation be a building within the framework of comes more effective as the PGA performance-based design) and it becomes larger (Figure 7), which will be investigated further. is typical of a sliding isolation sys- 36 In summary; the proposed hybrid sliding-rubber bearing isolation sys tem is quite effective in reducing the response acceleration of a Bushing mid 6 transformer-bushing system under Trans. top 5 uniaxial and biaxial earthquake simulator tests. It should be noted 4 that the seismic performance of an 3 actual transformer-bushing system Trans. bottom 2 equipped with the proposed isola tion system will be even better, Platform 1 because the isolation period of an actual transformer will be much 00 6 8 longer !;han that of the transformer bushing model used in the test due to a much heavier weight of the • Figure 8. Acceleration Maps: Triaxial Simulation, Hybrid Bearings, El actual transformer.Among ongoing Centro Record objectives of the project are nu merical studies and modifications During the initial phase of this to the design of the hybrid system study and parallel with the shake to improve its effectiveness under table tests, an extensive analytical triaxial motion. work was conducted. Using SDOF idealization and equilibrium of the forces involved, the differential Analytical Study equations of motion were estab lished and solved using IMSL rou tine IVPAG (Ersoy et al., 2001). SDOFModel Although the focus of this study is Over the past decade or so, there on application of FPS devices to have been several analytical and transformers, it also addresses ad experimental studies on the seis ditional parameters and aspects of mic performance of FPS isolators. response that either have not been These works proved the effective investigated before or have only ness of FPS in reducing inertia been considered on a limited ba forces within the displacement sis.Therefore, some of the :findings limit of the device. Practical appli are general and could have appli cations of the system in the design cations in the design of FPS bear of new structures and the rehabili ings for other structures as well. tation of existing ones, including Among the parameters considered the historic Ninth Circuit U.S. Court are ground motion characteristics, ofAppeals in San Francisco (Mokha, bi-directional motions, the effect of et al., 1996), have taken place. A vertical motion, and isolation ra more recent work by Almazan, et dius. aI., (1998) addresses several other Inertia reduction and the maxi important aspects of modeling and mum displacement of the system response such as constitutive rela were the criteria used in evaluat tionships (small vs.large displace ing the seismic response and the ment), and refinement of the effectiveness ofFPS bearings. Based structural model. on the results of the parametric 'E.;q;erimenta[ and:7I.J1.afyticaf Studj; of'Base- Iso[ation for 'EfectTi.C Power 'Equipmen.ts 37 interconnecting elements. To de 600 velop the knowledge base to achieve this goal, finite element 550 45~ models of typical transformers and ES25 ~ bushings were developed, as dis g SOD SO c: .Q C 475 ...... ! ...... + ...... :\ ... . cussed below, for time history Q) 55 g E 4SO ...... ,i, ...... '0 Q) analysis. ~ 425 .. ...L ...... L .... / 6Or!E ~ 400 .... j .. / .... . 65 ~ o 375 ; , ...... ; Q) 70 .£ Finite Element Model 3SO _ ...... ; ...... 75 325 A power transformer is com 300 +------i----i---+---+----+----+---+-----+----i~i----+--+--+---+ 80 05 08 1.0 1.3 1.5 1.8 2.0 2.3 2.5 2.8 30 3.3 3.5 38 4.0 posed of six parts: transformer tank, Radius of Curvature of Bearing (m) radiators, reservoir, core and coil, oil, and bushings. The transformer • Figure 9. Analytical Displacement-Inertia Reduction Chart for FPS System tank is the main structural compo (input PGA = 1.0 g) nent of power transformers. It has core and coil centrally placed study, charts for inertia reduction within it and the tank is completely and the maximum displacement ftIled by mineral oil. Radiators and were developed. Shown in Figure reservoir are appendages and they 9 is a chart corresponding to are externally attached to the trans ground motion with peak accelera former tank. tion of 1.0g. The challenge in de A finite element model of a sign will be the selection of radius power transformer (55 MVA, 230/ of curvature of the bearing such 135 kV) from a substation in New that there is a balance between the Jersey is shown in Figure 10. The desired inertia reduction and the transformer weighs 1,335 kN (300 displacement limits vis-a-vis bush kips), and the radiators (on the ing interaction with interconnect side) and reservoir (on the top) ing equipments. weigh 120 kN and 40 kN, respec As can be seen from figure 9, the tively. Thus, one can see the possi bearings can provide large inertia bility of torsional response and reductions. For example, for a sys higher demands on the supports in tem with radius of curvature equal light of the relative weights of the to 1.5 m (60" inches) the inertia appendages compared to the reduction is about 60%.That is, the weight of the transformer itself. transformer acceleration will be The finite element package O.4g when the peak ground accel ANSYS was used to develop the fi eration is 1.g. nite element models. The trans For this radius, the system period former tank was modeled by shell is 2.5 sec (i.e., frequency of 0.4 Hz) elements. Braces around the trans regardless of the weight of the former were modeled by offset transformer. However, the associ beam elements. Currently, the core ated large displacement needs to be and coil inside the transformer accommodated. To this end, there were modeled as mass elements. is a need for a simplified 3-D model Radiators and reservoir were mod to investigate the interaction be eled by 3-D solid elements.The con tween the transformer-bushing and tained oil inside the transformer 38 iii Figure 10. Fmite Element Model of a Transformer-Bushing System was modeled as solid with modu Conclusions lus of elasticity equal to the bulk modulus of the fluid. Research efforts over the past The seismic response of bushings several years have revealed that un was dominated by the behavior of derstanding the seismic interac the gaskets between the porcelain tions among key equipments of a units.The common fuilure mode in substation (transformers, bushings, volved movement of the upper por foundation, and interconnecting el celain unit relative to its support ements) is critical to conducting a flange, causing oilleakageTherefore, proper assessment of their seismic the analytical model for the bushings performance.Thus, the thrust of fu uses simple beam elements with ture efforts will be to evaluate seis equivalent density and stiffness to mic response, and propose design represent porcelain units, dome, and and rehabilitation guidelines, based all11Il.inum support. on system performance. This will Gaskets between these elements be achieved through further ex were modeled using nonlinear perimental tests and by developing axial and shear springs. For a fixed a simplified model that can accu transformer, the translational de rately represent critical elements of grees of freedom were removed at a substation in order to investigate the location of the supports. The their interactions in detail through soil-structure interaction may be in a parametric study. The model will vestigated using Winkler founda be developed based on the results tion elements to evaluate the level of ongoing 3-D finite element analy of stresses in the subgrade. Cur ses and the experimental results rently, time history analyses using conducted over the past several various earthquake records are be years. ing conducted. 'E0\Perimentaf amLfJl.nafgtica! Stud!! of'Base-lsofation for 'Electric Power 'Equipments 39 References------ Alma7.an, J. 1., De La Uera, J. C, and Inaudi, J., (1998), "Modeling Aspects of Structures Isolated with the Frictional Pendulum System," Earthquake Engineering and Structural Dynamics, 27, 845-867. ASCE Manuals and Reports on Engineering Practice No. 96, (1999), Guide to Improved Earthquake Performance of Electric Power L~}'stems, Edited by A. J. Schiff, ASCE, Virginia. EERI Newsletter, (1999), ''The Izmit (Kocaeli), Thrkey Earthquake of August 17, 1999," Vol. 33, Number 10, October. Ersoy, S., Saadeghvaziri,M. Ala, Liu, G. Y., and Mau, S.T., (2001), "Analytical and Experimental Evaluation of Friction Pendulum System for Seismic Isolation of Trans formers," Earthquake Spectra, Journal of Earthquake Engineering Research Institute, (tentatively accepted). IEEE, (1998), IEEE Std. 693-1997, Recommended Practices for Seismic Design of Substations, Piscataway, NJ, IEEE Standards Department. Mokha, A., Amln, N., Constantinou, M., and Zayas, V, (1996), "Seismic Isolation Retrofit of Large Historic Building.·;," J Structural Engineering, ASCE, Vol. 122, pp. 298-308. Murota, N., and Feng, M.Q., (2001), "Hybrid Base-Isolation of Bushing-Transformer Systems," Proceedings, 2001 Structures Congress, ASCE, Washington, DC, May 21-23. Saadeghvaziri, M. Ala, and Ersoy, S., (2001), "Evaluation of Seismic Response of Transformers and Effectiveness of FPS Bearings for Base-Isolation," Proceedings, 2001 Structures Congress, ASCE, Washington, DC, May 21-23. 40 Recommended Changes to the AASHTO Specifications for the Seismic Design of Highway Bridges (NCHRP Project 12-49) by. Ian M. Friedland.. Ronald 1. Mayes. and Michel Bruneau Research Objectives The ATC/MCEER Joint Venture is developing new specifications for the seismic design of highway bridges that can be recommended for incorpora tion into theAASHTO IRFD Bridge Design SPecifications. The recommended National Cooperative Highway Research Program ofthe specifications will be performance-based and address state-of-the-art aspects National Research of highway bridge seismic design, induding the latest approaches for repre Council's Transportation sentation of seismic hazard, design and performance criteria,improved analy Research Board sis methods, steel and concrete superstructure and ~Llbstructure design and Applied Technology Council detailing, and foundation design. (ATC)iJfCEERjoint Venture nAugust 1998, a joint venture of the Applied Technology Council (ATC) WiCbTegm ! and MCEER initiated work on a project to develop the next generation Ian iff. Friedland, Applied of seismic design specifications for highway bridges in the United States. Technology Council (EI) The project is sponsored by the American Association of State Highway Christopber Rajah, Applied Technology Council and Transportation Officials (AASHTO) and is being conducted by the Na (Administrative Officer) tional Cooperative Highway Research Program (NCHRP) of the Transpor Ronald Afayes, consultant tation Research Board. NCHRP Project 12-49, "Comprehensive (Technical Director) Specifications for the Seismic Design of Bridges," will result in the devel D0110ldAnt!ers011, CH2M opment of specifications and commentary which are expected to be in HiD, Inc. corporated into the AASHTO LRFD Bridge Design SPecifications. These Michel Bnmeau, Unit-ersity will be supplemented by a series of design examples demonstrating the atBu.ffolo Gregory Fenves, University application of key features of the new specifications. ofCalifornia at Berkeley The recommended specifications and commentary are currently being John K1zlicki, Modjeski and assessed by the AASHTO Highway Subcommittee on Bridges and Struc Masters, Inc. tures. The AASHTO Subcoffimittee is expected to make a decision regard John Manlier, University of ing their adoption in May 2001. During the course of the project, three Canterbury drafts of specifications and commentary will have been produced and re Lee Marsh, BERGER/ABAirJ viewed, prior to completion and submission of the :final draft to AASHTO. Consultants Geoffrey Martin, Uni1:ersity The new specifications are expected to incorporate all current and state ofSouth em California of-the-art practices in highway bridge seismic design, and will be perfor Andrzej No-wok, Uni1:ersity mance-based. They will address the latest approaches for representation ofMichigan of seismic hazard, design and performance criteria, analysis methods, steel Richard V. Nun, consultant and concrete superstructure and substructure design and detailing, foun Maurice POZlleT, Geo17U1tri... dation design, and soil behavior and properties. The specifications are Consultants also intended to address the differences in seismic hazard, soils, and bridge Andrei Reinborn, University construction types found throughout the United States, and therefore are at Buffolv 41 to be national in scope. These new "critical," and "essential" bridges for specifications will be a marked de defining performance levels. parture from the design philosophy, The upper level earthquake con approach, and requirements cur sidered in these provisions is des rently in use in the United States. ignated the Maximum Considered Project Engineering Panel Earthquake (MCE). In general, the Ian G. Buckle, University of MCE ground motions have a prob Nevada at Reno (chair) Technical Summary ability of exceedance of 3% in 75 Serafim Arzoumanidis, years, which is an approximate re Steinman Boynton A number of important changes Gronquist & Birdsall in philosophy and approach will be turn period of 2,500 years. How Mark Capron, Sverdrup Civil, included in these new specifica ever, MCE ground motions are Inc. tions, when compared to the cur bounded deterministically to values I. Po. Lam, Earth Mechanics, rent AASHTO highway bridge lower than 2,500-year ground mo Inc. seismic Qesign provisions. Included tions adjacent to highly active faults. Paul Liles, Georgia in these are the following: It should be noted that the 75 year Department of basis is equivalent to the theoreti Transportation cal design life of a highway bridge, Brian Maroney, California Design Criteria Department of as specified in the AASHTO provi Transportation A performance-based design cri sions. Joseph Nicoletti, URS teria has been included in the new The definitions of service levels Greiner Woodward Clyde specifications and commentary as shown in Table 1 are as follows: Charles Roetier, University of (see Figure 1). Specifically, the Immediate-Full access to nor Washington mal traffic shall be available follow Frieder Seihle, University of project recommends adoption of a California at San Diego dual-level performance criteria, and ing an inspection of the bridge. Theodore Zoli, HNTB two levels of seismic performance Significant Disruption - Limited Corporation objectives, as shown in Table 1. access (reduced lanes, light emer The recommended seismic per gency traffic) maybe possible after formance objectives are perfor shoring; however the bridge may mance-based and each State need to be replaced. determines the desired perfor The damage level definitions mance objective for any particular shown in Table 1 are as follows: bridge. This is a change from cur None - Evidence of movement rentAASHTO definitions of "other," may be present but there is no no table damage. 42 DESIGN I II APPROACHES I I I I Minima! Moderate SignITicant Damage Damage Damage o o I s: T s: I I I I m I z Com_lona! I Seismic I Cow_lona! Conlrcland Energy ConventionaJ :;;! Ductile Design fsoJation _ DUcIlle Repairability Dissipation Ductile :n With low (R<1.5) Design with Oesignwith -< R·f2Clor Merlium High I (R<1.5) R-Factor R-FactOT I 1.S II Figure 1. Design Approaches Minimal - Some visible signs of II Table 1. Perfonnance Objectives damage. Minor inelastic response may occur, but post-earthquake damage is limited to narrow flex ural cracking in concrete and the Rare Earthquake Service Immediate onset of yielding in steeL Perma 3% in 75 years Damage Minimal nent deformations are not apparent, Expe<.i:ed Earthquake Service Immediate 50% in 75 Years Minimal t\1inimal to None and any repairs could be made un Damage der non-emergency conditions with the exception of superstruc develop a set of geotechnical per ture joints. formance objectives that would be Significant - Although there is no similar to those being developed for collapse, permanent offsets may concrete columns. A two-daywork occur and damage consisting of shop was held to review initial draft cracking, reinforcement yield, and proposals and to refine recom major spalling of concrete and steel mended criteria and approaches. yielding and local buckling of steel The consensus of the workshop columns, global and local buckling was that the amount of acceptable of steel braces, and cracking in the foundation and abutment move bridge deck slab at shear studs on ment should be related to geomet the seismic load path is possible. ric and structural constraints by These conditions may require clo bridge type, rather than explicit sure to repair the damage. Partial values on foundation movements or complete replacement may be (see Figure 2). As a result, the rec required in some cases. ommended specifications propose constraints that would implicitly provide foundation design limits for Geom.etric and Structural seismic loads to meet the various Constraints performance objectives. Since this is the first time an attempt has been In the initial phases of this made at developing these con project, an attempt was made to straints, the specified values may :Rgcommewferf Cliailges to tfie JV2IS:J{'IO Specifications for tfie Seismic 'Design of:J{ignrua!f 'Brirfges 43 ~toa.rnmt Research • Approach fill • Approach slabs • The specifications being settlement • Approach fill de~'eloped under this • Bearing failure stabilization 0.083 feet 0.83 feet • Bearing type selection (0.03 meters) (0.2 meters) project draw heavily on the f To avoid results ofmany research vehicle impact studies conducted within Grade Break (2) the MCEER Highway Interior support • Strengthen foundation Use AASHTO None settlement • Bearing type selection "Green Book" Project, and on related Bearing failure • Longer approach slab requirements NSF-sponsored research, Approach slab to estimate ~ settlement allowable especially in the areas of break seismic hazard Alignment Offset Bearing failure _ Bearing type selection representation and .... Shear key failure • Strengthen shear key Abutment foundation • Strengthen foundation Shoulder geotechnical/soils failure Width peiformallce and response. (To avoid f vehicle impact) • Figure 2. Geometric Constraints on Service Level (partial listing, taken from Table C3.l0.1.2, Revised LRFD Design Specifications, Third Draft of Specifications and Commentary) require additional work in the fu lowed for immediate service. In the ture to reftne them. case of a signiftcant disntption, Geometric constraints generally post-earthquake use of the bridge relate to the usability of the bridge is not guaranteed and therefore no by trafflc passing on or under it. geometric constraints would be Therefore, such constraints will usu required to achieve these goals. ally apply to pennanent displace However, because life safety is at the ments that occur as a result of the heart of this performance level, ju earthquake. The ability to repair (or risdictions applying the provisions the desire not to be required to re should consider establishing some pair) such displacements should be geometric displacement limits. considered when establishing dis Structural constraints on displace placement capacities. When imme ments can be based on the require diate service is desired, pennanent ments of a number of stntctural displacements should be small or elements and can result from either non-existent, and should be at lev transient displacements due to els that are within accepted toler grOlmd shaking or pennanent dis ances for normal highway placements reSUlting from ground operations. When limited service movement due to faulting, is acceptable, the geometric con seiSmiCally induced settlements, lat straints may be relaxed. These may eral spreading, and so forth. Struc be governed by the geometry or tural damage to foundation types of vehicles that will be using elements is limited primarily to the bridge after an earthquake, and piles since footings and pile caps by the ability of these vehicles to are usually capacity protected. Al pass through the geometric ob though pile damage can often be stntction. Alternately, a jurisdiction avoided at a reasonable cost when may simply wish to limit displace piles are capacity protected, requir ments to a multiple of those al- ing this for all piles could lead to 44 overly conservative, and thus ex elastic (R = 1.0). This will result pensive,foundation designs. There in a performance that is equiva fore, it may be desirable to establish lent to an elastic (no damage) some lateral displacement limits for design for a 100 year flood, how piles based on a limited amount of ever, is less conservative but structural damage that is unlikely similar to an elastic design for a to compromise the structural integ 100 mph base wind design. rity of the bridge. These limits will • For a rare earthquake (3% prob be based on the type and size of ability of exceedance in 75 the piling, and the nature of the soil years), this design level is recom near the head of the pile. Caltrans mended in order to assure that has attempted to establish such lim a "no collapse" performance cri its for its standard piling based on teria for the MCE is satisfied. physical testing. Other jurisdictions Since seismic loads increase may attempt to do the same, or they much more significantly com may perform more complex analyti pared to wind and flood loads cal studies to establish similar dis as the return period increases, placement, capacity, and stiffness earthquake design needs to con limits. sider longer return period de sign events. A critical earthquake design issue Earthquake Return Periods in preventing collapse of a bridge Current AASHTO specifications is to ensure that deck displace consider a single-level earthquake ments, and hence seat widths for hazard design event, based on a the girders, can accommodate dis 500-year return period. At the time placements from events that have this event was incorporated into occurred and can be realistically the AASHTO specifications, it was expected to reoccur. A retUrn. pe the only event readily available as a riod of 500 years does not come design value for seismic hazard rep close to capturing the force and dis resentation in the United States. placement levels that may have oc The new provisions recommend curred during earthquakes in the a 2,500-year return period, which New Madrid region (1811-1812) provides an equivalent 3% probabil and Charleston, South Carolina ity of exceedance in 75 years, be (1886). A return period of the or used as the upper level design der of 2,500 years is therefore re event.A lower level event with an quired to obtain realistic approximate 100-year return pe displacement levels that may have riod, which would be based on a occurred during these earthquakes. 50% probability of exceedence in It is noted that the MCE ground 75 years, has also been recom motion maps incorporate determin mended. In brief, the reasons for istic bounds on the ground motions these recommended return periods adjacent to highly active faults. include: These bounds are currently appli • For a frequent or expected cable in California, along a portion earthquake (50% probability of of the California-Nevada border, in exceedance in 75 years), this is coastal Oregon and Washington, the design level under which the and in parts of Alaska and Hawaii. bridge should remain essentially These deterministic bounds are 'R,gc01mnerufecf Changes to tlie YlJiSJ-{'IO Specifications for tIie Seismic 'lJesign ofJ-(£gfiruay'Britfges 45 applied so that the ground motions conservatism in the design spectra do not become unreasonably large when compared to the ground mo in comparison to the ground mo tion response spectra beyond 1 sec tions that could be produced by ondperiod. maximum magnitude earthquake Analysis of ground motion data in on the faults. Deterministic bounds dicates that the acceleration spectra are defined as 150% of the median generally decreases with period in estimates of ground motions calcu the long period range as Iff or more lated using appropriate attenuation rapidly. The shape of the long pe relationships, and assuming the oc riod portion of the recommended re currence of a maximum magnitude sponse spectra is less conservative earthquake on the fault. However, than the current spectral shape and they are limited to not less than 1.5 g it is recommended that the shape of for the short-period spectral accelera the spectra decay as 1([ tion plateau and 0.6 g for l.O-second For periods longer than about 3 spectral accelerations. seconds, depending on the seismic environment, use of the Iff relation Design Spectra Shape ship for spectral acceleration may be conservative because the ground The long period portion of the cur motions may be approaching the rent AASHTO acceleration response constant spectral displacement re spectra is governed by the spectmm gion for which spectral accelerations shape and the soil factor, and it de decay as 1!f2. Either the Iff rela cays as 1ff2/3. There was consider tionship may be conservatively used able "massaging" of the factors that or a site-specific study conducted to affect the long period portion of the determine an appropriate long-pe currentAASHTO spectra in order to riod spectral decay. produce a level of approximately 50% Constmction of the design re sponse spectra requires the spectral acceleration value at 0.2 seconds and 1.0 second. The base curve con structed with these values is then Note: Sar is spectral acceleration on modified according to the 1994 rock Type B NEHRP short- (F) and long-period (F) soil site factors, as shown in Fig ure 3. Design Procedures and Soil response Response Modification Factors strum The recommended specifica tions provide for three levels of de sign and analysis procedures in the moderate-to-high seismic zones. 3 This is in addition to the current "no seismic" or low-level minimum requirements, which are defmed as • Figure 3. Response Spectrum COIl~truction Level 0: 46 Level 0 - These are the cur- Iii! Table 2. 1YPical Response Modification Factors rent AASHTO Zone 1 provi :>:: ;~,,; Suhslructure:Hement·, <.! ....." ;...... • ,. P"nnrmance Oiliective; sions, which require <:';>' :<.., ...... ,: ).< "'.;::;.:'.:"::': ..'{: .' '.: ;: :: ·life: SafetY'; : Operational . minimum seat widths and ~:'···.· •• ·..:};"··i'5; •. ·~r(~~:"r::(';;i:·· .: .:,'. ,.;,' V·$D~~· SOAP SDAp· specified design forces for :':::':.' .; '£ .', .;:.' 82~~AFt;; •. ·· 1::0·:· I E Wall Piers-larger dimension 2 3 1 1.5 :fIXed bearings of 10% dead Columns - Single and Multiple 4 6 1.5 2.5 Pile Bents and Drilled Shafts - Vertical load for Zone lA and 2S% 4 6 1.5 1.5 dead load for Zone lB. Piles - above ground Pile Bents and Drilled Shafts - Vertical 1 1.5 1 1 Level 1 - This requires no Piles - 2 diameters below ground Pile Bents and Drilled Shafts - Vertical formal seismic analysis but nJa 2.5 n/a 1.5 Piles - in ground requires the use of capacity Pile Bents with Batter Piles nla 2 nJa 1.5 design principles and mini Seismically Isolated Structures 1.5 1.5 1 1.5 Steel Braced Frame - Ductile mum design details. 3 4.5 1 1.5 Components Level2A - This is a one-step Steel Brace Frame - Nominally Ductile 1.5 2 1 1 design procedure based on an Components analysis method referred to as All Elements for Expected Earthquake 1.3 1.3 0.9 0.9 Connections (superstructure to the "capacity spectrum abutment; joints within superstructure; 0.8 0.8 0.8 0.8 method" and is applicable to column to cap beam; column to very "regular" bridges. This foundation) method has been incorpo- rated in some of the retrofit guidelines for buildings and pro quired if sacrificial elements exist vides the designer with the ability in a bridge). to assess whether or not a bridge Level3 -This is a two-step design designed for all non-seismic loads procedure using an elastic (cracked has sufficient strength and displace section properties) analysis with ment capacity to resist the seismic the multimode method of analysis loads. for the governing design spectra to Level 2B - This is a one-step de perform preliminary sizing of the sign procedure based on an elastic moment capacity of the columns. (cracked section properties) analy Capacity design principles govern sis using either the Uniform Load foundation and column shear de or Multimode method of analysis. sign. A pushover analysis is then The analysis is performed for the performed. The designer can governing design spectra (either change the design forces in the col SO%nS-year or 3%nS- year event) umns provided they are not low and the use of a conservative R-fac ered below those required by the tor. The analysis will determine the SO%nS-year event and that they moment demand at all plastic hinge satisfy the displacement demand. locations in the column. Capacity These provisions attempt to make design principles govern founda the Levell approach applicable to tion and column shear design. If as wide a range of bridges as pos sacrificial elements are part of the sible. Level 2A will be applicable design (Le., shear keys) they must to very "regular" bridges that can be sized to resist the SO%nS-year be idealized as Single-Degree-of forces, and the bridge must be ca Freedom systems, until more devel pable of reSisting the 3%nS-year opmentwork is done to extend the forces without the sacrificial ele method. ments (Le., two analyses are re- 'Rgc01mnerufea CIia11!fe5 to tlie Ji.:JISJf'YO Specifications for tIie Seismic 'Design of:Higfiway:Britfges 47 Base response modification (R are also capacity design require the seismic factors) are significantly more lib ments for detailing the connection design eral and comprehensive in the new to pile and column caps. There are provisions provisions than in currentAASHTO. no additional design requirements being developed These values, as shown in Table 2, for abutments or fotmdations with under this reflect the higher levels of seismic one exception - integral abutments project are demand and analyses required by must be designed for passive pres adopted by the recommended provisions. They sure. The premise for this is based AASHTO, in also require equivalently elastic on a parameter study that was con whole orin analysis for the 50%!75-year event. ducted under the project which part, they will R-factors for all connections require demonstrated that a foundation de essentially elastic behavior, but signed using a factor of safety of 3 become the should not be used where capac will perform well in these lower national ity design principles are used to de seismic zones, assuming that the standard under sign the connections. column connection to the pile cap which all is based on capacity design prin highway Levell, No Seismic Demand ciples. These procedures will be bridges in the permitted for use in sites with liq u.s. are Analysis uefaction potential, but will require designed." The "no seismic demand analy some additional considerations if sis" design procedures are an ex the predominant moment magni tremely important part of the tude, MB, for the site exceeds 6. recommendations because they are The limitations on the applicabil expected to apply as widely as pos ity of the "no seismic demand analy sible in Zone 2. The purpose ofthe sis" provisions started with the provisions is to provide the bridge defmition of a "regular bridge" as designer the ability to design stmc contained in the current AASHTO tures without the need to under seismic design provisions. Addi take dynamic analyses. The bridge tional limits were added based on elements are designed in accor engineering judgment, and these dance with the usual provisions, should be reevaluated in the future. but the primary seismic resisting el The limits as they currently exist ements are specifically detailed for consider (a) regularity, (b) axial load seismic resistance. limits, (c) load path and force shar Capacity design requirements ing,and (d) continuity. Table 3 pro exist for shear reinforcement and vides the deftnition of a "regular" conftning reinforcement at plastic bridge for these provisions. hinge locations in columns. There Span-to-span continuity must be provided by one or more of the fol • Table 3. Parameter Definitions for "Regular" Bridges lowing means: • A superstmcture that is mono lithic with the substmcture • A superstructure seated on bear ings that has transverse restraint • Simply supported girders seated on bearings must be ftxed at one support end, and transversely from span to span, excluding 4 4 3 2 abutments restrained at the other 48 • A superstructure supported on proposed steel seismic provisions isolation bearings that act in all developed as part of this project. directions that accommodate The workshop's participants con displacements sisted of representatives from academia, professional practice, Foundations State DOT's, FHWA, and the steel industry, with experience in seismic Provisions for spread footings, design and steel bridges.The objec driven piles, drilled shafts (or cast tive of the workshop was to iden in-drilled-hole piles), and abutments tify whether the proposed have been significantly improved provisions were sufficiently com over similar provisions in the cur prehensive and answered needs rentAASHTO specifications. These identified in past projects with are, in large part, based on a num workable solutions. ber of advances made through re The proposed seismic provisions search programs sponsored for steel bridges were generally well recently by the Federal HighwayAd received. It was also the consensus ministration, Caltrans, and others. opinion that seismic prOvisions for For abutments, explicit recognition steel structures were desirable, and of sacrificial elements has been pro that there may even be instances vided, including knock-off back where steel substructures may be walls and other similar fuse ele the preferred choice for seismic ments. energy dissipation (particularly when combined with other advan tages such as rapid construction). The final draft of the proposed Steel Bridges The current AASHTO Specifi cations do not have seismic re quirements for steel bridges, RlCSI.ab except for the provision of a con tinuous load path to be identi fied and designed (for strength) by the engineer. Consequently, within the scope of this project a. Eccentrically Braced Frame b. Shear Panel System (SPS) subtask, a comprehensive set of (EBF) Ductile Diaphragms Ductile Diaphragms special detailing requirements for steel components expected to yield and dissipate energy in a stable and ductile manner dur ing earthquakes were developed (including provisions for ductile moment-resisting frame sub structures, concentrically-braced c. Steel Triangular-Plate Added Damping and frame substructures, and end-dia Stiffness (fAD AS) Device Ductile Diagrams phragms for steel girder and truss superstructures). In July 2000, MCEER hosted a III Figure 4. Structural Model- Steel Structures workshop to review a draft of the 1\fcommemfed Cfianges to tfie 52liiS:ll'IO Specifications for tIie Seismic 'lJesign of:lligfiwag 'Bridges 49 to address the issues of soil lique Specifications reflects much of the faction and lateral spreading, espe expert opinions expressed during cially as this relates to the this workshop. It is also stnIctured infrequent but potentially large to allow the use of innovative steel upper-level 2,500-year event. based energy dissipation strategies (as shown in figure 4), provided seismic performance can be experi Summary mentally substantiated. The new specifications for the seismic design of highway bridges Special Considerations that have been developed under Regarding Seismic and NCHRP Project 12-49 are a signifi Geotechnical Hazards cant departure from the procedures and requirements found in current A number of special provisions AASHTO specifications. Work on have been included to address im NCHRP Project 12-49 was com portant considerations in seismic pleted in the spring of 2001, and hazard. Included in these are speci the AASHTO Highway Subcommit fications related to vertical accelera tee on Bridges and StnIctures will tions and their impacts on bridge review the provisions resulting performance, and near fault (and so from the project, with possible called "fling") effects. A significant adoption and publication by amolmt of work is also being done AASHTO in late 2001. 50 Literature Review of the Observed Performance of Seismically Isolated Bridges by George C. Lee, Yasuo Kitane and Ian G. Buckle Research Objectives Despite the use of seismic isolation for the protection of bridges for Federal Highway more than 20 years, the performance of these bridges during strong ground Administration shaking remains to be verified in the field. Numerous analytical and ex perimental studies have been completed in many countries to date dem onstrating the validity of the concept, but full-scale field verification under strong shaking has yet to be obtained.This is partly due to the fact that the world population of these bridges is relatively small, but in addition, only Ian G. Buckle, Professor, Department ajCilJil a small fraction of this population is instrumented with strong motion Engineering, UniL'ersity aj instruments. However, recent earthquakes (even though small-ta-moder Nevada, Reno ate in size) have given some insight into the actual performance of Yasuo Kitane, Graduate seismically isolated bridges, and the experience-database for these bridges Student, Department of is growing. This research focuses on assembling currently available infor Civil, Structural and mation on observed performance of isolated bridges and identifying ar Environmental eas where further research may be necessary to improve the state-of-the-art. Engineering, Unimrsity at Bllffal{} This paper is a summary of a comprehensive literature review recently George C. Lee, Director, undertaken on observed performance in recent earthquakes. Based on l}fultidisciplinary Center collected information, possible future research needs are indicated. for Earthquake Engineering Research, and Samuel P. Capen Professor ofEngineering, University he fundamental concept of seismic isolation is straightforward: to re at Bllffal{} T duce earthquake-induced forces in a structure by lengthening its natu ral period, or by adding damping to the structure, or both. Most isolation systems fall into this last category; i.e., they both lengthen the period and add damping. Although the concept of protecting structures from an earthquake, by introducing seismic isolators, was :frrst proposed almost 100 years ago, it is only recently that seismic isolation has become a practical strategy for earthquake-resistant design (Chopra,1995). In general, seismic isolators have the follOwing four major functions (Skinner, 1993; Kelly, 1997): • To transmit vertical load (e.g., dead load) from one part of a bridge to another, usually from the superstructure to the substructure, while al lowing thermal and other movement effects to occur (Le., to perform the same role as a conventional bearing). 51 • To isolate the part of the bridge Italy, Iceland, Japan, New Zealand, above the isolators by introduc Taiwan, and the U.S. (See, for ex ing flexibility in the horizontal ample, list of Seismically Isolated List ofSeismically Isolated plane, or by limiting the horizon Bridges in the United States at http:/ Bridges in the United States: tal shear force that may be trans /www.eerc.berkeley.edu/prosys/ http://www.eerc.berkeley.edu/ pros)'slusbridges.html mitted to the isolated part. usbridges.html.) Extensive research • To provide sufficient rigidity and development efforts have made under low level loads such as this rapid technology transfer pos wind and traffic loads, and mi sible. Numerous experiments in the nor earthquakes. laboratory and computer analyses • To introduce additional damp have been performed to investigate ing into the bridge system, so the effectiveness ofthe isolation tech that relative displacements nique, and have demonstrated (in across the isolators can be con theory) the advantages of isolation trolled. In some cases, damping design over conventional earthquake is provided directly by the isola resistant design. tors; in others, additional devices However, the number of are installed alongside the isola seismically isolated bridges that tors to provide this additional have experienced actual earth damping, such as viscous fluid quakes is very low, and the number dampers. of instrumented isolated bridges, The most common isolators in that have seen earthquakes, is even use today in the United States fall smaller. As a consequence, there is into one of two categories: elasto very little recorded data on the re meric-based (e.g., lead-rubber bear sponse of these bridges and the ef ings and high-damping rubber fectiveness of isolation remains to bearings), and friction-based (e.g., be conclusively proven in the field, Eradiquake bearings and friction particularly for large earthquakes. pendulum bearings). Nevertheless, collecting and syn In the 1970s, seismic isolation thesizing the performance data that systems began to be implemented is available is a worthy exercise in bridges as aseismic devices. The because it gives insight on field technology spread quickly for the experience as well as indicating next 30 years, and many applica where future research efforts might tions can now be found around the be directed.This paper is the result world and particular in Canada, of such an exercise.A comprehen- "',,",JIOO.&JLLL'_PoL"Y active areas 'Y~~t!c:t"J~~.!m~ bifQ,rnJlati.on Qn the respt>~e brid2~ valuable. Bridges 1 eaLt1iJ~q\Jak.~ and have been. " "'h~i"""""""'.,,""" oflnfornJaatloh "is. ~.~fedtiv(~ne:ss.of the isolation. ·st~:p'll(.1leVlelopjJJlg a databaSe ~",""",,~,,'''7 52 sive literature review has been con ducted on the observed perfor mance of seismically isolated bridges in recent earthquakes, and the results are presented below. Observations are also made and summarized. Observed Seismic Performance II Figure 1. MiyagawaBridge (Kawashima et aI., 1992) celerometers in the bridge super A database, summarizing the actual structure. As shown in Table 1, the performance of seiSmiCally isolated peak ground acceleration was small, bridges in real earthquakes, is grow the LRBs did not reach yield and the hIg gradually. This section summa nonlinear behavior of LRBs was not rizes some ofthis material for bridges significant. The maximum accelera- located in Japan, New Zealand, Tai tion in the deck was about 50% of wan,Thrkey and the U.S. Information that recorded 10 m below the ground is presented country-by-country,in al surface, which illustrates the effec phabetical order. tiveness of the isolation. Spectral analysis showed that vibrations with Japan frequendesinthe 3.0 to 5.0 Hz range, were successfully filtered out by the The Miyagawa bridge,as shown in isolators. However; it was also found Figure 1, was built in 1991 and is a that vibrations with frequencies of 3-span continuous steel plate girder 1.2 Hz were amplified by a factor of bridge with a total length of 108.5 2.0 (by the isolators) when com ill and effective width of 10.5 ill. pared to the ground motion. Nonlin- This bridge was the first in Japan ear behavior of the isolators may to be seismically-isolated. It uses suppreSs this amplification in larger lead-rubber bearings (LRBs) for the earthquakes but the importance of isolators, which are designed to considering the predominant ground yield at a lateral force of 12% of the motion period, when selecting the weight of superstructure. In Japan, isolated period, is illustrated it is common to restrain the trans (Kawashima et al., 1992). verse displacement of bridge isola The Varna-age bridge was built in tors, and therefore many bridges in 1993 and is a 6-span continuous Japan are isolated in the longitudi pre-stressed concrete box girder nal direction only.This 'partial' iso bridge with a total length of 246.3 lation is called'menshin' design, and m and an effective width of 10.5 m the Miyagawa bridge is an example to 13.5 m. The bridge is seismically of this technique. In 1991, the isolated with high damping rubber bridge was subjected to ground bearings, and is again isolated in the motions from an earthquake with longitudinal direction only a magnitude of 4.9 and an epicen (menshin design). Seismic re ter 30 km northeast of the bridge sponses of this bridge were re site. The seismic response of the corded during the 1994 bridge was recorded by three ac- Hokkaido-toho-oki earthquake Literature 'Rgoiew of tIie o6servetf Perfonnallce ofSeismica{fg !so[atea'l3ricfges 53 • Tahle 1. Peak Accelerations and Amplification Factors for Bridges in Japan Ii.··. >jff,;;·. i.C. i'::;.(.':;.j/,··. ' •. J''>; loneltlldiililfJ:l' mTia~j' }·;'>..i;'Yei'tical::·,' Deck 0.0051 g 0.0043 g 0.0027 g Peak Ace. Pier Crest 0.011 g 0.0067 g 0.0025 g Miyagawa U ndergrou nd' 0.012 g 0.0091 g 0.0030 g Bridge' Deck/Underground 0.431 0.483 0.897 Amplification Pier Crest/Underground 0.897 0.742 0.862 Factor Deck/Pier Crest 0.481 0.652 1.04 Deck 0.012 g 0.019 g --- Peak Ace. Pier Crest 0.038 g 0.014 g --- Yama-a?e U ndergrou nd3 0.0093 g 0.0085 Ii 0.047 g Bridge Deck/Underground 1.30 2.28 Amplification --- Pier Crest/Underground 4.05 1.71 Factor --- Deck/Pier Crest 0.320 0.133 --- Deck 0.098 g 0.12 g 0.058 g Peak Ace. Pier Crest 0.11 g 0.13 g 0.038 g Maruki-bashi U nderground4 0.065 g 0.059 g 0.038 g Bridge Deck/Underground 1.51 2.07 1.52 Amplification Pier Crest/Underground 1.69 2.19 0.997 Factor Deck/Pier Crest 0.890 0.947 1.52 Deck 0.19 g ------Pier Crest 0.20 g 0.36g 0.077 g Peak Ace. Footing 0.11 g 0.13 g 0.070 g Matsunohama Underground' 0.15 g 0.14g 0.12 g Bridge' Deck/Underground 1.30 ------Amplification Pier Crest/Underground 1.39 2.64 0.655 Factor Footing/Underground 0.717 0.933 0.595 Deck/Pier Crest 0.940 ------ NOTES 1. Transverse movement is restrained in these bridges 2. 10m below the ground surface 3. 5 m below the ground surface 4. 6 m below the ground surface 5. 1 m below the ground surface with magnitude 8.1, whose epicen tions.1t is a 3-span continuous pre ter was about 1,000 km away from stressed concrete box girder bridge the bridge. Recorded peak accel with a total length of 92.8 m and erations are shown in Table 1. Judg an effective width of 11.0 m. The ing from obtained acceleration bridge was excited by ground mo response records, none of the seis tions and its seismic response was mic isolators were exercised in recorded during the 1994 Sanriku their nonlinear range. Peak deck ac haruka-oki earthquake with magni celeration decreased to one-third of tude of 7.5. The epicenter was that at the pier top, in the longitu about 190 km from the bridge site. dinal direction, while it increased Recorded peak accelerations are by 30% in the transverse direction shown in Table 1. The peak deck (restrained direction). Spectral acceleration in the longitudinal di analysis of the recorded responses rection decreased by 11 % from that showed that the vibration compo at the pier top, due to the seismic nent, with a frequency equal to the isolators, and by 5% in the trans fundamental frequency of the pier, verse direction. The peak deck ac was not transmitted to the deck celerations were 1.7 times and 2.2 through the seismic isolators times of the peak ground accelera (PWRI, 1995; Unjoh, 1996). tions in the longitudinal and trans The Maruki-bashi bridge was built verse directions, respectively. in 1992 as a seismically isolated Judging from the recorded accel bridge with lead-rubber bearings of eration response, it can be con a circular shape. It is isolated in the cluded that seismic isolators longitudinal and transverse direc- remained elastic and did not yield. Therefore, the effectiveness of seis- 54 mic isolation is not very apparent responding to the frequency of the from the recorded data. Spectral pier (1.0 Hz), was not transmitted sting of analysis of the recorded data shows from the pier to the deck by the isolated btridge amplification factors (ground to isolators. These observations infer models at full deck) with two peaks at 2.5 Hz and that the isolation system func scale is now 16.7 Hz in the longitudinal direc tioned as expected (pWRl, 1995; possible due to tion (2.5 Hz is the fundamental fre Naganuma et al., 1996; Fujino et al., m-nltiple shake quency of the entire bridge, and 1997; Kouno et al., 1997). table facilities 16.7 Hz is the frequency of the with bin.xial bridge pier itself). Moreover, it was New Zealand motions un-de,. also observed that the vibration the NSF-NKES component with the frequency The Te Teko (Rangitaiki River) in:itititi'27e~" equal to the fundamental fre bridge was built in 1983 and is a 5- quency of the pier, was filter~d out span pre-stressed concrete U-beam between the pier top and the deck bridge with a ,total length of 103 by the seismic isolators, in the lon m. The bridge is seismically isolated gitudinal direction (pWRI, 1995). by lead-rubber bearings at each pier The Matsunohama bridge was sub and plain rubber bearings at each ject to the 1995 Hyogoken-nanbu abutment. It was subject to the (Kobe) earthquake of magnitude 1987 Edgecumbe (Bay of Plenty) 7.2 and epicenttal distance 35 km earthquake with magnitude of 6.3 from the bridge site. The bridge at a distance of 15 km from the was built in 1991, and is a curved bridge. The bridge was not instru 4-span continuous steel box girder mented but the peak ground accel bridge with a total length of 211.5 eration, recorded about 11 km m, a maximum width of 21.94 m, south of the bridge, was 0.33g. One and a radius of curvature of 560 m. of two bearings on the abutment It is a seismically isolated bridge but was dislocated due to a construc only in the longitudinal direction tion defidency (inadequate fasten (menshin design). Both ends ofthe ing detail for the isolator) and as a bridge are supported by pivot consequence, the bridge sustained roller bearings, and two lead-rub minor damage. The superstructure ber bearings are installed on each suffered a small permanent dis of the three middle piers. Re placement, a plastic binge began to corded peak accelerations are form in one pier that spalIed a small shown in Table 1. The peak accel amount of cover concrete, and a eration of the deck was a little sacrifidal knock-off device was smaller than that of the pier top in pushed about 80 to 100 rom from the longitudinal direction, in which its original position. Based on the seismic isolation was effective. fact that the bridge suffered only Judging from the displacement re minor damage despite the severe sponse obtained by integrating the ground shaking, it can be said that acceleration records, the isolators the bridge performed well, and that reached yield. The maximum dis it may not have been damaged at placement was about twice the all if the construction defidendes yield displacement, and about 8% had been avoided by better detail of the design displacement for the ing (Mckay, 1990; Robinson, 1992; isolators. Spectral analysis showed DIS, 1996). that the vibration component, cor- Literature !Rg;view of tfi.e 06se17lea Perfonr.ance ofSeismica[{g Isofatecf 'Britfges 55 • Figures 2 and 3. Bai-Ho Bridge (top) and Isolation SYl>1em of the Bai-Ho Bridge (bottom) Taiwan lated in the transverse direction. Construction of the bridge was The Bai-Ho bridge, as shown in nearly completed but it was not in figure 2, is a 3-span continuous non service at the time of the 1021 prismatic prestressed concrete (1999) Gia-I earthquake of magni girder bridge with a total length of tude of 6.0. Eleven accelerometers 145 m and a maximum width of had been mounted in the bridge 16.1 m. The bridge is seismically prior to the earthquake and re isolated with two lead-mbber bear corded peak accelerations and am ings installed on each pier (see fig plification factors are shown in ure 3) and two PTFE coated rubber Table 2. No damage was suffered. bearings at each abutment. Shear The peak acceleration in the deck keys and specially designed steel in the longitudinal direction was rods are installed on both abut slightly higher than that in the ments to restrict the transverse foundation, while the peak accel movement of the superstmcture. eration in the transverse direction The bridge is therefore partially iso- was 2.5 times that of the founda tion. Significant elongation of the • Table 2. Peak Accelerations and Amplification Factors in the Bai-Ho Bridge stmcture period was observed in the longitudinal direction due to the nonlinear behavior of the LRBs. The bridge performed very well in this direction. On the other hand, two dominant vibration modes were observed in the transverse di Pier Top/Foundation 0.991 1.10 rection and the peak transverse re Deck/Pier T 1.19 2.26 sponse was governed by its second 56 mode. It is clear that the transverse restraint provided at the abutments adversely affected the response in this direction (compared to that in the longitudinal direction). This bridge also experienced the Chi Chi earthquake on September 21, 1999, but no records were obtained at the time (Chang et al., 1999). Thrkey The Holu viaduct on the Trans European Motorway (see figure 4) in the Kaynasli valley was still un II Figure 4. Bolli Vmduct on the Trans European Motorway (Xiao et aL, 1999) der construction when the Duzce earthquake CM=7.2) occurred in ANATOUAN MOTORWAY - TURKEY November 1999 . Many of the struc GUMUSOVA - GEREDE tural elements were complete at VIADUCT.! TYPICAL MODULE the time, but guardrails and paving F.sed,... remained to be installed.The bridge , S ". P is composed of two parallel viaduct * structures with pre-stressed con crete hollow T-girders and 58 monolithic column footings. It has a total length of 2,313 m and a width of 17.5 m.A typical module of the bridge (as shown in figure 5) is a lO-span viaduct with a maxi mum span length of 39.6 m and a maximum pier height of 49 m.The PTFE bearing and the crescent III Figure 5. Typical Module of the Bolu Viaduct (Xiao et al., 1999) moon-type steel energy dissipation deVJice are shown in Figures 6 and site have been estimated as over Ig 7, respectively. The bridge is (the strong motion station in equipped with fiat stainless steel Duzce, near the epicenter and PTFE sliding bearings, a multidirec about 5 km to the west of the via tional crescent moon-type steel duct, recorded accelerations in ex energy dissipation device at the cess 1.0g.) central support ofthe 100span unit, The following observations were viscous connecting devices at made by Xiao and Yaprak (1999). other supports, steel strands re Groundruprure was observed near strainers at the expansion jOints be the south side of the bridge, and tween the 10-span units, and the fault crossed under the bridge bi-directional steel energy dissipa near Pier 46. Soil around the foot tion devices at the two abutments. ing of the pier was disturbed sig The bridge was not instrumented nificantlyand the pier appeared to at the time of the earthquake, but have rotated about 12 degrees peak ground accelerations at the about its vertical axis. Damage to Literature !I(g:view of HIe 06servei Per/onnance ofSeismica[[g Isoratea 'BritftJes ; 7 0000 .!. ( { ~ .... - r* •...... fe ., :- L •• _ 0- •• ---" I '" • Figure 6. PTFE Bearing (Xiao et aI., 1999) • Hgure 7. Energy Dissipation Device(Xiao et aI., 1999) the columns of the viaduct was lim their concrete pedestals. Cracks ited to fine horizontal flexural occurred at the ends of many gird cracks in the lower portion of sev ers. Severe concrete spalling and eral piers. Most of the structural crushing was observed for at least damage was sustained by the super one pre-stressed concrete girder. structure and the abutments. Both The south side wing walls of the superstructures (of the two east abutments of the two bridges bridges) had a westward perma sustained severe damage due to nent displacement relative to the pounding by the girder ends A per piers with offsets at all girder ends. manent southward transverse off Some of the outer-most pre set of about 250 mm was observed stressed concrete girders slid off at the ends of both the south and north bridges. Concrete crushing and inclination of the abutments was observed at the west end of the viaduct, caused by pounding between the girders and abutments (Xiao et aI, 1999). Stainless steel plates, bearing ma sonry and anchor plates, and PTFE bearings fell from their supports at almost all piers. Judging from the scratch signs on the stainless steel plates, these bearings were prob ably dislodged at a very early stage in the earthquake (see figure 8). This is consistent with the expec- • Figure 8. Bearing Failure of the Bolu Viaduct (Xiao et aI., 1999) 58 tation that the bridge encountered ground acceleration at the base of a near-fault or on-fault pulse-type the bridge columns was 0.090 gin motion during the earthquake, and the longitudinal direction and that it experienced excessive dis 0.050 g in the transverse direction. placement. Energy dissipation de Significant amplification occurred vices at the east abutment at these low level accelerations in sustained distortion of the steel the steel superstructure of the yielding dements and rupture of bridge, due to the lock-up action the anchor bolts of the devices. No of the abutments, which prevented span collapsed despite experienc the isolators from acting. No dam ing a level of ground shaking be age, however, was sustained (DIS, yond the original design level 1996). during this earthquake. In addition, The Eel River bridge was retrofit this viaduct survived the previous ted with lead-rubber bearings in earthquake (August 17) without 1987. Two truss spans of the bridge damage. The magnitude of the Au were seismically isolated, and each gust earthquake was 7.4, and the span has a span length of 90.0 m peak ground accelerations were (Buckle et al., 1987). The bridge estimated as 0.39 g (longitudinal), was subjected to earthquake 0.31 g (transverse), and 0.50 g(ver ground motions during the 1992 tical) (Xiao et al, 1999). Cape Mendocino (petrolia) earth quake, whose magnitude was 7.0 and epicenter was located 22 km United States from the bridge. The bridge was The Sierra Point overpass, U.S. not instrumented but peak ground 101, is a 10-span simply supported accelerations obtained at the composite girder bridge with a Painter Street Overpass, a similar skew angle of 59 degrees at the distance from the epicenter, were north abutment and 72 degrees at 0.55 g in the longitudinal direction, the south abutment. Its total length and 0.39 g in the transverse direc and width are 184.8 m and 35.1 m, tion. As a consequence, it is be respectively. The bridge was con lieved that this is the strongest structed in 1956, and retrofitted ground motion to have been felt by with lead-rubber bearings in 1985. a seiSmically isolated structure in It was the first bridge to be retro the United States. Judging from the fitted in the U.S. using isolation and displacement of the guardrail at avoided the need to jacket the non tached to the truss span, the maxi ductile columns and strengthen the mum relative displacement footings. However, the abutments between the isolated trusses and were not modified for the larger the abutment appear to have been clearances required for an isolation of the order of 200 mm (longitudi retrofit. They are therefore ex nally) and 100 mm (transversely). pected to engage during low-to The bridge came to rest in its origi moderate shaking, and to be nal position and structural damage damaged in strong shaking. was light and limited to concrete The seismic response of this spaUing at joints (DIS, 1992; DIS, bridge was recorded by five accel 1996). erometers during the 1989 Loma Prieta earthquake. The peak Literature 2?g.vie'lIJ of the 06served Perfonn.ance ofSeismicalfg Isofated 'Bridges 59 • Figures 9 and 10. Thjorsa Bridge in the South Iceland Lowland<;(top left) and close-up of the isolation bearings-both ends of the truss tipped away from the river by 2 cm (bottom right) Observations tively small and the isolators have either not engaged or have • Seismically isolated bridges have remained in their elastic (and been researched and investi stift) range. Recorded perfor gated by academia and engi mance of an isolated bridge neers for many years. Due to this during strong ground motion is extensive effort, seismic isola not yet available and thus the tion design has become a effectiveness of seismic isola practical option for earthquake tion, especially during a large resistant design. earthquake, remains to be • Results from numerous com proven in the field. puter simulations and shake • An exception to the previous table experiments have shown item might be the performance the advantages of seismically iso of several isolated bridges dur lated bridges compared to non ing the June 17 and 21,2000, isolated bridges. However, few earthquakes in Iceland, which isolated bridges have experi measured 6.6 and 6.5 on the enced actual earthquakes and Richter scale, respectively. These been able to demonstrate their bridges include the Thj6rsa effectiveness, especially under bridge (see Figures 9 and 10), the strong ground shaking. Further Sog River bridge, the St6ra Laxa more, the majority of bridges River bridge, and the Oseyrar that have been subject to earth River bridge (Higgins, 2000). quakes, have not been instru Lead-rubber bearings are in mented and performance must stalled in all four bridges, which be inferred or extrapolated from were subjected to large ground nearby sites. Also, in the major motions during these earth ity of cases, the earthquakes that quakes. The peak ground accel have been captured are rela- eration was 0.84g at one of the 60 bridge sites, but no significant to this concern.An urgent need damage was found. Recorded is a rigorous investigation of the responses are currently being Bolu VIaduct to determine the analyzed at the Earthquake En cause for the damage, and iden gineering Research Center, Uni tify changes, if any, that should versity of Iceland CEERCUI). be made to U.S. design and Results are expected later this manufacturing criteria for isola year (2001). tors. • Our experience-database of iso • It is clear from several of the lated bridges is very sparse. above case studies that small Many more seismically isolated earthquakes rarely yield the iso bridges need to be instrumented lators, which then perform in with strong motion arrays, in their elastic and stiff range. Ac order to address this urgent cordingly, the observed amplifi need and increase the likelihood cation factors are higher than that field verification will be would be expected for an iso available in the near future. lated bridge and this highlights • In the absence of field data, test the need for a class of isolators ing of isolated bridge models at that have the 'intelligence' to full scale, in structural testing optimize their properties ac laboratories, could be under cording to demand. taken. Such work has not been • Various types ofseismic isolators previously possible on a mean have been proposed, developed, ingful scale, but the current de and tested to date. However, velopment of multiple shake only a few of these have been table facilities with biaxial mo- . implemented in the field due to tions, under the NSF-NEES initia concerns about performance tive, makes this option a under extreme events and their practical reality. durability and maintenance. • The performance of isolated Comparative performance of bridges in the near field of ac new devices, using either elas tive faults has been a concern tomeric isolators or friction pen since the Northridge earthquake dulum devices as benchmarks, in 1994. The damage sustained should be undertaken to con by the Bolu VIaduct inTurkey, de tinue to advance the state-of-the scribed above, gives substance art. References Buckle, LG. and Mayes, R.L., (1987), "Seismic Isolation of Bridge Structures in the United States of America;~ Proceedings of the Third US. Japan Workshop on Bridge Engineering, Public Works Research Institute, Ministry of Construction, Tsukuba, Japan, pp. 379-401. Chang, K. c., Wei, S. S., Chen, H. w., and Tsai, M. H., (1999), Field Testing and Analysis of a Seismically Isolated Bridge, Report No. CEER R88-07, Center for Earthquake Engineering Research, National Taiwan University, Taipei, ROC (in Chinese). Literature 1(eview of tlie Obse17Jed Perfonnance ofSeismicaffg Isofatelf 'Bridges 61 References (COll't) ------ Chopra, Anil K., (1995), Dynamics of Structures: Theory and Applications to Earthquake Engineering, Prentice Hall, NJ. Dynamic Isolation Systems, (1992), Petformance of the Eel River Bridge in the Petrolia Earthquakes of April 25-26, 1992, Dynamic Isolation Systems, Inc., CA. Dynamic Isolation Systems, (1996), Earthquake Petformance of Structures on Lead Rubber Isolation Systems, Dynamic Isolation Systems, Inc., CA. Fujino, Y., Yoshida, J., and Abe, M., (1997), "Investigation on Seismic Response of Menshin Bridge Based on Measured Records," Proceedings of the 24th Symposium on Earthquake Engineering Research, pp. 325-328 (in Japanese). Kawashima, K., Nagashima, H., Masumoto, S., and Hara, K., (1992), "Response Analysis Buckle, 1. G. and Mayes, R. 1., (1987), "Seismic Isolation of Bridge Structures in the United States of Anlerica," Proceedings of the Third U.S.-Japan Workshop on Bridge Engineering, Public Works Research Institute, Ministry of Construction, Tsukuba, Japan, pp. 379-401. Kelly, Janles M., (1997), Earthquake-Resistant Design With Rubber, Second Edition, Springer-Verlag, UK. Kouno, T. and Kawashima, K., (1997), "Vibration Characteristics of Menshin Bridge Based on Measured Records," Proceedings of the 24th Symposium on Earthquake Engineering Research, pp. 329-332 (in Japanese). Mckay, G.R., Chapman, H.E., and Kirkcaldie, D.K., (1990), "Seismic Isolation: New Zealand Applications," Earthquake Spectra, Vol. 6, No.2, pp. 203-222. Naganuma, T., Horie, Y., Kobayashi, H., and Sasaki, N., (1996), "The Analysis of Seismic Response of the Bridge with Menshin Bearings During the Hyogo-ken Nanbu Earthquake," Proceedings of the Fourth U.S.Japan Workshop on Earthquake Protective Systems for Bridges,Japan, December 9 and 10, 1996, Technical Memorandum of PWRI, No. 3480, Public Works Research Institute, Japan, pp. 259-265. Public Works Research Institute (1995), Study on Seismic Response Characteristics of Menshin Bridges Based on Measured Records, Technical Memorandum of PWRI, No. 3383, Public Works Research Institute, Japan. Robinson, W.H., (1992), "Seismic Isolation of Bridges in New Zealand," Proceedings of the Second U5.Japan Workshop on Earthquake Protective Systetns for Bridges, Japan, December 7 and 8, 1992, Technical Memorandum of PWRI, No. 3196, Public Works Research Institute, Japan. Skinner, Robert Ivan, Robinson, William H., and McVerry, Graeme H. (1993), An Introduction to Seismic Isolation, John Wiley & Sons Ltd, England. Unjoh, S., (1996), "Analysis of the Seismic Response of Yam a-age Bridge," Proceedings of the Fourth US. Japan Workshop on Earthquake Protective Systems for Bridges, Japan, December 9 and 10,1996, Technical Memorandum of PWRI, No. 3480, Public Works Research Institute, Japan, pp. 149-163. Xiao, Yan and Yaprak, Tolga Taskin, (1999), Performance of a 2.3 km Long Modern Viaduct Bridge During the November 12, 1999 Duzce, Turkey, Earthquake, U.S.C Structural Engineering Research Report, Report No. U.S.C-SERP 99/01, Department of Civil Engineering, University of Southern California, Los Angeles, California. 62 A Risk-Based Methodology for Assessing the Seismic Performance of Highway Systems by Stuart D. Werner Research Objectives Federal Highway The objective of this research is to develop, apply, program, and dissemi Administration nate a practical and technically sound methodology for seismic risk analy sis (SRA) of highway-roadway systems. The methodology's risk-based framework uses models for seismology and geology, engineering (struc tural, geotechnical, and transportation), repair and reconstruction, system anaiysis,and economics to estimate system-Wide direct losses and indirect Craig E. Taylor, President, losses due to reduced traffic flows and increased travel times caused by Natural Hazards earthquake damage to the highway system. Results from this methodol iJfanagement Inc. ogy also show how this damage can affect access to facilities critical to James E. ~Ioore ill, Associate Professor, emergency response and recovery. Departments ofCivil Engineering and Public Policy and frimiagement, ast experience has shown that effects of earthquake damage to high Universit;' ofSouthern Pway components (e.g., bridges, roadways, tunnels, etc.) may not only California include life safety risks and post-earthquake costs for repair of the com Sungbin Cho, Research ponents. Rather, such damage can also disrupt traffic flows and this, in Associate, School ofPolicy, turn, can impact the economic recovery of the region as well as post Planning, and Development, Unil/ersity of earthquake emergency response and reconstruction operations. Further Solttbem California more, the extent of these impacts will depend not only on the seismic Jon s. Walton, Vice response characteristics of the individual components, but also on the President, MacFarland characteristics of the highway system that contains these components. Walton Enterprises System characteristics that will affect post-earthquake traffic flows include: Stuart D. Werner, Principal, (a) the highway network configuration; (b) locations, redundancies, and Seismic Systems and traffic capacities and volumes of the system's links between key origins Engineering Consultants and destinations; and (c) component locations within these links. From this, it is evident that earthquake damage to certain components (e.g., those along important and non-redundant links within the system) will have a greater impact on the system performance (e.g., post-earth quake traffic flows) than will other components. Unfortunately, such sys tem issues are typically ignored when specifying seismic performance requirements and design/strengthening criteria for new and existing com ponents; i.e., each component is usually treated as an individual entity only, without regard to how its damage may impact highway system per- 63 formance. Furthermore, current tion are contained in the report by criteria for prioritizing bridges for Werner et al. (2000). seismic retrofit represent the im • Ian G. Buckle, University portance of the bridge as a traffic ofNevada, Reno carrying entity only by using aver Methodology • Ian Friedland, Applied age daily traffic count, detour Technology Council Description length, and route type as param • John Mander, University' The SRA methodology (Figure 1) eters in the prioritization process. ofCanterbury, New can be carried out for any number Zealand (formerly ofthe No attempt is made to account for of scenario earthquakes and simu University at Buffalo) the systemic effects associated with lations, in which a "simulation" is • Masanobu Shinozuka, the loss of a given bridge, or for the defined as a complete set of system University ofSouthern combination of effects associated SRA results for one particular set California with the loss of other bridges in the of input parameters and model un highway system. certainty parameters. The model To address these issues, current Other Contributors and input parameters for one simu and recent highway research • Howard Hwang, Cehter lation may differ from those for for Earthquake Research projects conducted by MCEER and other simulations because of ran and Information, funded by the Federal Highway dom and systematic uncertainties. University ofMemphis Administration (FHWA) have in For each earthquake and simula • John Jernigan, Ellers, cluded tasks to develop a SRA Oakley, Chester and Rike, tion, this multidisciplinary process methodology for highway systems. Inc., MemphiS, Tennessee uses geoseismic, engineering This paper describes this method • Arch Johnson, University (geotechnical, structural, and trans ology, presents results from a dem ofMemphis portation), network, and economic • W.D. Liu, Parsons onstration application of the models to estimate: (a) earthquake Brinckerhoff, Inc., methodology to the highway sys effects on system-wide traffic flows formerly with Imbsen & tem in Shelby County, Tennessee, and corresponding travel times, Associates, Inc., and summarizes plans for the fur Sacramento, California paths, and distances; (b) economic ther development, application, pro • David Perkins, Arthur impacts of highway system damage gramming, and dissemination of the Frankel and Eugene (e.g., repair costs and costs of travel methodology. Further details on (Buddy) Schweig, U.S. time delays); and (c) post-earth- Geol.ogical Survey this methodology and its applica- • Maurice Power, Geomatrix Consultants, Inc., San Francisco • Sarah H. Sun, Shelby County Office ofPlanning This and Devel.opment SRA methodology will provide cost and risk in • Edward Wasserman, formation that will facilitate more rational evaluation Tennessee Department of of alternative seismic risk reduction strategies by deci Transportation sion maket'S from government and tran.sportation agen • T. Leslie Youd, Brigham cies involved with improvement and upgrade of the Young University highway-roadway infrastructure~ emergency response planning, and transportation planning.. Such strategies can include prioritization and seismic strengthening measures for existing bridges and other components, establishment of design criteria for new bridges and other components, construction of additional roadways to expand system redundancy, and post-eariliquake traf fic management planning. 64 quake traffic flows to and from im seismic hazards, bridge vulnerabili portant locations in the region. Key ties, and transportation network to this process is a GIS data base analysis have been developed. that contains four modules with These developments, which are data and models that characterize summarized below; are now incor the system, seismic hazards, com porated into a new pre-beta soft ponent vulnerabilities, and eco ware package named REDARS 1.0 nomic impacts of highway system (Risks due to Earthquake DAmage damage (Figure 1b). These are in to ,Roadway Systems). corporated into the four-step SRA procedure shown in Figure 1a. This SRA methodology has sev eral desirable features. First, it has a GIS database, to enhance data management, analysis efficiency, Initialization of Analysis and display of analysis results. Second, the GIS database is Next Earthquake ! modular, to facilitate the incorpo System Analysis for Scenario Earthquake E", H GIS Databasa I ration of improved data and mod and Simulation n(m) els from future research efforts. Next Simulation Third, the procedure can develop ! aggregate SRA results that are ei Incrementation of ther deterministic (consisting of Simulations and Scenario Earthquakes a single simulation for one or a few scenario earthquakes) or + probabilistic (consisting of many I Aggregate System Analysis Results simulations and scenario earth quakes). This range of results fa cilitates the usefulness of SRA for (a) Overall Four-Step Procedure a variety of applications (e.g., seismic retrofit prioritization and Transportation Cost Module System Module Models criteria, emergency response Network Inventory Effects ofTrave! Time Traffic Data Increases planning, planning of system ex O-DZones Data pansions or enhancements, etc.). Trip Tab!es Vehicle TypesIOccupancy Network Analysis Models Unit Costs Finally, the procedure uses rapid engineering and network analy I Steps 1-4 of I. sis procedures, to enhance its SRA Procedure ft· possible future use as a real-time Component Module Data predictor of component damage Hazards Module Structural states, system states and traffic Seismo-Tectonics Repair Coste Topography Traffic Slatas impacts after an actual earth Soil Conditions Models Ground Motion Attenuation loss quake. Geologic Hazard Models Functionality Model Uncartainties Uncarlainties Recent Developments (b) GIS Database Improved procedures for char acterizing scenario earthquakes, • Figure 1. Risk-Based Methodology for Assessing Seismic Performance of Highway Systems 5'issessing the Seismic Perfonnance ofJ-figfi'llJaq Sqstems 65 Scenario Earthquakes Idriss type methods to assess liq ~toOWnmt uefaction potential during each Resemv:b SRA of a highway system with earthquake and (c) for those sites • Seismic Vulnerability of spatially dispersed components re with a potential for liquefaction Existing Highway quires use of scenario earthquakes during the given earthquake, esti Constrnction (Project 106) to evaluate the simultaneous effects mation of lateral spread displace of individual earthquakes on com ment and vertical settlement using • Seismic Vtdnerability ofthe ponents at diverse locations (in National Highway System methods byYoud and byTokimatsu (TFA-21 Project) cluding systemic consequences of and Seed (1987) respectively. damages). Earthquake models now being incorporated into REDARS are adaptations of work by Frankel Component Models et al. (1996) which was developed Component models for highway under the United States Geological system SRA develop traffic state fra Survey (USGS) National Hazard gility curves, which estimate the Mapping Program. Frankel et al. probability of a given traffic state models for the Central (CUS) are (Le., open lanes at various times summarized later in this paper. after the earthquake) as a function Adaptation of models for California of the level of ground shaking or (which also builds on work by the permanent ground displacement at California Division of Mines & Ge the component site. Thus far, this ology) is now lmderway. All adap research has focused on develop tations feature a "walk-through" ing such models for bridges only. analysis, which is a natural way to These models estimate the bridge's assess system loss distributions and damage state (damage types, loca their variability over time. tions, and extents) under a given level of grolmd shaking or displace Seismic Hazards ment, and then obtain correspond ing traffic states by using The ground motion models for expert-opinion damage-repair mod the SRA procedure include rock els. The SRA methodology now in motion attenuation characteristics cludes three options for modeling representative of the region where damage states of bridges due to the system is located, as well as am ground shaking: (a) an elastic ca plification of rock motions due to pacity-demand approach by local soil conditions. For the Cen Jernigan (1998); (b) a simplified tral United States, the Hwang and but rational mechanics-based Lin (1997) rock motion attenuation method by Dutta and Mander relationships and soil amplification (1998) that develops rapid esti factors for NEHRP site classifica mates of damage states based on tions meet these requirements. bridge-specific input parameters Models appropriate to other re inferred from the FHWA National gions of the country are now be Bridge Inventory database; and ( c) ing incorporated. Liquefaction user-specified fragility curves, hazard models are based on work which can be developed for any by Youd (1998), and include: (a) bridge in the system, but are most geologic screening to eliminate appropriate for complex or un sites with a low potential for lique usual bridges. In addition, the SRA faction; (b) use of modified Seed- methodology includes a first-order 66 model for estimating bridge dam ! J I age states due to permanent • erics~ V ground displacement. /\./ Road Ndwcm V /'VW*-< i 1+ Transportation Network ~~~~ Analysis The SRA procedure contains two transportation network analysis . methods. For deterministic SRA for a limited number of scenario earth quakes and simulations, a User Equilibrium (DE) method is used. This is an exact mathematical solu tion to an idealized model of user III Figure 2. Shelby County Tennessee Highway System Model (7,807 Links and 15,604 Nodes) behavior, which assumes that all us ers follow routes that minimize their travel times. For probabilis tic SRA involving many earth quakes and simulations, a new Oriif,n Cestlnatlcn Zones m Slanc!ard Zaflf! Associative Memory (AM) transpor fEJKe>JZone N tation network analysis procedure A is used. This method provides rapid estimation of network flows, rep resents the latest well-developed technology foc estimating traffic flows, is GIS compatible, and uses transportation system input data that are typically available from Metropolitan Planning Organiza tions. It is derived from the artifi cial intelligence field, and provides rapid and dependable estimates of Ii Figure 3. Origin-Destination Zones in Shelby County, flows in congested networks for Tennessee given changes in link configuration due to earthquake damage (Moore is located in the southwestern coc et ai., 1997). ner of Tennessee, just east of the Mississippi Rivec. Its highway-road way system contains a beltway of Demonstration highways that surrounds the city of Application Memphis, two major crossings of the Mississippi River, and major roadways that extend from the cen System Description ter of Memphis to the north, south, and east (Figure 2). Traffic demands The SRA methodology was used on the system are modeled by trip in a demonstration application to tables that define the number of the highway system in Shelby trips between all of the origin-des- County,Tennessee. Shelby County Jfssessil1£! tfie Seismic Performance of:;{igfiway Su,stems 67 Ground MotiOn (g) Permanant Ground • 0.13 -0.16 Displacement (m) • 0.16 -0.19 0.00 -0.15 0.19 -0.22 4 0.15·0.30 • 0.30-0.45 • 0.22-0.25 • 0.45-0.80 Soli Type Soli Type _E, 0 _eo • Figure 4. Ground Motion Hazard. Spectral Acceleration at • Figure 5. liquefaction Hazard. Permanent Ground Displace Period of 1.0 Sec. ment. Bridge Damage State • None County's projected network for the • Slight • Moderate year 2020 that was provided by the • Extensive • Collapse Shelby County Office of Planning SoIl Type :"'0_e and Development; (b) soils input data, in terms of NEHRP soil classi fications and initial screening for liquefaction potential, were based on local geology mapping carried out by the Center of Earthquake Research and Development at the University of Memphis; and (c) at tribute input data for each of the 384 bridges in the network were • I<1gure 6. Bridge Damage States based on data compilation efforts by Jernigan (1998). tination (O-D) zones in the COlIDty. Figure 3 shows these O-D zones, as Scenario Earthquakes well as those zones for which post This SRA was conducted as a earthquake travel times were moni walk-through analysis with a dura tored in this SRA. tion of 50,000 years. Earthquakes occurring during each year of this Input Data duration were estimated by adapt The input data for this SRA are as ing the Frankel et aI. (1996) mod follows: (a) system input data de els of the region. This generated scribing the roadway network ge 2,321 earthquakes with moment ometry, traffic capacities, O-D magnitudes ranging from 5.0 to 8.0. zones, and traffic demands were de Each earthquake was located into veloped from a working file for the one of the 1,763 micro zones (with lengths and widths of about 11.1 68 km) that encompassed the sur rounding area. Typical Results for One Scenario Earthquake and Simulation To illustrate the form of the re ::::"... sults for one earthquake and simu lation, we consider a scenario + earthquake with moment magni j~ tude of 6.9 centered about 65 km ) northwest ofdowntown Memphis. I~,I I .. :. For this event, ground shaking haz . ... ards and liquefaction hazards were estimated by the previously noted methods by Hwang and lin (1997) and Youd (1998) respectively. (Fig ures 4 and 5). Next, the Dutta Mander (1998) approach was used to estimate bridge damage states (Figure 6), and associated system states at various times after the earthquake were developed by ap plying the first-order repair model given in Werner et aI. (iOOO) to these damage states (e.g. , Figure 7). Network analysis procedures sum marized earlier in this paper were II! Figure 7. System States at 7 D~s (top) and 60 days(bottorn) after then applied to each system state, Earthquake to obtain corresponding system wide traffic flows and travel time delays. Fmally, simplified economic Exposure Times analysis methods adapted from Q) 100 years g 111 California Department offranspor m "'0 tation models and summarized in (]) -~~ Werner et al. (2000) were used to ~ 0.01 ~.¥10 years ~::::: ... estimate corresponding economic a ...... losses (due to commute time in g 11m, ------:0 ~'~ creases only). ttl ..0 ~-- e 0.0001 0.. Economic Losses o 20Q 400 ern roo 10m 120l 1400 1GOJ After results of the type shown loss in Dollars (M) above are developed for each sce nario earthquake and simulation during each year of the • Figure 8. Economic Losses due to Earthquake-Induced Increases in walkthrough, they can be aggre- Commute Tnnes in Shelby County, Tennessee 5Issessing tfie Seismic Peiformance ofJligfi'lOalf Sv,stems 69 • Table 1. Incrl'asc in Access Times to Locations in Shelby County due to Damage to Highway System from Earthquake with Magnitude 6.8 centered 66 km Northwest of Downtown Memphis gated to obtain probablistic esti earthquake damage to highway sys mates of economic losses. Figure tems can affect post-earthquake 8 shows results of this type for ex traffic flows and travel times. It is posure times of 1, 10,50, and 100 a technically sound and practical years. Deterministic estimates of approach that will enable decision economic losses can also be ob makers to consider system-wide tained for selected individual earth traffic effects when evaluating vari quake events. ous seismic risk reduction options for highway components and sys Travel Tunes for Selected tems. Although the basic SRA method Locations ology is in place, further work re For emergency planning pur mains before it can be disseminated poses, it may be of interest to esti and applied to highway systems na mate how travel times to and/or tionwide. For example, the from selected key locations in the REDARS 1.0 pre-beta software region may be affected by earth package that is based on this meth quake damage to the highway sys odology is now being indepen tem. Such results can be developed dently validated and applied. This as aggregated probabilistic curves will help to identify future direc (similar in form to Figure 8) or as tions for further development of deterministic estimates for selected this software. Additional improve earthquake events (see Table 1). ments now being planned include: (a) incorporation of models for es timating scenario earthquakes and Conciusions/Future ground motion hazards nationwide; Directions (b) development of models for es timating system-wide landslide and The risk-based methodology de surface fault ntpture hazards; (c) de- scribed in this paper estimates how 70 velopment of improved compo walls, and culverts; and (e) devel nent repair models; (d) develop opment of the system module to ment of vulnerability/fragility accommodate post-earthquake traf models for retrofitted bridges as fic demands that differ from pre well as other highway components earthquake demands. such as tunnels, slopes, pavements, References------ Dutta,A. and Mander,].R., (1998), "Seismic Fragility Analysis of Highway Bridges," Proc. of lNCEDE-MCEER Ctr-Ctr Workshop on Earthq. Engrg. Frontiers in Transp. Sys tems, Tokyo Japan, June 22-23. Frankel, A., Mueller, c., Barnhard, T., Perkins, D., Leyendecker, E.Y., Dickman, N., Hanson, S., and Hopper, M., (1996), National Seismic Hazard Maps, June 1996 Documen tation, Preliminary Report prepared at USGS, Denver CO, July 19. Hwang, H.H.M. and lin, H., (1997), GIS-Based Evaluation of Seismic Performance of Water Delivery Systems, University of Memphis, Memphis TN, Feb 10. Jernigan, ].B., (1998), Evaluation of Seismic Damage to Btidges and Highway Sys tems in Shelby County, Tennessee, Ph.D. Dissertation, Univ. of Memphis, Memphis TN, Dec. Moore, ].E. II, Kim, G., Xu, R., Cho, S., Hu, H-H, and Xu, R, (1997), Evaluating Sys tem ATLHIS Technologies via Rapid Estimation of Network Flows: Final Report, California PATH Report UCB-ITS-PRR-97-54, Dec. Tokimatsu, K and Seed, H.B., (1987), "Evaluation of Settlements in Sand due to Earth quake Shaking," J of Geotech. Engrg., ASCE, Vol. 113, No.8, pp. 861-878. Werner, S.D., Taylor, C.E., Moore, ].E. II, and Walton, ].S., (2000), A Risk-Based Meth odology for Assessing the Seismic Performance of Highway Systems, Report MCEER-OO-D014, Multidisciplinary Center for Earthquake Engineering Research, Uni versity at Buffulo, (Dec. 31). Youd, T.L., (1998), Screening Guide for Rapid Assessment of Liquefaction Hazard at Highway Bridge Sites, MCEER-98-0005, MultidiSciplinary Center for Earthquake Engineering Research, University at Buffalo, May. .Jlssessing the Seismic Performance of J-{igfi:-Ulag Slfstems 71 Page Intentionally Left Blank Analysis, Testing and Initial Recommendations on Collapse Limit States of Frames by Darren Vian, Mettupalayam Siz!aselvan, iJificbel Bruneau and Andrei Reinbom Research Objectives The objective of this study is to develop a set of guidelines and analytical tools for use by practicing engineers to determine the collapse limit state of structures. A program of shake table testing of simple frames through col . National Science FOllndatkm, Earthquake Engineering lapse was carried out, and analytical strategies to capture this behavior have Research Centers Program been developed and made available on the web via MCEER's users network. The research performed here helps to develop a better understanding of the collapse mechanism and pro"ides tools for further investigation. s inelastic behavior is more extensively relied upon in the dissipation .eanbTeam of seismic input energy; the destabilizing effect of gravity becomes A Da17fm Vian, Graduate more significant in the structural evaluation of existing structures. How Student, Mettupa!ayam ever, practicing engineers have limited confidence in the adequacy of cur Sivase!1Jan, Graduate rently available analytical tools to accurately predict when collapse will Student, Michel Bnmeau, occur (i.e., the collapse limit state). As a result, there is a need to investigate Professor, and Andrei the seismic behavior of structures to enhance our understanding of the Reinbo171, Professor, condition ultimately leading to their collapse, and to ensure public safety Department a/Civil, during extreme events. While many experimental studies and theoretical Stntctural and Environmental damage models support these calculated values,it remains that few experi Engineering, University at mental studies have pushed the shake table tests up to collapse. Buffal.o This paper presents the results of research to provide some ofthat data through a program of shake-table testing of simple frames through col lapse, developments of analytical strategies to capture this behavior, and recommendations for deSign. Note that every effort was made to ensure that the experimental data is fully documented (geometry; material prop erties, initial imperfections, detailed test results, etc.); it will be made broadly available on MCEER's Users Network such that these tests can be used at a later time by other researchers as a benchmark to which analytical mod els can be compared. An example of such use is the development made by the authors, described herein, using the benchmark test results. Experimental Program Fifteen specimens, each consisting of four columns, were tested to failure in the course of this research. These were subdh'ided into three groups of five with slenderness ratios of 100, 150, and 200. The dimen- 73 cut from hot-rolled steel plate and then milled to size. Mass applied 35':6m~mm ~Jf·~mommal to each specimen column varied MCEER Users Network: o.o 0 from approximately 150 kg to bttp.!lcivil.eng. buffalo.edul 385 kg. Predicted fundamental pe users_ntwk riod of vibration for the specimens, using nominal dimensions, varied from 0.191 sec up to 1.098 sec con sidering the P-~ effect. Note that the specimens were not intended to be scaled models of actual stmc tures; therefore unscaled ground motions were used. A sample column layout is shown • Figure 1. Typical Specimen in Figure 1. A range of values for axial capacity versus demand, P,I , Pn was used for each slenderness sions and mass used were varied ratio, where P u is the weight of the within each group. Nominal speci mass plates used in the test, and Pn men column widths ranged from is the axial capacity of all columns 2.8 mm (1/8 in) to 9.8 mm (3/8 in). in the specimen, calculated using Column heights ranged from theAISC-LRFD specifications (AISC 91.7 mm (5.41 in) to 549.9 mm 1993). This range of values is (21.65 in). Individual columns were shown in Figure 2. A number of initial imperfections were carefully measured and documented, and 1.0 movement in the transverse direc I Specimen numbers next to value I 0.9 tion was prevented by flexible 15 a: 0.8 5 14 braces verified to have no impact ~ 0.7 10 on behavior in the principal direc 0.6 f 9 !il- 0.5 tion, as reported in Vian and 3 ~ 0.4 13 Bmneau (2001). c 8 ~ 0.3 4 7 12 A schematic of the test setup and 8 0.2 2 6 .11 0.1 instnunentation is shown in Figure 1 0.0 3. Some special attachments and 75 100 125 150 175 200 225 modifications to standard displace Slenderness ("-I,) ment transducers were developed (Vian and Bmneau, 2001) to ensure • Figure 2. Specimen Axial Load-to-Strength Ratio versus Slenderness reliable measurements up to and rl1e.~X:~ri1Iienta1~taand.an~lytical models from this :rest!~<;~ can ~ rlsetl by ot~¢rresearchers as a benchmark for c~mparis()1i. Eventually; the research will enable struc tural engineers to adequately predict when collapse of a givenstructw.-ewill.O(:cur, thus ensuring public safety dur ing strong earthquakes. . 74 LEGEND o ..l::....DISPLACEMENTTRANSOUCER ~ ACCElEROMETER STRAIN GAGE ~ "m" D'~~ffi DHOREAST '1,_~~1::'moo HORWEST Ace EAST I ACCWES I I I I ---r,-_T I A1N 7 , I COLUMN HSGHT = l , ...... - QLI, SDOF SHAKING TABLE DIRECTION OF SHAKING IIilI Figure 3. Schematic of Test Setup and Instrumentation (Looking West) through collapse, but these details mated peak inelastic response. are beyond the scope of this paper. Among all the data recorded and A free vibration test was first per stored, it is noteworthy that the ac formed on each specimen. Subse tual total horizontal displacement of quently applied were a number of a specimen was calculated using the ground motions progressively in correction shown in Figure 4. creasing in magnitude from approxi mately two-thirds of the estimated peak elastic response to the esti- Ul"llll ROD ...... r---+----I ~- vt 'r' ~ L ITransducer I Tube I II Figure 4. Definition of Variables used in Horizontal Displacement }'l.nafys~ 'Testing anti Initia[ 2?g.commenaations on Co[[apse Limit States of :Frames 7; Summary of Test Results shear, Vyo' versus displacement ductility, Il, as well as similar plots The fundamental period of vibra of base shear versus percent tion of each specimen is obtained drift, Y (=urcl-ma,!Lavg)' experimentally via Fourier Spec Some typical results are shown in trum Analysis of the time history Figure 5. data of free-vibration tests. Interest ingly, the damping ratio was ob served to vary as a function of the Behavioral Trends amplitude of linear elastic response. The value of the stability factor, in Results from the tests to collapse dicated in the preceding discussion, include: has a significant effect on the re • Seismic response of shake table sponse of the structure. In practical tests including time history plots bridge and building structures, e is of target table acceleration,mea unlikely to be greater than 0.10, and sured and ftltered table accelera is generally less than 0.060 (MacRae tion, total mass acceleration, et. al., 1993). Specimen 1 is the only relative mass displacement, as one here that has a e value near that well as a plot of estimated base suggested practical range for the sta shear including P-~, V *, versus p bility factor; with a value of 0.065. relative displacement. Specimens 2,6, and 11 have stability • Time history plots of relative dis factors slightly larger than the likely placement for each test in the upper limit, atO.123, 0.101, and 0.138, schedule for comparison of dis respectively. All other specimens placements throughout the have a value ofe ~ 0.155. range of progressive collapse. Three dimensionless acceleration • Plots of estimated base shear,V/, parameters were compared with normalized by the plastic base Trial 10 Trial 11 2 1·· 1·· 0 0 ;t ~>. iCa. 0 ;'a. ::> -1 .- '~5L---.J. 2 5 ----'O--2.:':5,.--~ -2 ·2 -25 -20 0 5 -30 -25 -20 ·15 ·10 Ductility. J.L Ductility. J.L Ductility. J.L • Figure 5. Typical Results from Tests to Collapse 76 five dimensionless displacement parameters. The following general 3.0 observations can be made: 2.5 • The elastic spectral acceleration, \ Sa' ductility, Il, and percent drift, II y=o.Ol75£,·8118 20 \ f-- Z ,,(, were observed to have inverse \ r HR =O.8604 '\ ) II relationships with In support e. ~ of this observation, these vari 1. 0 ables are plotted in Figures 6 1\ I ~ and 7 versus the stability factor o. 5 >'-. for the next to last test (given r-- • o. 0 subscript "final"). This suggests 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 that the structures may be less Stability Coefficient, e able to undergo large inelastic excursions before imminent in II Figure 6. Spectral Acceleration versus Stability Coefficient stability as the stability factor in creases. • Specimen 1 was the only speci men that underwent both a duc tility greater than five (20.35), and a drift larger than 20% ofthe 25 specimen height (64%), prior to I. I collapse, as shown in Figure 7. 20 Recall that this is the only speci _ 15 y=o.4503£1.0465 § 2 men that has a value of e less I R =O.715 than 0.1. 10 Overall, these Figures show a high \. I dependence of ultimate inelastic ~ behavior upon the stability factor for 'r~ • • I a P-L1 affected structure. For the 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 specimens tested in this research, Stability Coefficient, e those that had a value of e equal to 70% I • 1 or greater than 0.1, tended to have a 600/0 I relatively low level ofineIastic behav SO% ior before collapse of the structure. y=o.0386,,-O·6638 ( RZ=O.4568 Structures with e < 0.1 were able to \ ) ,~ withstand ground motions with 0 I higher spectral accelerations, expe 200/0 I' rience larger values of ductility, and I t- I. I 100/0 • accumulate larger drifts, than those '"• 10- • I .1 -.. i with e > O. L The more slender struc 0% I 0.3 0.5 0.6 0.7 tures, characterized by a larger e 0.0 0.1 0.2 0.4 value, will undergo relatively small Stability Coefficient, e inelastic excursions prior to collapse. III Figure 7. Plots of Displacement Ductility and Drift versus Stability Coefficient }In.al!Jsis; 'Testing ami Initia{ :R.gcommencfations on Co{(apse Limit States of :Frames 77 Analytical Modeling ibility approach. The solution pro cedure associated with the model and Verification using allows verifying their performance Data Generated From up to complete collapse. the Experiments Data from the experiments per Element Equations formed in this research can be used Figure 8 shows the deformed in the verification of time history shape of an Euler-Bernoulli beam analysis programs in the modeling in co-rotational coordinates at of inelastic structural behavior up tached to the initially straight to collapse. In these studies, an at centerline of the beam. The non tempt was made to formulate a linear strain displacement relation flexibility-based planar. beam-col ships are given (Huddleston, 1979) umn element, which can deform by: inelastically lmtil it looses stability de or deteriorates and cannot sustain dx = (1 + E) (1) gravity loads. The formulation has no restrictions on the size of rota tions, using one co-rotational frame d~ = (1 + E) cos e (2) dx for the element to represent rigid body motion, and a set of co-rota dry = (1 + E) sine (3) tional frames attached to the dx integration points, as customarily where (~,rl) is the coordinate of used to represent the constitutive a point which was at (."(;,0) before equations. The formulation shown deformation, 6 is the angle made in the next section provides an en by the tangent to the center-line hancement to the existing models with the horizontal, e is the axial by adding the inelastic behavior in strain of the centerline and is the an element, which is stable under curvature. static and reversible loads with large deformations using the flex- • Figure 8. Euler Bernoulli Beam Subjected to Large Defonnation 78 Considering a small perturbation instead of stress-strain relationships. about this deformed position, the Simeonov and Reinhorn (2001) incremental compatibility condi derived a smooth three-dimen tions are given by: sional plasticity model for arbi trarily shaped yield functions and de (). - = eifJ + 1 + £ ifJ (4) used it to represent the constitu dx tive behavior of beam-column cross-sections. This model is based ~; = e cos e - [(1 + £) sin e]e (5) on a parallel-spring plasticity model (Park and Reinhorn, 1986, and Nelson and Dorfinann, 1995) with := =eSine+[(l+£)COSe]e (6) an extension of Bouc-Wen hyster etic model (Wen, 1976, Sivaselvan Integrating these equations over and Reinhorn,2000) and its gener the length o! the element and per alization to multiple dimensions by forming a series of integrations by Casciati (1989). The model deter parts, the following variational mines the generalized force,F, vec equation is obtained: tor (forces and moments) as: F= Fe +Fh (9) Fe =af. (10) Fb =(I-a)Ko[I-BH1H2]E (11) (J) where, Fe and Fb are the elastic 'U where I = axial displacement, and hysteretic components of F, a 'U left 'U is 2 = rotation at end and 3 = the diagonal matrix of post-yield rotation at right end as shown in rigidity ratios, f. is the vector of Figure 8, ¢ = (1+c) ¢ and is found strains and curvatures,Ko is the ini to be the work conjugate of the co tial tangent rigidity matrix, I is the rotational moment. This agrees identity matrix and Hi and ~, the with the result of Reissner (1972). step functions for yielding and for b is a matrix given by: loading reversals, respectively, are described below. B is the force sine sine cosO ------moment interaction matrix: b= gel) gel) g g B = (r;)(r;)T (1- a)Ko Tj ---1 (12) gel) gel) (~~r (I-a)Ko(~) (8) where @(FJ is the yield function. Hj(FJ is the step function the de Constitutive Model noted yielding given by: In this work, the inelastic behav (13) ior of the members is captured in a global sense.The relationships be where N is a parameter that gov tween stress resultants (axial force, erns the smoothness of the transi bending moment, etc.) and gener tion from the elastic to the plastic alized strains (centerline strain, cur state. When N tends to infinity then vature, etc.) are used directly the model collapses to a bilinear Ylnafysis, 80 The results obtained from the cur rent formulation and those from the Illite element analysis program _ New Formulation ABAQUS formonotonic loadings,are 500.0 --- Abaqus shown in Figure 11 with good agree ment. Figure 12 shows monotonic deflections of columns leading to complete loss of strength assuming actual elastic material behavior. The slight difference between the results stems from the representation of moment-curvature-behavior in the current approach vs. ABAQUS. The present fOrmulation uses the smooth ness parameter N=2 to represent the II Figure 11. Analytical Results for Linear Elastic Material post-yield behavior.While this is suf ficient for structural steel sections, a more accurate description is re quired for a square section. Efforts are underway to incorporate the ex ----····-··---~-·-···--·-·1 act moment-curvature equation pre .--:-~~:~~~=~------'------:------1 sented by Stronge andYu (1993). ~ ... ~...... In the dynamic analysis, the speci Q.~ 11~ lTD tf If .. -.. ;.. - ... -.-.-.... ~.~~:'-,...... -.... ;... -.-.-.. ----.t .. ---.. --.-~ men was subjected to histories ; : ...... ; ~ ~ used in the experiment without .... --.... ~ .... -.-.-.. -.... ~ ... --.---.--~--- .... -.--.- ~-----.-.--... -~...... ----J.---.-.... --~ i rreoretica~Respons~in~ 1 ...... ~ i considering the initial state of the ·········---·-·t-·---·---~~p~a,,~-Ftc:"-~t~'f------t------~~~",-<---.-~ model (previously subjected to se ries of base motions) and the re sult is shown in Figures 13 and 14. The time history results differ from those obtained in the experi II Figure 12. Static Analysis with Nonlinear Material ment (see Figure 14). However, when the initial deformation ofthe model is considered, the analysis 60.0 model shows same pattern as the _ 50.0 E experiment as shown in Figure 15. .§. E 40.0 CD E 2l 30_0 Comparison with g-'" 15 20_0 NCHRP 12-49 ]! 2 10.0 -.:: o Proposed P-11 Limits :t: f\ /\ ~ f\f\AFJvI 0.0 " v ~v v 1 1! The National Cooperative High ·10_0 way Research Program (NCHRP), Time (s) Project 12-49, under the auspices of the Transportation Research Board, is investigating seismic de II Figure 13. Horizontal Displacement from Dynamic Analysis sign ofhridges from all relevant as pects. At the conclusion of this Ylnafgsis, 'Testing ana Initia{ ~commentfations on Co[fapse Limit States of :Frames 81 The displacement of a pier or Horizontal Displacement bent in the longitudinal and 60 , transverse direction must satisfy 50 proposedAASHTO LRFD Equation 3.10.3.9.4-1 : .sE 40 ~ 30 I E Am';OoZ5.C{:}H (16) a) Computational b) Experimental shear coefficient, Cs *, is plotted as (with initial offset) a function of the maximum drift, y. Results for specimens with e < 0.25 • Figure 15. Displacement Response at the Verge of Collapse (1,2,4,6,7, 11, and 12) are shown. During the initial tests, when the proposed limit was satisfied, none project, proposed revisions to the of these specimens failed. Due to current LRFD specifications for repeated inelastic action, the cumu highway bridges will be presented lative drifts of the structure in to theAmericanAssociation of State creased, eventually causing Highway Transportation Organiza progressive collapse and violating tions (AASHTO) for review and the proposed limit. Collapse always possible implementation. Included occurred only after the limit was in the proposed revisions are how exceeded in a prior test, thus vali additional demands from P-Ll affect dating the proposed criterion. As structural performance. The most shown in Figure 16, the remaining recent proposed provision, as of specimens, for which e ~ 0.25, this writing, states: never satisfied the drift criteria, even for those tests that remained 82 1.00 T··;,······;······;·······;;······"·······;· .... T······;··· .. '-·/.·~,..···,;······.;,...... ;,······;"N=C=H=R=P=1=2=-A=. 9=p=lr=o=r:o=s=ed=P=-=A='=im=@";; I i"'_,,_,, __ ,,_:...... j ~r.__ 1-L"...:.. ..;:, .• y~~;;=i:!.=.~;~~~:=,:===~?. 0.75 -l1··.·.·..... ".,•. ~"'~.d~lI~:~::>;II.:"':::.::>:-;·i .. "::.II:;:IIl":~:·'" -+Jr:".·.;:.V.~:It..d,'.. "'.::>.~' ".111+Il_+ ..... :~: ., A : ; ;; /' ' :. : : ./'" /; ,;, , .. ++···;·····+ .... ·;-.. ·r;···;·· .. +· .. _·;; ~oW 050-l~;·····: ...JI!lP.,""; ·~~~~/~~·++···+·;······:;·-·+·+ : +.... ,-++·+.... ; : 025 0% 10% 20% 30% 40% 50% 60'% drift,'Y (a) c; vs drift - e< 0.25 LOO ; J ~ ;• L ; NCHRP 12-49 proposed P- A limitl" • : I ..... T ; : f ; : : : 0.75 •• I.. . , . ~ /. V!olate • lir~it ; ...... patisfies limit 1. ~ ~w '/; ,- o j 0.50 f J ::1·:]: '/ t::;::: J ' -T '/ : [j~ F~E J ; : 0.25 '/ .... j • II:! :r'; ... Specimen 5b ..... Specimen 8 '/i;tf; : -!t- Specimen 9 --;:a- Specimen 10b /W ...... Specimen 13 ~ Specimen 14 1/.' :• ; - NCHRPUmit 0.00 10% 20"10 30% 40% 50% 60% 70% drifi,'Y C; vs drift - e2: 0.25 III Figure 16. Test Results Comparison with NCHRP 12-49 Limits in the elastic range. The stability factor for these specimens, how Conclusions ever, is well above the practical The experimental data generated range discussed previously; there by this project provides a well fore, the limit violation is of no con documented database of shake sequence. table tests of a SDOF system sub jected to earthquakes of progres- JInafysis, 'Testing and Initia[ !R.f;commenaations on Co[fapse Limit States of:.Frames 83 sively increasing intensity up to col tant points that must be considered lapse due to instability. This data in the design of slender steel struc will be useful for, and shared with, tures. The stability coefficient, e, other researchers who may wish to has the most significant effect on validate or develop algorithms ca the behavior of the stmcture. As e pable of modeling inelastic behav increases, the maximum attainable ior of steel frame structures up to ductility, sustainable drift, and spec and including collapse. The data tral acceleration, which can be re presented here will also be located sisted before collapse, all decrease. on the Internet (with all interme When this factor is larger than 0.1, diate data files) for immediate ac the ultimate values of the maxi cess by other researchers. mum spectral acceleration, dis An attempt was made to analyti placement ductility, and drift cally model the collapse using an reached before collapse are all advanced flexibility based formu grouped below values of 0.75 g,5, lation with large deformation in and 20%, respectively. Stability co elastic behavior. The verifications efficient values less than 0.1 tend and the refmement of the model to increase each of those response are done using the experimental values significantly. The research data. Parametric studies can be per performed here helped develop a formed with the model to support better understanding of the col further codes and standards devel lapse mechanism and provides the opment efforts. tools for further work to modify or The research presented here improve current design standards. demonstrated a number of impor- References------ AlSC, (1993), Load and Resistance Factor Design Specification for Structural Steel Buildings, American Institute of Steel Construction, Chicago, IL, Decemberl. Brenan, K. E., Campbell, S.1., and Petzold, 1. R., (1996), Numerical Solution of Initial value Problems in Differential-Algebraic Equations, Philadelphia, Society for Industrial and Applied Mathematics. Casciati, E, (1989), "Stochastic Dynamics of Hysteretic Media," Structural Safety, Vol. 6, No. 2-4, pp. 259-269. Huddleston,]. v., (1979), "Nonlinear Analysis of Elastically Constrained Tie-Rods under Transverse Loads," International Journal of Mechanical Sciences, Vol. 21, No. 10, pp. 623-630. MacRae, G. A., Priestley, M.].N., Tao, J., (1993), P-Li Effects in Seismic Design, Report No. SSRP-93/05, Department of Applied Mechanics and Engineering Sciences, University of California, San Diego. Nelson, R. B. and A. Dotfmann, (1995), "Parallel Elastoplastic Models of Inelastic Material Behavior," Journal of Engineering Mechanics, Vol. 121, No.10, pp. 1089-1097. Park, Y]. and Reinhom, A.M., (1986), "Earthquake Responses on Multistory Buildings Under Stochastic Biaxial Ground Motions," Proceedings of the Thb'd US. Conference on Earthquake Engineering, Charleston, South Carolina, Vol. 2, pp. 991-1001,August 1986. 84 References (Cont'd) ------ Reissner, E., (1972). "On One-Dimensional Finite-Strain Beam Theory: the Plane Problem," Journal ofApplied Mathematics and Physics (ZAMP), Vol. 23, No.5, pp. 795-804. Simeonov, v.: K., (1999). Three-Dimensional Inelastic Dynamic Structural Analysis of Frame Systems, Ph.D. Dissertation, Civil, Structural and Environmental Engineering, University at Buffalo. Simeonov, v.: K, Sivaselvan and A. M. Reinhorn, (2000), "Nonlinear Analysis of Structural Frame Systems by the State-Space Approach," Computer-Aided Civil & InfrastntctureEngineering, Vol. 15, No.2, pp. 76-89. Simeonov, v.: K and Reinhorn, A.M., (2001), Three-Dimensional Inelastic Dynamic Structural Analysis of Frame Systems, Technical Report MCEER-01-xx, Multidisciplinary Center fur Earthquake Engineering Research, University at Buffalo (in re\iew). Sivaselvan, M. v.: and Reinhorn, A.M., (2000), "Hysteretic Models for Deteriorating Inelastic Structures," Journal of Engineering Mechanics, Vol. 126, No.6, pp. 633- 640. Stronge, w.]. and Yu, TX., (1993), Dynamic Models for Structural Plastidt); London, New York, Springer-Verlag. Vtan, D. and Bruneau, M., (2001), Experimental Investigation Of P-Delta Effects To Collapse During Earthquakes, Technical Report MCEER-01-0001, Multidisciplinary Center fur Earthquake Engineering Research, University at Buffalo. Wen, Y. K, (1976), "Methods of Random VIbration of Hysteretic Systems," Journal of the Engineering Mechanics Division, Vol. 102, No.2, pp. 249-263 . .9l.nafgsis, 'Testing anti Initia{ :R,gcommenaations on Coilapse Limit States of:Frames 85 Page Intentionally Left Blank Centrifuge-Based Evaluation of Pile Foundation Response to Lateral Spreading and Mitigation Strategies by Ricardo Dobry, Tarek H Abdoun and Tbomas D. O'Rourke Research Objectives A first objective of the research is to identify and quantify the mecha National Science Foundation, nisms and parameters determining the hazard to deep foundations and Earthquake Engineering superstructure caused by the lateral spreading.Two main situations have Researcb Centers Program National Science Foundation: been identified, depending on pile foundation bending response controlled U.S.-Japan Cooperatiz'B by the pressure of the liquefied soil or by that of a shallow nonliquefied Research for Mitigation of layer. Field evidence, centrifuge results, and analyses have shown the shal Urban Earthquake low nonliquefiable layer to be more critical and more amenable to retrofit Disasters ting. A second objective of the research is to develop and evaluate strategies Federal Highway to retrofit deep foundations by decreasing the pressure exerted by this Administration nonliquefied shallow layer, with emphasis in cost-effective and advanced materials. A third objective is to develop fragility curves for nonretrofitted andketrofitted foundations. he effects of liquefaction on deep foundations are very damaging and T costly. Permanent lateral ground deformation or lateral spreading is a Ricardo Dobry, Professor; main source of distress to piles, either alone or in combination with iner Tarek H. Abdoun, tial superstructural forces and moments arising during shaking and acting Research Assistant on a soil already weakened by rising water pore pressures. Cracking and Professor, Yinjuang rupture of piles at shallow and deep elevations, rupture of pile connec Wang, Graduate Student, tions, and permanent lateral and vertical movements and rotations of pile Ricardo Ramos, Graduate Student, Cilrfl and heads and pile caps with corresponding effects on the superstructure Environmental have been observed (Figure 1).This has affected buildings, bridges, port Engineering Department, facilities and other structures in Japan, the U.S. and other countries in Rensselaer Polytechnic cluding the 1989 Loma Prieta, California and the 1995 Kobe, Japan, earth Institute quakes (Hamada and O'Rourke, 1992; O'Rourke and Hamada, 1992; Thomas D. O'Rourke, Tokimatsu et al., 1996; Dobry and Abdoun, 2001). Thomas R. Briggs Professor Examination and analysis of case histories have revealed important as ofEngineering and Siang Huat Gob, Research pects of the phenomenon and highlighted its complexity: It is essentially Assistant, Department of a kinematic soil-structure interaction process involving large ground and Civl and Environmental foundation permanent deformations, with the deep foundation and su Engineering, ComeO perstructure responding pseudostaticallyto the lateral permanent displace University ment of the ground. While in some cases the top of the foundation displaces laterally a dis tance similar to that in the free field, like in Figure 1 where both the 87 grOlmd and foundation moved hori ure 1). More damage tends to oc zontally about 1 m, in others it cur to piles when the lateral move moves much less due to the con ment is forced by a strong straining effect of the superstruc nonliquefied shallow soil layer • Mourad Zegbal, Assistant ture, or of the deep foundation's (end-bearing pile No.2 in Figure 1), Professor and Vivian lateral stiffness including pile than when the foundation is freer Kallou, Graduate Student. groups and batter piles. The foun to move laterally and the forces Civil and Environmental dation may be exposed to large lat acting on them are limited by the Engineering Department, eral soil pressures, including strength of the liquefied soil (float Rensselaer Polytechnic especially passive pressures from ing Pile No.1 in Figure 1). institute the nonliquefied shallow soil layer Lateral spreading has been iden • Masanori Hamada, riding on top of the liquefied soil. tified as a major hazard to pile foun Professor, Waseda In some cases, this soil has failed dations of hospital buildings, and University, japan before the foundation with negli centrifuge modeling as ·a key tool gible bending distress and very to identify and quantify mecha small deformation of the founda nisms, calibrate analyses and evalu tion head and superstructure ate retrofitting strategies for pile (Berrill et al., 1997); while in oth foundations. Figure 2 shows the ers the fotmdation has failed first 100 g-ton RPI geotechnical centri in bending (Figure 1) and/or has fuge used for this research, which experienced excessive permanent is located at the RPI campus in Troy, deformation and rotation at the pile NewYorkThis centrifuge, originally heads. The observed damage and commissioned in 1989 with sup cracking to piles is often concen port from MCEER (then NCEER), trated at the upper and lower has in-flight earthquake simulation boundaries of the liquefied soil capability allowing base shaking to layer where there is a sudden be applied to the base of the model. change in soil properties, or at the It was recently selected by NSF to connection with the pile cap (Fig- gether with other earthquake en- Engineering design rtrms~foun.datiol1andconsu1ting en~ gineers, hospital authoritieS, SUite tt'atlsp()t}ationdepart ments, and port and harh?" authO~tifS 88 Pile 1 Pile 2 Program 2: Seismic Retrofit of Hospitals 8.5 to 9.0m Liquefiable 1 Layer Finn Base (;;:1.5 m, say 1.5 m or 11.5 m below GS.) III Figure L Damage to pile foundations due to lateral spreading under NFCH building, 1964 Niigataearthquake,Japan (Hamada, 1992, 2001) gineering experimental sites started in 1995 with support from throughout the U.S. to form the NCEER and NSF and has continued George E. Brown, Jr. Network for since then with current support EarthquakeEngineeringSimufution from both MCEER and NSE The (NEES, www.eng.nsfgov/nees).Ad technical discussion below is di ditional information on the centri vided in three parts: case of pile fuge equipment used in this bending response to lateral spread research, results from other ing controlled by the pressure of projects and the basic principles of the liquefied soil, case of response centrifuge modeling, can be found controlled by shallow nonIiquefied at the RPI web site (www.ce. soil layer; and pile retrofitting strat rpi.edu/centrifuge), which also has egies and results. useful links to other relevant web sites; see also summary articles by Dobry et al. (1995) and Dobry and Abdoun (1998,2001). In addition to the centrifuge experiments themselves done at RPJ, this cen trifuge-based research has included other analytical, laboratory, case history review and retrofitting strat egy components, conducted either at Cornell University or in close co operation between the RPI and Cornell teams. The RPI-Comell jOint centrifuge-based research on II Figure 2. 100 g-ton geotechnical centrifuge with in-flight lateral spreading effects on piles shaking capability at RPI Pile Foundation Response to Lateral Spreading and ilHtigation Strategies 89 Pile Bending Response deformations and strains attained .su.s in the experiments (Figure 5). RPIIOO g-ton geotechnical Controlled by the In the test of Figure 3, the soil centrifuge facility: Liquefied Soil proftle consists of two layers of fine http://wuw.ce.rpi.edll/ Nevada sand saturated with water: Figure 3 shows centrifuge pile centrifuge a top liquefiable layer of relative The complete visualization or Model 3, simulating the bending density, Dr = 40% and 6 m protoype "movie" produced from the response of a pile foundation sub thickness, and a bottom slightly data recorded in the jected to the lateral pressure of a cemented nonliquefiable sand layer centrifuge test illustrated liquefied soil due to lateral spread by the single frame of having a thickness of 2 m. The pro ing. These and other experiments Figure 5 can be viewed in totype single pile is 0.6 m in diam were conducted using the rectan this web site. To see this eter, 8 m in length, has a bending movie, after entering the gular, flexible-wall laminar box con stiffness, EI = 8000 kN-m 2 ,and· IS tainer sketched in Figure 3. This web site, go to Research free at the top. The pile model is and then to Visualization. laminar box is comprised of a stack instrumented with strain gages to of up to 39 rectangular aluminum measure bending moments along NEBS Initiative: rings separated by linear roller its length, and a lateral LVDT at the http://www.eng.nsfgov/nees bearings, arranged to permit rela top to measure the pile head dis tive movement between rings with placement.The soil is instrumented minimal friction. In Model 3 as well with pore pressure transducers (pi as in all other lateral spreading ezometers) and accelerometers, as spreading experiments, the laminar well as with lateral LVDTs mounted box and the shaker under it are in on the rings of the flexible wall to clined a few degrees to the proto measure soil deformations in the type horizontal direction to free field. A prototype input simulate an infinite mild slope and accelerogram consisting of 40 sinu provide the shear stress bias soidal cycles of a peak acceleration needed for a lateral spread.The flex of 0.3 g was applied to the base, ibility of this box container is dem which liquefied the whole top layer onstrated by the large permanent in a couple of cycles and induced a permanent lateral ground surface displacement in the free field of aboutO.8m. 50 g Results of this experiment are shown in Figures 4 and 5.As soon as the top sand layer liquefied at the beginning of shaking, it started Z=Om moving laterally downslope throughout the shaking, with the maximum displacement at all times Z=6 measured at the ground surface, Z=8 and with this surface ground dis placement increasing monotoni cally with time to its final value DH • Strain Gage • Pore Pressure __ LVDT • Accelerometer = 0.8 m at the end of shaking. The Transducer maximum bending moment along the pile at any given time occurred • Figure 3. Lateral spreading pile centrifuge model in two-layer soil profile at the interface between the two (Abdoun, 1997) soil layers, that is at a depth of about 90 6 m. Figure 4 shows the time his tory of this prototype bending moment for Model 3, measured at 80 E 60 z = 5.75 m; the plot reveals that the .3- moment increased to a maximum ci. 40 (J) Mmax = 110 kN-m at a time, 6 20 e = 17 sec, with the moment de a 45 creasing afterwards despite the Pile head continuation of shaking and the 30 continuous increase of the soil de 15 formation in the free field.The pile a head displacement in the same fig 120 ure also reached a maximum at about 17 sec and decreased after 80 wards. Clearly at this time the liq 40 uefied soil reached its maximum a strength and applied a maximum a 10 20 30 40 lateral pressure to the pile, with the soil flowing around the pile, exhib iting a smaller strength and apply III Figure 4. Prototype lateral displacement of soil and pile and ground surfuce, ing a smaller pressure afterwards; and pile bending moment at a depth of 5.75 m in model ofFigure 3 (Abdonn, as a result, the model pile bounced 1997) back and the bending moments decreased. The two photos in Fig ure 6 - taken after the centrifuge side of the pile for direct compari tests - illustrate this flow of lique son between ground and pile dis fied soil around the pile in other placements as well as to visualize two models where colored sand the larger movement ofthe liquefied had been placed around the pile. Figure 5 summarizes the state of the system during a repeat of Model 3, at the tim~ when the pile RPI Input Motion head displacement and the bend r ing moment at a depth of about 6 u m attained their maximum values. This is a frame taken from the visu alization of the experiment pro duced from the measurements (the whole visualization may be viewed at the RPI centrifuge web site).The displaced shape of the box con tainer indicates the lateral spread ing in progress, with concentration of permanent shear straining in the lower part ofthe liquefied soil; this box shape was obtained from the lateral LVDTs placed on the side II Figure 5. Frame taken out of visualization of two-layer centrifuge model of walls. This distorted shape is also Figure 3, produced from the recorded data (KallOll et al., 2001; to see copied as a white mesh to the right whole visualization, visit http://U'tlJW.ce.rpi.eduicentrijuge) Pile Foundation Response to Late-ral Spreading and i~:Iitigation Strategies 91 Direction of the free field lateral.. displacement • Figure 6. Photos showing flow of liquefied sand around the pUe in the downslope direction in two layer centrifuge models (Abdoun, 1997). The photos were taken after the test in models where colored sand had been placed in a circular ring around the pUe soil flowing arOlmd the pile, com sand in that part of the shaking cycle. pared with the displacement of the At other times corresponding to dif pile itself. The blue color in the up ferent parts of the shaking cycle, the per part of the loose sand layer indi whole layer is blue and hence com cates complete liquefaction as pletely liquefied. measured by the piezometers, while In addition to Model 3 summarized the green color in the lower part of in Figures 3 to 5, similar centrifuge the layer indicates lower excess pore tests of a single pile with a pile cap, pressure due to dilative cyclic stress with densification around the pile to strain response of the liquefied simulate pile driving, and with 2:x2 pile groups indicated that, while Mm.1X still occurs at a depth of about 6 m sometime during the shaking, the value of Mmax increases with the area of pile foundation exposed to the soil lateral pressure and decreases in the pile groups due to the contribution to moment of the axial forces in the piles (frame effect). Simple limit equi librium calculations with a constant assumed maximum pressure of the liquefied soil along the pile, PI ,indi cate that values of PI of the order of 10 kPa explain well all measured trends and values of M in this se- • Figure 7. Concept used to develop undrained triaxial extension model for rnax the lateral loading of liquefied soU on the pile (Goh and O'Rourke, 1999; ries of centrifuge tests. Goh,2001) 92 III Figure 8. Comparison between predicted and measured pile bending moment in centrifuge model of Fig. 3 at the lower boundaIY of liquefied soil using trla,\ial e,uension undrained loading approach (Goh and O'Rourke, 1999; Goh, 2001) The physical origin and basic Pile Bending Response mechanisms determilling the behav Controlled by Shallow ior ofthe liquefied soil,including the Nonliquefied Layer lateral pressure on pile fuundations and values such as PI and Mmax mea Figure 9 shows centrifuge Model sured in these centrifuge tests,are not 2, where a strong shallow yet well understood and are the sub nonliquefied soil layer increases sig ject of intense research. The Cornell nificantly the bending response of team has proposed the explanation the pile foundation to lateral sketched in Figure 7, with PI and Mmax spreading. The shallow top layer controlled by the peak undrained shear strength of the saturated sand loaded in the extension mode (Goh and O'Rourke 1999; Goh, 2001). Based on p-y curves generated analytically from triaxial exten sion tests conducted at Cornell using the same Nevada sand and relative density ofthe cen trifuge tests, nonlinear Beam on-Wmkler -Foundation (BWF) -200 analyses ofcentrifuge Mode13 Mcmant (Jd'J-m) -were able to predict closely the measured bending re sponse (Figure 8). II Figure 9. Lateral spreading pile centrifuge model in three-layer soil profile (AMoun, 1997) Pile Foundation Response to Lateral Spreading and iliIitigation Strategies 93 1-'- Model 2 ~ Model 2m I 0 0 0 0 2 2 2 2 :[ s:. 4 4 4 4 Co -CII "C 6 6 6 6 ~ 8 8 8 8 10 10 10 10 -100 -50 0 50 100 -200 0 200 -200 0 200 -200 0 200 Moment ( kN-m) • Figure 10. Measured bending moment response along pile in lateral spreading centrifuge models without (Model 2) and with (Model 2m) inertial loading (Wang, 2001) consists of a 2-m thick (in proto nonliquefiable layer did not fail.The type units), free draining, slightly values of maximum bending mo cemented sand. Model 2m, not ments at 2 m and 8 m are close to shown here, is similar to Figure 9 300 kN-m, much greater than those but with a mass added above measured in 2-layer tests such as ground to evaluate the combined shown in Figure 3, which did not effects of lateral spreading and in exceed 170 kN-m even when a pile ertialloading. cap was added. Figures 10-11 summarize the The shapes of ,the bending mo main characteristics of the bending ment profiles at various times pre response of Model 2 (only lateral sented in Figure 10 indicate that the spreading), which is also typical of deformed shape of the pile had a other pile models tested in this 3- double curvature caused by the top layer soil profile. The same as in and bottom soil layers loading the Model 3 discussed before, the 6-m pile in opposite directions. This thick noncemented sand layer liq double curvature was confirmed by uefied early in the shaking after the fact that when the top soil layer which the lateral spreading in failed, the pile head and cap creased monotonically, reaching a "snapped" in the downslope direc value DH =0.7 m at the end of shak tion (Figure 11), showing that at ing (Figure 11). The pile bending very shallow depths, the pile was moments in the top 2 m first in pushing the soil rather than the creased with time of shaking and other way around. Both the passive then decreased after passive failure failure of the top layer and the of the top nonliquefiable layer moment concentrations at the top against the pile (Figure 10); while and bottom boundaries of the liq the bending moments near the uefied layer indicated by the figures bottom increased monotonically are consistent with the experience and never decreased, as the bottom from earthquake case histories. 94 These moment concentrations are Despite the rapid change in shal also predicted by theory (e.g., low bending moments due to the Meyersohn, 1994; Meyersohn et al., mass, when the top soil layer failed 1992; Debanik, 1997). Another in in passive in Model 2m, the pile teresting aspect of Figures 10 and head and cap "snapped" in the 11 is that the bending moments downslope direction, exactly the vary linearly within the liquefied same as in Model 2 (Figure 11), layer, suggesting that they are es showing that the soil failure mecha sentially controlled by the loading nism was still controlled by lateral of the top and bottom layers, with spreading. the pressure of the liquefied soil Another factor which has been being negligible.The values ofMI]m,'< studied in the centrifuge for the 3- at z = 2 m and z = 8 m are higher layer soil model is the influence of than the corresponding values of the superstructural stiffness that M= at z = 6 m for the 2-layer soil field case histories has shown to be profiles, such as in Figure 4, which important. This has been done by were controlled by the strength of the addition oflateral and rotational the weaker liquefied soil. The au springs above ground connected to thors have successfully calibrated the pile head, such as spring k in a limit equilibrium method to pre Figure 12 (Ramos, 1999). AB ex dict Mma.,,< in some of these 3-1ayer pected, the analysis of these cen pile centrifuge models, after incor trifuge results has required porating basic kinematic consider significant kinematic consider ations to allow for the change in ations and parameters, even when pile curvature (and hence of the simple limit equilibrium calcula sign of the passive soil pressure on tions are conducted. On the other the pile) within the top nonliquefiedsoillayeL The comparisons in Figures 10 and 11 between Models 2 and 2m Measured Soil Disp - Estimated Pile Disp • reveal interesting aspects of the role played by superstructural in Model 2 Model 2m 0 0 ertia in the lateral spreading pro I / Slightly cess. For depths greater than 2 or cemented sand / / 3 m, the effect of lateral spreading 2 z predominates and the inertial load E E ing due to the mass can be ignored. ~4 4 J:: J:: C. C. However, at shallow depths of less Q) Q) '0 '0 than 2 m, that is in the top '0 en 6 0en 6 nonliquefiable layer; the bending Nevada sand Nevaaasand moments of the two centrifuge (Or=40%) (Or=40%) 8 8 models are very different, with Slightly Slightly those of Model 2m changing rap cemented sand cemented sand 10 10 idIywith time due to the combined 0 ZO 40 60 80 100 0 ZO 40 60 80 100 effect of inertia and lateral spread Displacement (cm) Displacement (cm) ing.However,even in Model 2m the maximum moments still tend to concentrate at the upper and lower E Figure 11. Snapping of pile in downslope direction in centrifuge models boundaries of the liquefied layer. without (Model 2) and with (Model 2m) inertial loading (Wang, 2001) Pile Foundation Response to Lateral Spreading and Mitigation Strategies 95 • Input motion: 0.3 9 - 40 cycles Z=Om - 2 Hz • Free rotation: Z=2 Model 3free - 24 Model 3free- 00 Z=8 Model 3free- 00 Z =10 cap Input Motion • Fixed rotation: • Strain Gage. Pore Pressure .. LVDT • Accelerometer - Model 3fixed - 0 Transducer • Figure 12. Lateral spreading pile centrifuge model incorporating effect of superstructural stiffness (Ramos, 1999) hand, some aspects of the analysis rial that will yield under constant become simpler compared with lateral soil forces (Figure 13a).This the case of k = 0 (Figure 3), in that would decrease both bending mo if the value of k is large enough, ments and foundation deforma there is no double curvature of the tions while allowing the ground pile at very shallow depths, and no lateral spreading to take place with "snapping" of the pile in the out interference from the founda downslope direction as in Figure tionAs this retrofitting scheme also 11. That is, the constraining effect decreases the lateral resistance of of spring k forces the lateral pres the foundation to inertial loading, sure of the nonliquefied layer on the desired frangible material se the pile to act in the same lected, while yielding to static force downslope direction at all depths should remain resilient under the between 0 and 2 m. transient inertial loading. Alterna tively, the trench surrounding the foundation with frangible material Pile Retrofitting may be located at some distance Strategies and Results from the foundation so as to in crease the resistance to inertial Both case histories and centri loading (Figure 13b). fuge models have shown the great A series of centrifuge models of importance of the shallow a single pile with pile cap in the 3- nonliquefiable soil in increasing the layer soil profile were conducted bending response of the pile foun using the retrofitting setups of Fig dation.Therefore,a promising reha ure 13, labeled respectively Strat bilitation approach of existing egy 1 and Strategy 2. These foundations is to replace the shal experiments are listed in Table 1, low soil in a trench around piles which include also the benchmark and pile cap by a frangible mate- nonretrofitted Models 2 and 2m, 96 Z=Om Z=2 Z=8 Z=10 a) Model 2r1 (Strategy 1) Z=Om Z=2 Z=8 Z=10 • Pore Pressure • Strain Gage Transducer -lilt- LVDT • Accelerometer b) Model 2r2 (Strategy 2) 1\ Figure 13. Lateral spreading pile centrifuge models to evaluate retrofitting strategies (Wang, 2001) already discussed. Models 2r1, Figures 14 and 15 illustrate mea 2mrla and 2mrlb were done with surements and observations ob Strategy 1, without and with a mass tained from Models 2r1 and 2r2. above ground, and Models 2r2 and The free field lateral ground dis 2mr2 were conducted with Strat placements during shaking in cen egy 2. In both cases, a soft clay was trifuge tests without and with pile placed in a trench either directly foundation retrofitting were essen around or at some distance from tially the same (Figure 11), consis the foundation. In future tests the tent with the assumption that they use of an artificial frangible mate represent truly free field response. rial with higher resistance to tran Figure 14 compares the bending sient loading is planned (W"ang, moment response without and 2001). with retrofittingAs expected, there Pile Foundation Response to Lateral Spreading and lliitigatiol1 Strategies 97 _IModel21 A IModel2r11 .IModeI2r2i1 0 0 0 0 2 2 2 2 g DH=60cm J:: 4 4 4 4 Q. -CII 't:I 6 6 6 6 ·0 CI) 8 8 8 8 10 10 10 10 -200 0 200 -200 0 200 -200 0 200 -200 0 200 Moment (kN-m) • Figure 14. Measured bending moment response along pile in lateral spreading centrifuge models without (Model 2) and with (Models 2r1 and 2r2) foundation retrofitting (Wang, 2001) is a dramatic reduction in the mo ing moment at the lower boundary ments in the top 2 m of pile in con of the liquefied layer, at about 8 m tact with the nonliquefiable soil. depth. Similarly, the pile head dis The maximum moment there was placements at the end of the tests close to 300 kN-m in Model 2 and were reduced by a factor of two by becomes about 10 kN-m after ret retrofitting (from 85 to 40-50 cm, rofitting.A smaller reduction is also with DH = 70 to 80 cm for the soil observed for the maximum bend- in the free field). The photo of Model 2rl in Figure 15, taken after the test, illustrates the correspond ing "crunching" of the soft clay against the pile cap in the upslope side, and opening of a gap downslope between soil and foun dation. However, the counterpart to this reduction in permanent bend ing response to lateral spreading of the pile foundation was an increase of transient pile accelerations and displacements, especially in the tests incorporating inertial loading (Models 2mrla,b and 2mr2, not shown), due to the reduced lateral ground support in the top 2 m of Direction of the free field lateral displacement ~ the foundation; future tests will address this problem. • t1gure 15. Plan view of retrofitted pile cap and ground after the test, Mode12rl (Wang, 2001) 98 II Table 1. Program of Centrifuge Tests to Evaluate Retrofit Strategies 1 and 2 (Wang, 2001) 1':t~sfNb. 1"t~~\:,., .. '·~l.@~:.I:··R~~bfiitirig .~ ·····f~·~.C~mQ1~rifs! 2 Yes No No 2rl Yes No Yes 2r2 Yes No Yes 2m Yes Yes No 2mrla Yes Yes Yes 2mrlb Yes Yes Yes Repeat of 2mrl a 2mr2 Yes Yes Yes Conclusions and shallow nonliquefiable soil layer riding on top of tl;.ie liquefied soil Future Research in increasing foundation bending Case histories during earth response has been clarified. Retro quakes have shown the signifi fitting strategies are being evalu cance of lateral spreading in ated in the centrifuge, aimed at causing damage to deep founda mitigating the effect of lateral tionsandsupportedstructtrresdu~ spreading associated with the pres ing earthquakes.The complexity of sure of this shallow layer while pre the problem requires use of cen serving needed lateral resistance to trifuge physical modeling to clarify inertial loading. Additional work is mechanisms, quantify relations and needed to understand and quantify calibrate analysis and design pro the response of nonretrofitted and cedures. Centrifuge results so far retrofitted pile foundations, with have clarified the deep foundation centrifuge model experiments response, have shown signillcant combined with case studies and agreement with field experience, theory, toward improving the state and are being used to calibrate limit of-practice of seismic design and equilibrium and Beam-on-Winkler retrofitting of deep foundations Springs (p-y) analytical methods. against liquefaction. Specifically, the importance of the References'------ Abdoun, T. H., (1997), Modeling of Seismically Induced Lateral Spreading of Multi LaJ1er Soil Deposit and Its Effect on Pile Foundations, Ph.D. Thesis, Dept. of Civil Engineering, Rensselaer Polytechnic Institute, Troy, NY. Berrill, J. B., Christensen, S. A., Keenan, R. J., Okada, W. and Pettinga,]. K., (1997), Lateral spreading Loads on a Piled Bridge Foundation, Seismic Behavior of Ground and Geotechnical Structures, (Seeo E Pinto, ed.), pp. 173-183, Balkema, Rotterdam. Debanik, c., (1997), Pile Response to Liquefaction-Induced Lateral Spread, Ph.D. Thesis, Cornell University, Ithaca, NY. Dobry, R., Taboada, V. and liu, L., (1995), "Centrifuge Modeling of liquefaction Effects During Earthquakes," Keynote lecture, Proc. 1st IntI. Conf on Earthquake Geotechnical Engineering (K. Ishihara, ed.), Tokyo, Japan, Nov. 14-16, Vol. 3, pp. 1291- 1324. Pile Foundation Response to Lateral Spreading and il,Htigation Strategies 99 References (Cont'd)------ Dobry, R. and Abdoun, T. H., (1998), "Post-Triggering Response of Liquefied Soil in The Free Field and Near Fotmdations," State-of-the-art paper, Proc. ASCE 1998 Specialty Conference on Geotechnical Earthquake Engineering and Soil Dynamics (P. Dakoulas, M. Yegian and R. D. Holtz, eds.), University of Washington, Seattle, Washington, August 3-6, Vol. 1, pp. 270 - 300. Dobry, R. and Abdoun, T. H., (2001), "Recent Studies on Seismic Centrifuge Modeling of Liquefaction and its Effect on Deep Foundations," State-of-the-Art Paper, Proc. 4th Inti. Conf on Recent Advances in Geotechnical Earthquake Engineering and Soil Dynamics and Symposium to Honor Prof W.D.L. Finn (S. Prakash, ed.), San Diego, CA, March 26-31, SOAP-3, Vol. 2. Goh, S. H., (2001), Soilpile Interaction During Liquefaction-Induced Lateral Spreading, PhD Thesis, Cornell University, Ithaca, NY. Goh, S.-H. and 0' Rouike, T. D., (1999), "Limit State Model for Soil-Pile Interaction During Lateral Spread," Proc. Seventh US.;lapan Workshop on Earthquake Resistant Design of Lifeline Facilities and Countermeasures Against Soil Liquefaction, Seattle, WA, August 15-17, Technical Report MCEER-99-0019, Multidisciplinary Center for Earthquake Engineering Research, University at Buffalo, (O'Rourke, Bardet and Hamada, eds.), pp. 237-260. Hamada, M., (1992), "Large Ground Deformations and their Effects on Lifelines: 1964 Niigata Earthquake," Ch. 3 of Case Studies of Liquefaction and Lifeline Pmjormance During Past Earthquakes, Vol. 1: Japanese Case Studies, (Hamada and O'Rourke, eds.), 3-1 to 3-123. Hamada, M., (2001), Personal Communication. Hamada, M. and O'Routke, T D., (eds.) (1992), "Case Studies of Liquefaction and Lifeline Performance During Past Earthquakes;' Vol. I:Japanese Case Studies, Technical Report NCEER-92-0001, National Center for Earthquake Engineering Research, University at Buffalo, February. Kallou, V.,Abdoun, T. H., Zeghal, M., Oskay, C. and Dobry, R., (2001), Visualization of Centrifuge Models of Lateral Spreading and Pile Bending Response (in preparJtion). Meyersohn, W D., (1994), Pile Response to Liquefaction Induced Lateral Spread, Ph.D. Thesis, Dept. of Civil and Environmental Engineering, Cornell University, Ithaca, New York. Meyersohn, W D., O'Rourke, T. D. and Miura, E, (1992), "LaterJI Spread Effects on Reinforced Concrete Pile Foundations," Proc. 5th US-japan Workshop on Earthquake Disaster Prevention for Lifeline Systems, Tsukuba, pp. 173-1 %. O'Rourke, T D. and Hamada, M., (eds.), (1992), "Case Studies of Liquefaction and Lifeline Performance During Past Earthquakes;' Vol. 2: United States Case Studies, Technical Report NCEER-92-0002, National Center for Earthquake Engineering Research, University at Buffalo, February. Ramos, R., (1999), Centrifuge Study of Bending Response of Pile Foundation to a Lateral Spread Including Restraining Effect of Superstructure, Ph.D. Thesis, Dept. of Civil Engineering, Rensselaer Polytechnic Institute, Troy, NY Tokimatsu, K., MiZlmo, H. and Kakurai, M., (1996), "Building Damage Associated with Geotechnical Problems," Soils and Foundations, pp. 219-234, January. Wang, Y., (2001), Evaluation of Pile Foundation Retrofitting Against Lateral Spreading and Inertial Effects During Liquefaction Using Centrifuge Models, MS Thesis, Dept. of Civil Engineering, Rensselaer Polytechnic Institute, Troy, NY. 100 Other Publications Abdoun, T.H., Dobry; R., O'Rourke, T.D. and Chaudhurl, D., (1996), "Centrifuge Modeling of Seismically-induced Lateral deformation During Liquefaction and its Effect on a Pile Foundation," Proc. Sh.:th Japan-US. Workshop on Earthquake Resistant Design of Lifeline Facilities and Countenneasures Against Soil Liquefaction, Technical Report NCEER-96-0012, Multidisciplinary Center for Earthquake Engineering Research, University at Buf:fulo, pp. 525c539. Abdoun, T.H. and Dobry; R, (1998), "Seismically Induced lateral Spreading ofTwo-layer Sand Deposit and its Effi:ct on Pile Foundations," Proc. Inti Con! Centrifuge'98 (T. Kimura, O. Kosakabe and]. Takemura, eds.), Tokyo, Japan, Sept. 23-25, Vol. 1, pp. 321-326. Dobry, R., (1994), "Foundation Defurmation Due to Earthquakes," Froc. ASCE Specialty Conference on Settlement '94, College Station, TX, June 16-18, pp. 1846-1863. Dobry; R, (1995), "Liquefaction and Defurmation of Soils and Foundations under Seismic Conditions," State-of-the-art paper, Proc. Third Intl. Conf on Recent Advances in Geotechnical Earthquake Engineering and Soil Dynamics (5. Prakash, ed.), St. Louis, MO,April 2-7, Vol. ill, pp. 1465-1490. Dobry, R.,Abdoun, T.H. and O'Rourke, T. D., (1996), "Evaluation of Pile Response Due to Liquefaction-Induced Lateral Spreading of the Ground," Froc. 4th Caltrans Seismic Research Workshop, Sacramento, CA, July, 10 pages. Dobry; R., Van Laak, P., EIgamal, A.-W, Zimmie, T. E and AdaIier, K, (1997), "The RPI Geotedmlcal Centrifuge Facility;" NCEER Bulletin,April, pp. 12-17. Elgamal, A.-W, Dobry; R., Van laak, P. and Nicolas-Font,]., (1991), "Design, Construction and Operation of 100 g-ton Centrifuge at RPI," Proc. Intl. Con! Centrifuge'91 (H.-y. Ko and E G. Mclean, eds.), pp. 27-34. 0' Rourke, T., Meyersohn, W D., Shiba, Y. and Chaudhurl, D., (1994), Evaluation of Pile Response to Liquefaction-Induced Lateral Spread, Tech. Rept. NCEER-94-0026, Multidisciplinary Center fur Earthquake Engineering Research, University at Buffulo, pp. 445-455. Ramos, R., Abdoun, T. H. and Dobry, R., (1999), "Centrifuge Modeling of Effect of Superstructure Stiffness on Pile Bending Moments Due to Lateral Spreading," Proc. Seventh IT S.-Workshop Workshop on Earthquake Resistant Design of Lifeline Facilities and COllnte17neasures Against S0l7 Liquefaction, Seattle, WA,August 15-17, 10 pages. Ramos, R.,Abdoun, T. H. and Dobry, R, (2000), "Effect of Lateral Stiffness of Superstructure 011 Bending Moments of Pile Foundation Due to Liquefaction-induced Lateral Spreadillg;" Froc. 12th Wo1"ld Con! on Earthquake Engineering, Auckland, New Zealand, Jan. 30 - Feb. 4, 8 pages. Taboada, Y.M. and Dobry; R, (1998), "Centrifuge Modeling of Earthquake-Induced Lateral Spreading in Sand," J. Geotechnical and Geoenl'ironmental Engineering, ASCE, Vol. 124, No.12, pp. 1195-1206. Van Laak, P., Eganlal, A.-W. and Dobry, R., (1994a), "Design and Performance of an EIectrohydraulic Shaker fur the RPI Centrifuge," Proc. Inti. Conference Centrifuge'94, Singapore, August 31-Sept. 2, (C.R Leung, RH. Lee and T.S. Tan, eds.), A.A. Balkema, Rotterdam, pp. 139-144. Van Laak, P., Taboada, Y., Dobry, R. and ElgamaI, A.-W, C1994b), "Earthquake Centrifuge Modeling Using a Laminar Box;' Dynamic Geotechnical Testing II CR. J. Ebelhar, V. P. Drnevich and B. L. Kutter; eds.),ASTM S1P 1231, pp. 370-384. Van Laak, P., Adalier; K., Dobry, R. and ElgamaI, A..-W, (1998), "Design of RPl's Large Servohydraulic Shaker; Proc. IntI. Con! Centrifuge'98 cr. Kimura, O. Kosakabe and J. Takernura, eds.), Tokyo, Japan, Sept. 23-25, VoL 1, pp. 105-110. Pile Foundation Response to Lateral Spreading and ivIitigation Strategies 101 Page Intentionally Left Blank Analysis and Design of Buildings with Added Energy Dissipation Systems by Michael C. COllstantinou, Gal]' E Dargush, George C. Lee (Coordinating Autb01~, Andrei it!. Reinhorn and Andrew S. Whittaker Research Objectives TIus article is a summary of the progress made during 2000-2001 by National Science Foundation, MCEER researchers working in the subtask of facilitating technologies. Earthquake Engineering These research progresses should be viewed from the context that the Research Centers Program long term (3-4 year) objective is to complete an MCEER monograph on Analysis and Design of Buildings with Added Ene1'gy DiSSipation Sys tems. Although each individual researcher is advancing the state-of-the art knowledge with rus respective graduate students and research Michael C. Constantinoll, collaborators, it is the total systems integrated effort directed toward the Professor and Chair, practicing professional that underpins the projects witllin this group. TIus Department ofCh'll, undertaking is possible by using the "Center Approach" in earthquake Structural and Environmental engineering research. Engineering, University at Buffaln, Panos Tsopelas, Assistant Professor, Catholic University of CEER's research program 2 on the seismic retrofit of hospitals concen America, Wilbelm M trates. on developing cost-effective retrofit strategies for critical fa Hammel, Senior cilities using new and emerging materials and enabling technologies. Consultant, KBJJG, Current emphasis is given to hospitals that should remain functional dur German); andAni N. ing and immediately after damaging and/or destructive earthquakes. Sigaher, Ph.D. candidate, Department ofChi!, The major disciplinary components of this program are geotechnical, Structural and structural, nonstructural, advanced technologies and rugh performance Environmental materials, socio-economic issues and systems integration. One important Engineering, University at research task concerning the methods of analysis and design of buildings Buffalo with added emerging materials (e.g., composite infill walls) andlor en abling technologies (e.g., damping devices) is carried out by a group of Gary E Dargnsb, Associate research investigators under the title "FacilitatingTechnologies." T11is task Professor. Enmesh San!, Ph.D. Candidate, and is the heart of the systems integrated approach concerning the perfor Enjesb Rndhakrislman, mance of a system (i.e., the performance of hospital buildings and con Graduate Student, tents with added earthquake protective systems so that the medical Department ofCivil, functions can be carried out). To develop retrofit strategies for buildings Structural and by using added materials and/or enabling devices, practicing engineers Environmental need to know how to choose appropriate technologies to satisfy building Engineering, Unit'ersityat performance indices andlor objectives cost-effectively for given earthquake Buffaln risk. In view of FEMA 273/274 and NEHRP 2000, which encourage the Research Team continued on page 2 103 use of emerging technologies to and nonstructural components. achieve building performance ob MCEER researchers have contrib jectives, there is a need to develop uted to FEMA 273/274 in two ar principles and design guidelines for eas: modeling and analysis, and George C. Lee, Director, these engineers. This is the funda seismic isolation and damping sys MCEER, and Samuel P. tems (see Constantinou et aI., 1998 Capen Professor of mental rationale of the research Engineering, Department task on facilitating technologies. andTsopelas et aI., 1997). ofCivil, StnJctural and MCEER investigators have made The 2000 NEHRP Recommended Environmental many key contributions to the two Provisions for Seismic Regulations Engineering, University at most advanced codes and guide for New Buildings and Other Struc Buffalo, lJ-fai Tong, Senior lines related to the implementation tures includes robust procedures for Research Scientist, MCEER of passive energy dissipation sys the analysis and design of passive en Yihui Wu, Research tems: FEMA 273/274 Guidelines ergy diSSipation systems (Appendix Scientist, MCEER,and Wei Liu, PIl.D. candidate, for the Seismic Rehabilitation of to Chapter 13) using force-based Department ofCivil, Buildings, published in 1998, and methods of analysis that are consis StnJdural and the NEHRP 2000 Guidelines for tent with those methods used for the Environmental Seismic Regulations for New analysis and design of conventional Engineering, University at Buildings and Other Structures, construction. The development of Buffalo that will be published in the next force-based analysis and component few months. checking methods for highly nonlin Andrei Reinhorn, Professor, Department ofChil, The FEMA 273 Guidelines and ear or velocity-dependent energy Structural and 274 Commentary represented the dissipation devices proved to be a Environmental culmination of more than a decade most demanding task. MCEER re Engineering, University at of work funded by the Federal searchers have developed the tech Buffalo, Katrin Emergency Management Agency, nical underpinnings of the methods Winkelmann, Visiting the National Science Foundation presented in NEHRP 2000 (see Researcll Associate, and other agencies. These docu Ramirez et al., 2000). Department ofCMI, StnJctural and ments provide structural engineers Since the publication of FEMA Environmental with new information on proce 273/274 in 1998, many practicing Engineering, University at dures for the analysis, evaluation, engineers have been interested in Buffalo from tbe University and design of existing and retrofit using these guidelines to retrofit ofDannst£ldt and construction. Information is also existing buildings with base isola Mettupalayam presented on performance-based tion and/or energy dissipation de Sivaselvan, Ph.D. earthquake engineering, modeling vices. As a result, several major candidate, Department of Civil, Structural and and analysis, steel, concrete, and retrofit projects have been com Environmental timber structures, foundations, pleted and their results publicized, Engineering, University at Buffalo Andrew S. Whittaker, Associate Professor, Oscar ~~~•• ~~~:~= .. RJlmirez, Graduate howfo s~lectme.lfe~fp~~~1'~r~n~iiit4is~ip~ti0~ d~iCe sys Student, andJuan Gomez, temfo1:".~(~v~~.if~·9~~9P~g(~·i;~~'$~iiS~~)ow-mediiun Graduate Student, height, s!eeI,~~WfoK~d::cc5ntt .<." ., . ·O( "',tltDber, rela Department ofCivil, tive "... ·s··j····etrli;;hi ':'i'\;ki;";~):·;:.,'>~:rrhlid for '. en StnJdural and .. eartZq;:f~tty~' .,'·@j~~~~~~~drit¢sT~ucli a:;:ow Environmental to a struc Engineering, University at optimallydistrl~ute datr.ptii~ll~~¢~stJ:#:~ughout ture and how co~ider niUiti~pet'(tjtmance'llidices are also Buffalo to of critical'importance~;{~ ..... 104 particularly in the area of base iso is considerably more complicated lation systems, in which MCEER than to model or to test the behav researchers (most notably, M. ior of a single device. Constantinou and A. Whittaker) First, existing buildings them have acted as consultants. Because selves have different dynamic char • MartinJobnson, Group base isolation technology has a acteristics and are complicated to Manager,EQE longer history of practical imple model. The performance of the International mentation for buildings and devices cannot be generalized bridges, more projects have been based on one simple structure. • Ali Karakaplan, President, IARSA, Inc. successfully completed. However, Many multiple degree of freedom only limited reports are available in (MDOF) systems cannot be treated • Charles Kircher, the open literature on passive en as single DOF systems (decoupling Principal, Charles Kircher ergy dissipation devices applied to assumption) with sufficient accu and Associates buildings. This is primarily due to racy. To develop Simplified finite the relatively short history in ex element (FE) analysis models for • Douglas Taylor, perience of practical applications complicated MDOF systems is itself President, Taylor Devices, Inc. and the fact that many available a challenge. devices/approaches are reported in When energy dissipation devices • u.s. Department of research publications, many devel are implemented in these MDOF Commerce oped and/or improved by MCEER structures, the total building-device researchers. systems are generally nonlinear sys During the past two decades, tems. Much creativity and funda many advances have been reported mental research in structural on the performance and vibration dynamics principles have to be pur reduction properties of passive sued to develop a reasonably energy devices. As noted earlier, simple and accurate analysis and MCEER researchers have played a design procedure. The current, significant role in the development MCEER studies may be grouped of a variety of devices. This "device into three separate categories. based" line of pursuit will be con Type I Projects: New Devices and tinued as new ideas, materials and/ Systems -continued development or devices become available. of new ideas in passive energy dis Anew systems-based approach is sipation devices and semi-active now being undertaken to provide systems. The semi-active systems answers to the many questions on can extend the range of traditional choosing appropriate devices. Re passive energy dissipation systems garded as the "building-device sys and can be combined to provide tem based" consideration, these an added fail-safe feature to these studies emphasize the perfor systems. mance of the building with added Type n Projects: MDOF Modeling passive energy dissipation devices. ofBuilding-Device Systems-This cat This new systems-based approach egory of studies is concerned with is the central theme of the research the behavior and design of the task on facilitating technologies, building-device system that must and this paper focuses on these be studied by using multiple DOF new approaches. models, including when the sys To develop analytical models or tems are nonproportionally to carry out experimental observa damped. These studies provide tions for the systems-based study quantitative information on the er- Analysis and Design ofBuildings witlJ Added Energy Dissipation Systems 105 rors when such buildings are ap which offer certain advantages,either proximated by decoupled single in terms of cost of the energy dissi DOF systems. pation devices, or in terms of archi Program 2: Seismic Retrofit of Type III Projects: Analysis and De tectural considerations such as open Hospitals sign Software-MCEER will con space requirements. Particularly, stiff tinue its development of analysis structural systems under seismic • Task 2.4a Techno/{)gy and design software for buildings load or structural systems under Portfolio - Structural with added energy dissipation sys wind load undergo small drift and tems. Some projects are fundamen the required damping forces are tal in nature to establish new large. This typically results in larger approaches while others empha damping devices and accordingly, size the development of user greater cost. In other cases, energy friendly simplified procedures for dissipation devices cannot be used design professionals with accept in certain areas due to open space able accuracy. requirements and the ineffective The findings from research in ness of damping systems when these three categories will culmi installed at near-vertical configura nate in a monograph on Analysis tions. and Design of Buildings with Two recently developed configu Added Energy Dissipation System rations, the toggle-brace and the for the design professional. The scissor-jack energy dissipation sys following sections provide brief de tem configurations, offer advan sCriptions of the progress made in tages that overcome these current projects. limitations. Both utilize innovative mechanisms to amplify displace ment and accordingly lower force Scissor-Jack Seismic demand in the energy diSSipation Energy Dissipation devices. However, they are more System complex in their application since they require more care in their (Type I Project) analysis and detailing. The theory Energy dissipation systems are and development of these systems being employed in the United has been described in States to provide enhanced protec Constantinou et al. (2001) and tion for new and retrofit building Constantinou and Sigaher (2000). and bridge construction. The hard This section briefly presents these ware utilized includes yielding steel new configurations and compares devices, friction devices, viscoelas them with the familiar chevron tic solid devices and mostly, so far, brace and diagonal configurations. viscous fluid devices. The toggle-brace and scissor-jack Engineers are familiar with and systems are configurations for mag have extensively used diagonal and nifying the damper displacement chevron brace configurations for the so that sufficient energy is dissi delivery of forces from energy dissi pated with a reduced requirement pation devices to the structural for damper force. Conversely, they frame (Soong and Dargush, 1997; may be viewed as systems for mag Constantinou et at., 1998). New con nifying the damper force through figurations have been developed shallow truss configurations and 106 then delivery of the magnified and diagonal configurations have f force to the structural frame. less than or equal to unity. Figure 1 illustrates various For the purpose of comparison, MCEER Users NetllJork: damper configurations in a framing consider the case of the use of a http://cilJileng.buffalo.edu/ system. Let the interstory drift be linear viscous damper with C =160 users_ntwk o U, the damper relative displace kN-s/m (= 0.9 kip-s/in) in the fram- ment be u ' the force along the axis ing systems of Figure 1 with weight mARC Users Group: D http://citrileng.buffolo.edu of the damper be ~ and the damp W = 1370 kN (= 308 kip) andT = ing force exerted on the frame be 0.3 second. The resulting damping F. It may be shown that ratios are shown in Figure 1. The effectiveness of the toggle-brace (1) and Scissor-jack systems is clearly (2)" where f = magnification factor. Expressions for the magnification /=cos6 factor ofvarious configurations are shown in Figure 1. The significance of the magnification factor may be best demonstrated in the case of /=1.00 linear viscous dampers, for which (3) whereun = relative velocity be tween the ends of the damper along the axis of the damper. The damping ratio under elastic condi tions for a single-story frame (as shown in Figure 1) with weight,w, _ ___sin-'-'6-=2_ . 6 / +sm 1 and fundamental period,T, is: cos(61 +62 ) f3= Cof2gT (4) 47rW That is, the damping ratio is pro portional to the square of the mag nification factor. The toggle-brace and SCissor-jack systems can achieve magnification factors larger than unity. The systems can be typically configured to have val ues f = 2 to 3 without any signifi cant sensitivity to changes in the geometry of the system. By con trast, the familiar chevron-brace • Figure 1. Effectiveness of Damper Configurations in Framing Systems Analysis and Design of Buildings with Added Energy Dissipation Systems 10-/ tects. As an example, Figure 2 illus trates the scissor-jack system tested at the University at Buffalo. The open bay configuration, the slen derness of the system and the small size of the damper are apparent. Damping Ratio as a Seismic Response Re duction Measure of Non-Proportionally Damped Structures (Type II Project) In FEMA 2731274 (1997), an im portant design parameter for added • Figure 2. Tested Scissor-Jack Damper Configura devices is the effective damping tion ratio. As suggested in FEMA 273/ 274, displacements are reduced as the effective damping ratio is in demonstrated. It should be noted creased. Some believe that placing that the configurations for these more dampers at the level of maxi two systems are identical to those mum inter-story drift will achieve tested at the University at Buffalo. optimized damping effects. This is It is clear in the results of Figure 1 true for proportionally damped or and in equations (1), (2) and (4) that single degree of freedom (SDOF) the toggle-brace and scissor-jack con systems. It has been shown by Lee figurations may provide substantial et al., (2001) that for mUltiple de energy dissipation capability with the gree of freedom (MDOF) systems, use oflow output force devices. This higher damping ratios could, in may result in an important cost ad some cases,increase the seismic re vantage in systems that undergo sponse. For structures with added small drifts such as stiff structural passive energy dissipation or seis systems under seismic load and most mic isolation devices, the damping stmctural systems under wind load. is no longer proportional nor neg Such cases of small drift lead to a re ligibly small.To study the damping quirement for increased volume of effects on MDOF building struc fluid viscous devices and accordingly tures, the analysis method should increased cost. The use of the new consider the effects of non-propor configurations eliminates the neces tional damping. sity for large volume damping de For non-proportionally damped vices and may result in reduced cost. MDOF systems, the damped mode Moreover, the scissor-jack system shape is not orthogonal to the may be configured to allow for damping matrix in the n-dimensional open space, minimal obstruction of physical domain.The complex mode view and slender configuration, shapes are orthogonal in the state which are often desired by archi- space domain, where real valued 108 modal superposition methods can not be directly applied.The "double ,e results where u: and u~ are real modal superposition" approach de from mDdeling value response associated with the veloped by Gupta and law (1986) is modal shape, as defined by equa- and an~lysis of used in this analysis, where the re a d CR-lifo-rnia sponse is the superposition of'modal tion (8), and ljf i and ljf: are dis hospital displacement" and "modal velocity:' placement and velocity mode builiUng will These modes are nottheun-damped shapes, respectively. They are real p-rGvide the system modes, nor the complex sys value mode shapes and are calcu enginee':t'ing tem modes.The conventional modal lated by the state vector eigen-equa analysis routine of fast calculation tion. These mode shapes are cDmmu-nity witb can be used in this approach and the determined not only by the system a tbrec damping effects can be more readily mass and stiffness but also by the dimensional explained by using structural dy added damping matrix. In some n8-nUneR? namic parameters. cases, the added damping will anaJysis dominate the mode shapes (as plR-ifo1l"m tbR-f Theoretical Background when the fundamental modal t:u17enfly does shape collapse due to added damp not exist." In a proportionally damped sys ing.). X. is the modal response of I tem, the damping ratio is used to the following "modal" equation. describe damping effects. However, In equation (8), ~i and illi are the damping can affect the response of "modal damping ratio and circular an MDOF non-proportionally frequency of mode i", and ug is the damped system in many ways, in ground motion. cluding modal response, modal shape and natural frequencies, and Xi +2~iO)iXi+0);Xi=-Ug (8) damping ratio. The damping ratio and natural For an MDOF system subjected circular frequency ~i' ill; in the to earthquake excitation, the equa above equation are determined by tion of motion can be written as: the complex state vector eigen value solution,like the mode shape. MU+CU+KU=-MUbUg (5) Because the state vector will vary where M, C and K denote mass, for different damping devices damping and stiffness matrices, re added, the modal shape, circular spectively; U is the relative displace frequency and damping ratio will ment vector; Ub is a displacement not be constant. Thus, the system vector obtained by statistically dis response defined by equation (6) placing the support by unity in the cannot be always reduced by in direction of the input motion; and creasing the damping ratio. u g is the ground displacement. The double modal superposition method obtains the response by Case Studies equation (6): To illustrate how system charac N teristics change and their impact on U = LUi, Ui = U~ - U: (6) seismic response due to added I damping, a four DOF frame struc ture was analyzed. The first two vi- Analysis and Design ofBuildings with Added Energy Dissipation Systems 109 added device plus 2%). As listed in Table 1, dampers were installed on every floor independently, every Model _I two floors, every three floors or on 3 J all four floors. There were 13 dif Mode 2 / ferent damper configurations, with (; four damping ratios. 02 u:: As noted above, the fIrst mode of J the complex damping ratio may be I different from the first mode of the controlling effective damping ratio. The fIrst complex modal damping OL-----~----~----~----~ ratio for all 52 cases is listed inTable .1 ·0.5 0 0.5 2. For small damping ratios (5%), the complex damping ratio is al most the same, however; when the • Figure 3. First 1\\'0 Modal Shapes damping ratio is increased to 10%, case 4 shows a dramatic change, bration modes of the frame are while the other cases stay the same. shown in Figure 3. Cases 3,4,7,10, and 12 have very The added damping was limited different values when the control to a maximum of 30% of the effec ling ratio increases to 20%. Finally, tive damping ratio (defined in the fIrst complex damping ratio for FEMA 273 as a proportional ratio). cases 1, 3,4,7 and 12 becomes 99% Fifty-two linear viscous damper when the ratio is increased to 30%. configurations were examined.The The change in natural frequency 52 configurations were arranged by in the first mode is given in Table considering all possible damper 3. Both damping ratio and fre locations and their combinations, quency changes are related to the then changing the damping param modal change for different damper eters to make the fIrst modal effec configurations. tive damping ratio 5%, 10%, 20% To illustrate the non-proportional and 30% (damping ratio by the mode shape change, the first dis- • Table 1. Damper Location • Table 2. Damping Ratio Comparison (First mode) 110 iii Table 3. Natural Frequency Comparison (First mode) (in Hz) i'CaSe' \1 :' 2: 3,< I 4< '5'" ;.61.;'7' !A3 .\ 9;"·1 1{}" !11 ! 12: \13 5% 2.03 2.02 2.03 2.05 2.02 2.02 2.03 2.02 2.02 2.02 2.02 2.03 2.02 10% 2.06 2.04 2.07 1.17 2.03 2.07 2.02 2.03 2.03 2.03 2.03 2.05 2.02 20% 2.22 2.12 2.24 0.50 2.06 2.07 2.23 2.03 2.07 2.08 2.09 2.14 2.03 30% 1.96 2.29 1.50 0.32 2.12 2.15 1.41 2.04 2.14 2.17 2.20 1.30 2.04 placement mode shape at the con ing is shown in Figure 6. Except trolling effective damping ratios of for case 4, the response decreases 5%, 10%, 20% and 30% are shown as the damping ratio increases. in Figure 4. The figure shows that, When the controlling damping ra as the damping ratio becomes tio is higher than 15%, different higher, dramatic changes can result damper configurations will result in cases 1, 3, 4, 7 and 12 (the funda in different response reductions mental mode collapsed).When the even though they may achieve the fundamental mode shape collapses, same effective damping ratio value. the second mode shape becomes It may also be seen from the figure dominant in the system response, that the maximum responses with as shown in Figure 5. a 30% effective controlling damp The relationship between the ing ratio for some cases are higher maximum displacement response than those with 20%. Table 4 lists and the controlling effective damp- the maximum displacement re- , ..:--case4 r ci 3 , C , '" ~'" 2 00 02 0.4 06 0.8 t2 1.4 a) Damping =5% ci ci c c '"en 2 enf1l 2 ::zal ::zal Displacement Displacement c) Damping = 20% d) Damping = 30% iii Figure 4. Displacement Modal Shapes Analysis and Design ofBuildings witb Added Energy Dissipation Systems 111 sponses for these two damping ra tio values, calculated for the El Centro earthquake. As seen in the 1st mode 1017Hz, complex damping ratio=99% table, the damping increases in two ways: the first is where the re sponse increased after more damp ers were added and the configuration was unchanged 2nd mode 2.09Hz, complex (cases 3 and 4). The second case is damping ratio=4% when more dampers are added in different locations and configura o~--~--~--~--~----~--~~ -0.2 o 0.2 0.4 0.6 0.8 1.2 tions, as shown in case 1 at 20%, Displacement case 3 at 30%, case 4 at 30% and a) Case 4, 10% Damping Ratio the rest of the cases are at 20%.This conclusion has also been obtained 1st mode 1.41 Hz, complex using other earthquake records, damping=99% such as from the Northridge earth quake. Continuing Effort It is important to understand O~----~----~----~------~--~ o 0.2 0.4 0.6 O.B the limitations of using the damp Displacement ing ratio as the key seismic re b) Case 7,30% Damping Ratio sponse reduction measurement. For many non-proportionally 4.5 damped structures, the evaluation 1st mode 1.30Hz, complex damping=99% of performance should be based on response history analysis. In o 3.5 C this regard, a simple method to III 3 cope with the various non-pro :2'" 2.5 portional damping effects will be 2nd mode 2.37Hz, complex damping=21% valuable for engineering applica 1.5 tions. This team is currently at -h'=-2 --~---='02=---O.4'---O.6:---0:":.B:--~--""'t~.2---,J'.4 tempting to develop such a Displacement method. c) Case 12, 30% Damping Ratio • Figure 5. Mode Collapse • Table 4. Maximum Displacement Response at the Top Level of the Frame Using Standard El Centro Earthquake Record 112 Maximum displacement and damper configuration 0.9,----~--~--~--_=__..-~- 0.8 case 4 O]w.-___- L E ~ 0.0 oCD 025~--~1O--~15~-~~~--~~--~30 Damping 1% II Figure 6. Maximum Displacement Response structure to be non-proportionally Optimal Design of damped, a "systems approach" con Damping Devices for sidering the change in characteris Multi-story Steel tics is necessary. The earthquake ground accelera Frru11'es Based on tion can be approximated as a fi Multi-performance nite Fourier series expansion, and Indices the seismic response as the linear combination ofall responses to the (Typie III Project) single frequency exdtation. Thus, to fuid the effective configuration, FEiVIA 273, NEHRP 2000 and the it is necessary to know the domi Blue Book 1999 provide primary nant frequency response (or domi design guidelines for buildings nant mode), the modal with supplemental damping de composition, and its mass partid vices. These procedures can ad pating factors. The selection of the dress building design for different response will follow the perfor performance objectives using lin mance index. With this knowledge, ear or nonlinear methods after the an optimization scheme for damp damping device configuration has ing devices can be developed. been determined. However, the rule or procedure of how to opti mally distribute these damping de Analysis and Design Procedure vices throughout the building and The design procedure is outlined how to consider multi-perfor as follows: mance indexes is still not available. These procedures are very critical 1. Calculate the characteristics of for practicing engineers. Since add each potential configuration ing damping devices causes the Analysis and Design of Buildings with Added Energy Dissipation Systems 113 illustrates the modes that are im 2. Compare these with the origi portant in the response of a given nal structural system's character mass pOint. istics, and identify for each 5. The damping distribution may configuration: be very different if the optimi a. Natural frequencies (Some zation target is selected as the modes may collapse into each story drift or acceleration. The other due to added damping. drift response is dominated by This is particularly critical the major modes. Thus, with a when damping becomes non higher modal participation fac proportional. It follows that tor and a higher damping ratio, the natural frequencies may the drift responses will be fur change significantly.) ther reduced.This is not the case b. Modal damping ratio (100% for acceleration response, which critical damping usually im includes more higher mode con plies a mode collapsed, which tributions. may be a problem for modal 6. For acceleration response, if a response reduction) damper is placed between the c. Mode shapes th i and i+ 1th floors, then the i+ 1th d. Modal loading factors (varia floor acceleration is usually tion in this characteristics may lower than the ith floor. In addi strongly influence the abso tion, i-I th floor acceleration will lute acceleration response) be lower than that of ith floor. e.Phase differences between the modes. 3. For a mass point of interest, Case Study check which natural frequen An II-story steel frame building cies will contribute most in the structure was used as an optimal response, by using sinusoidal ex design example. The building's citations of different amplitudes. typical plan and north-south eleva Examine each corresponding tion are shown in Figures 7 and 8. steady state response vector to This building had been designed identify the largest contribution with supplemental damping de vector. A general excitation can vices. The typical device configu be Simplified as a sum of this si ration in the frame is shown in nusoidal excitation of various Figure 9. A total of 120 dampers amplitudes, multiplied by cer were added to the building, with tain envelopes. Thus, the re 24 dampers on the first story, and sponse will be dominated by the gradually decreasing as the story combination of n basic vectors. height increases. 4 The vectors computed above can If the performance index is se be further split into modal contri lected as optimal story drift, the butions. In this regard, compari first story mass contributes less to sons can be made among modal the dominant mode of the system shapes,loading factors,natural fre response. Better results may be quencies (which provides infor obtained if the dampers in this mation on relative phase lags) and story are redistributed. The opti damping ratios (which provide in mized distribution removes most of formation on dynamic amplifica the devices from the first story, and tion factors). This information 114 1-- f-H . III Ilg . , , != f- f-H II Figure 7. Floor Frame PIan II Figure 8. Elevation Frame PIan II Figure 9. Damper Distribution adds them to the 6 th to 11th stories. Computational The same total number of damp ers is used as in the original damper Aseismic Design and distribution. With this optimized Retrofit for Passively distribution, the total system damp Damped Structures ing is increased by 4%. More im portantly, the damping ratio of the (Type III Project) dominant mode is increased by Over the past two decades, con 19%, and the largest story drift, siderable effort has been directed which occurred in the 6th floor, is toward the development and en reduced by an average of 11% for a hancement of protective systems series of spectra-compatible earth for the control of structures under quake. If the performance index is seismic excitation. In the area of selected as acceleration, the opti passive energy dissipation systems, mal distribution will be different. applications typically involve me tallic yielding dampers, friction Current Efforts dampers, viscous fluid dampers or viscoelastic dampers (e.g., Soong The design and analysis proce and Dargush, 1997; Constantinou dure proposed here is based on fre et. al., 1998). Although the intro quency domain analysis and time duction of these new concepts and domain verification. Since response systems presents the structural en spectrum or ,time history analysis gineer with additional freedom in are preferred in structural design, the design process, many questions the method is currently being im also naturally arise. In the case of proved and optimized using re passive energy dissipation systems, sponse spectrum-based objective these questions range from perfor functions. mance and durability issues to con cerns related to the sizing and placement of damping elements. One promising direction for fu ture research involves the further development ofthe FEMA 273/274 Analysis and Design ofBuildings with Added Energy Dissipation Systems 115 and NEHRP 2000 design guidelines Holland (1962, 1992) also devel based upon additional numerical oped a unified theory of adaptation simulations and practical experi in both natural and artificial sys ence. Alternatively, one may envi tems. In particular, Holland brought sion a dramatically different design ideas from biological evolution to process for passively damped struc bear on the problem. Besides pro hIres by adopting a computational viding a general formalism for approach. Such an approach studying adaptive systems, this led should incorporate the dynamics of to the development of genetic al the problem, the uncertainty of the gorithms. seismic environment, the reliability of the passive elements and per Computational Framework for haps also some key socioeconomic factors. With these requirements Aseismic Design in mind, one can conceptualize With these ideas, one can now aseismic design as a complex adap envision a new aseismic design tive system and begin to develop a approach based upon the creation general computational framework of an artificial complex adaptive that promotes the evolution of ro system. The primary research ob bust, and possibly innovative, de jective is to develop an automated signs. system that can evolve robust de signs under uncertain seismic en Complex Adaptive Systems vironments. With continued development, the system may also There is a broad class of systems be able to provide some novel so in nature and in human affairs that lutions to a range of complex involve the complicated interac aseismic design problems. tion of many components or agents. Figure 10 depicts the overall ap These may be classified as complex proach for computational aseismic systems, particularly when the in design and retrofit (CADR), borrow teractions are predominantly non ing terminology from biological linear. Within this class are systems evolution. DeSign involves a se whose agents tend to aggregate in quence of generations within a se a hierarchical manner in response quence of eras. In each generation, to an uncertain or changing envi a population of individual struc ronment. These systems have the tures is defined and evaluated in ability to evolve over time and to response to ground motion realiza self-organize. In some cases, the tions. Cost and performance are system may acquire collective used to evaluate the fitness, which properties through adaptation that in tum determines the makeup of cannot be exhibited by individual the next generation of structures. agents acting alone. Key character Performance is judged by perform istics of these complex adaptive ing nonlinear transient dynamic systems are nonlinearity, aggrega analysis. Presently, this analysis uti tion, flows and diversity (Holland, lizes either ABAQUS (2000) or an 1995). Examples include the hu explicit state-space transient dy man central nervous system, the namics code (tda). The implemen local economy, a rain forest or a tation of the genetic algorithm multidisciplinary research center. 116 Computational Aseismic Design and Retrofit Evaluate Era Evaluate Generation SEQUENTIALt Evaluate Individual Structure Realize Define Structure Ground Mation{s) / ~ I Evaluate Performance Determine Cost (abaqus arlda) M~=t ~ Determine Fitness II Figure 10. Overall Framework for Computational Aseismic Design and Retrofit controlling the design evolution is later in the design process. Cur accomplished within the public rently, a two-surface cyclic plastic domain code Sugal (Hunter, 1995). ity model is applied for the primary structural system and metallic plate Model Problem: Five-Story dampers, while a coupled thermoviscoeIastic model with in Steel Moment Frame elastic heat generation is used for Consider an example of a five the viscoelastic dampers. Both story steel moment frame retrofit interstory drift and story accelera with passive energy dissipators as tion limits are set in order to estab shown in Figure 11. Three differ lish acceptable performance. ent types of dampers are available: As a specific example, consider metallic plate dampers, linear vis the application of the CADR strat cous dampers, and viscoelastic egy to a typical five-story steel mo dampers. For each type, five dif ment frame based upon Era 1 (i.e., ferent sizes are possible. Conse quently, a 20-bit genetic code is employed to completely specify the dampers used in each story of Computational Aseismic Design and Retrofit any particular structureAE a . Thus, Example: Five-story steel moment frame retrofit w/passive energy dissipators for this problem, the set a contains Problem Definition: 220 (Le., more than one million) possible structures. Figure 11 also defines a hierarchical approach in which different structural models with varying levels of complexity are utilized in each era. The idea is GenelicCode :W-bit design Descriptor 00 NoneiSizs A 00 Size B :xx ;:: 01 Metalh-c to first use simple models to widely XYW XYW XYW ""YY XYW 10 Viscous Yf=~~~~:g Story: 1 2 3 4 5 llVE 11 Size E explore the design space and then to employ more complicated and computationally expensive models II Figure 11. Problem Definition for Five-story Steel Moment Frame Analysis and Design ofBuildings witb Added Energy DisSipation Systems 11-/ lumped parameter) simulations. (tpea) dampers and viscoelastic Let kjand ~ represent the f'story (ve) dampers are available and the elastic stiffness and story weight, 20-bit genetic code defmed in Fig respectively. The baseline frame ure 11 is applied. Hypothetical model has uniform story weights device cost data for various size Wi =W=125 kips fori =1,2, .. .5 and dampers were set as indicated in story stiffness kl=Is=k) =193 kip/in, Table 5. Each increment in damper k4 =147kip/in,ks =87 kip/in. The size corresponds roughly to a dou first two natural frequencies are bling of the damping capacity. 1.07 Hz and 2.72 Hz. A two-surface For the automated design, a popu cyclic plasticity model is employed lation of N p = 40 individual struc- to represent the hysteretic behav tures was evolved for a total of ior of the primary stmcture. Ng = 40 generations. Within each A retrofit strategy is now devel generation, each structure was sub oped to protect this structure situ jected to a total of Ns = 10 seismic ated on firm soil in a simplified events. Crossover and mutation hypothetical seismic environment operators were used to evolve new that can be represented by a uni stntctures from an initially random form distribution of earthquakes pool. At the end of each genera with magnitude 7.2 ~ Mw ~ 7.8 tion, one-half of the structures were and epicentral distance replaced with potentially new indi 20 km ~ r ~ ·30km. Each ground viduals. As generations pass, gener motion realization is generated ac ally speaking, the average fitness cording to the model of increases, indicating that the popu Papageorgiou (2000) for eastern lation becomes enriched with more U.S. earthquakes. For the retrofit, robust structures. However, the evo it is assumed that linear viscous lution of average fitness is not mono (visc) dampers, metallic yielding tonic, because the genetic algorithm continues to explore the design space for better structures. Table 5 • Table 5. Five-Story Steel Moment Frame -Baseline (Case 1) also presents the five structures that have appeared most frequently in the population. These are high fitness designs that have survived over many generations. The table data includes the total number of earthquakes that each of the five structures has expe rienced and the success (or survival) rate. Notice, according to Table 5, that the high fitness designs most often utilize viscous dampers and that the largest dampers are placed on the flfst story. In four of the high fitness designs, size C dampers ap pear in the fourth story, suggesting perhaps that the second mode re sponse also requires damping. 118 Concluding Remarks simple and accurate analysis and design procedure for use by the If;;ucb' ... In this research, a new computa practicing professionals. Two on tional aseismic design and retrofit creg,tivityand going MCEER projects are de (CADR) approach is advocated. fundamental scribed in the following sections. This approach centers on the de -reseR-rch in One is fundamental in nature to velopment of an artificial complex st-ructu,?al establish new approaches while adaptive system within which ro the other emphasizes the develop dynRmics bust aseismic designs may evolve. ment of user-friendly simplified principles bn.ve AB a first phase of this research pro procedures for the design profes to be pursued to gram, a genetic algorithm is applied sionals. develop a for the discrete optimization of a simple and passively damped structural system, accn-ra'te subjected to an uncertain seismic Nonlinear Structural Analysis anaJysis and environment. The results of pre by the State Space Approach liminary applications,involving the design seisniic retrofit of multi-story steel The State Space Approach (SSA) p?l!Jcedu-re fOT moment frames, suggest that con is an alternative approach to the p?;PJ.cticing tinued development of the ap formulation and solution of initlal pTDfessionals." proach may prove beneficial to the boundary-value problems involving > engineering community. Current nonlinear distributed-parameter efforts are underway to work with structural systems.The response of several MCEER Industry Partners to the structure, which is spatially enhance the CADR software and to discretized following a weak formu develop applications associated lation, is completely characterized with critical facilities. by a set of state variables.These in clude global quantities such as nodal displacements and velocities Development of and element (or local) quantities Analysis Tools for such as nodal forces and strains at the integration POints.The nonlin Engineering ear evolution of the global state Community variables during the response of (Type III Project) structures is governed by physical principles, such as momentum bal Any implementation of protec ance, and the nonlinear inelastic tive systems in design ofnew build evolution of the local variables is ings or bridges, or in their retrofit, governed by constitutive behavior. requires modeling and analysis of The essence of the SSA is to solve integral systems including the the two sets of evolution equations structures and the devices. When simultaneously in time using direct these multiple-DOF systems are numerical methods, in general as a implemented with energy dissipa system of differential-algebraic tion devices, the total building-de equations. The proposed method vice system in general is nonlinear. ology results in a more consistent Much creativity and fundamental formulation with a clear distinction research in structural dynamics between spatial and temporal principles have to be pursued in discretization. order to develop a reasonably Analysis and Design of Buildings with Added Energy Dissipation Systems 119 Objectives and results which allows an elastic beam to be A material nonlinear three-di bent into a circle. Figure 14 shows mensional beam column element the collapse pattern of a simple and a fully geometric (equilibrium structure while Figure 15 shows and nonlinear strain-deformation that all the models are developed relations) and material nonlinear using the macro model approach two-dimensional beam-column el in which structures are represented ement have been developed in this by beam-column elements with framework based on a flexibility hysteretic degradation. formulation. A general three-dimen sional interactive constitutive Three Dimensional Inelastic macro-model has been developed. In this model, hysteretic degrada Dynamic Analysis of Structures tion can also be modeled using suit with Protective Systems: able constitutive equations IDARC3D Version 2.0 (Sivaselvan and Reinhorn, 2000). The nonlinear analysis of inelas The reSUlting platform can study tic structures with energy dissipa structures near collapse. The basic tion systems and base isolations approach has been used to model was the subject of research and a structure, which collapsed in development throughout the exist shake table under severe lateral ence of the Center's activities. The buckling (see Vian et al.,in this vol research work, both analytical and ume). experimental, resulted in a series The above models and solution computer platforms, IDARC and procedure have been implemented 3DBASIS, now available nationally in an object-oriented computer and internationally to the public at program that uses the graphical large through a dedicated Users user interface (GUI) of the com Group (bttp://civileng.buffa/o.edu). mercial structural analysis program, (See also Park et.aI., 1987, Reinhorn LARSA. Figure 12 shows the re et. aI., 1988, Kunnath et. aI., 1989, sponse of a three-dimensional Nagarajaiah et.al., 1989,Nagarajaiah frame with hysteretic behavior to et. aI., 1991, Tsopelas et. aI., 1991, bi-axial non-proportional loading. Kunnath et. aI., 1992, Nagarajaiah Figure 13 shows results of analysis et. aI., 1993, Tsopelas et. aI., 1994, of extremely large deformations, W8<1' .. lIn! Ux~nJ u.pn) typo (b) Non-proportional (c) Force-dlsp. In Long. Olr. (d) Force-dlsp. In Trans. Olr. Excitation History (a) 0 Frame • Figure 12. Inelastic Response of Frame Using 3D Interaction Elements 120 .;.~;,v0&;¥£'~'C!';;'i:?' ,uwLD:u.s r-I~n~~ _ •.".". (a) SDOF system (b) Force-!llsp. (b) Fon:a-!lisp. Beam Model Response Ca) Response II Figure 13. Eccentric Axially III Figure 14. Dynamic Collapse using III Figure 15. Hysteretic Degradation Loaded Elements Geometric Nonlinear Beam Reinhom et. aI., 1994, Valles et. aI., is added, the developer has to deal 1996, and Reinhom et. aI., 1998.) only with the data, and routines of The authors undertook an expan the new class itself.Also, changes sion to the three-dimensional sys to one class will not affect the rest tems of the code of IDARC in a of the program because of the en redevelopment effort using an ob- . capsulation. This Simplifies the in ject-oriented approach. The result tegration of new parts. ing software architecture will In the second step, components enable progressive growth by easy are incorporated in the new plat addition of new models from other form IDARC3D Version 2.0: (i) en research tasks and provides the out 'ergy dissipation systems / dampers come to the engineering commu and (li) a computing core capable nity at large. The current focus of of nonlinear analysis.The new pro work is to provide the tools for gram structure provides possibili modeling damping and other ad ties to easily add new elements vanced systems using a unified ap such as base isolators, adapted from proach to nonlinear systems. The the platform 3D-BASIS also devel work done in cooperation with oped by Reinhorn and LARSA, Inc., a software developer, Constantinou in multi-annual enables creation of a user friendly projects. To ensure convenience, and acceptable analysis platform. IDARC3D Version 2.0 operates on a PC and has a graphical user inter Objectives and results face for input and output (I/O).The The redevelopment includes I/O is decoupled from the core of three main steps. The :frrst step is the program so that it can be to create a flexible and extendable changed without interfering with setup for the platform using an the actual program. object-oriented finite element pro The third step is to model and gramming approach. This modular evaluate a structure with protective framework clearly separates the dif systems and to verify the results of ferent elements of the program by the nonlinear analysis with encapsulating data in classes. IDARC3D 2.0 against results from Gasses are black boxes, which pro other standard analysis programs or vide easy to use interfaces through experiments. out the program.Thus if a new class Analysis and Design ofBuildings witb Added Energy DiSSipation Systems 121 The implementation of the new lation examples. The developer's design is realized using FORTRAN manual, which describes the modu 90 to be consistent with the previ lar setup and provides the reader ous and existing IDARC programs. with the necessary information to Much of the existing code is reused change or extend the program, will in the new platform to minimize be published through the MCEER programming new code. Although Networking activities. FORTRAN 90 is not an object-ori A California hospital building is ented programming language, ob being modeled for research by ject-oriented features can be other center investigators, and a simulated with a reasonable effort. benchmark physical model tested Also, source code written in FOR on the shake table at University at TRAN 90 and C++ (as an example Buffalo is being analyzed to provide for an object-oriented program the first example cases. The result ming language) can be used to form of this development will provide one platform together. the engineering community with a The documentation developed three-dimensional nonlinear analy for the new platform includes sis platform that currently does not guidelines to further develop the exist. system, a users manual and instal- Rererences------ ABAQUS, (2000), Theory Manual, Version 6.1, Hibbitt, Karlsson and Sorensen, Pawtucket, RI. Constantinou, M.e., Tsopelas, P., Hammel, W. and Sigaher, A.N., (2001), "Toggle-Brace Damper Seismic Energy Dissipation Systems;' Journal of Structural Engineering, ASCE, Vol. 127, No.2, pp. 105-112. Constantinou, M.C., and Sigaher, A.N., (2000), "Energy Dissipation System Configurations for Improved Performance;' Proceedings, 2000 Structures Congress, ASCE, Philadelphia, PA, May. Constantinou, M.C., Soong, T.T., and Dargush, G.E, (1998), Passive Energy Dissipation Systems for Structural Design and Retrofit, MCEER-98-MNOl, Multidisciplinary Center for Earthquake Engineering Research, University at Buffalo. Federal Emergency Management Agency, (1997), NEHRP Guidelines for the Seismic Rehabilitation of Buildings, FEMA 273, Federal Emergency Management Agency, October. Federal Emergency Management Agency, (1997), NEHRP Commentary on the Guidelines for the Seismic Rehabilitation of Buildings, FEMA 274, Federal Emergency Management Agency, October. Gupta A. K. and Law ]. W., (1986), "Seismic Response of Non-classically Damped Systems," Nuclear Engineering and Design 91, pp. 153-159. Holland, ].H., (1962), "Outline for a Logical Theory of Adaptive Systems; J Assoc. Compo Mach., Vol. 3, pp. 297-314. Holland, ].H., (1992), Adaptation in Natural and ArtifiCial Systems, MIT Press, Cambridge, MA. 122 References, (Cont'd)..... ------ Holland, ].H., (1995), Hidden Ordet·, Addison-Wesley, Reading, MA. Hunta; A., (1995), Sugal Programming Manual, Version 2.1, University of Sunderland, England. Kunnath, S.K., Reinhom, A.M. and Lobo, R.E, (1992), IDARC Version 3.0: Inelastic Damage Analysis of Reinforced Concrete Structures, Technical Report NCEER 92- 0022, Multidisciplinary Center for Earthquake Engineering Research, University at Buffalo. Kunnath, S.K. and Reinhorn, A.M., (1989), Inelastic Three-Dimensional Response Analysis of Reinforced Concrete Building Structures (IDARC-3D): Part 1 - Modeling, Technical Report NCEER-89-QOll, Multidisciplinary Center for Earthquake Engineering Research, University at Buffalo. Lee, G. C, Tong, M .. and Wu, Y.H., (2001), "Some Design Issues for Building Seismic Retrofit Using Energy Dissipation Devices," Inn. Journal of Structural Engineering and Mechanics, (in press). Nagarajaiah, S., ii, C, Reinhom, A.M. and Constantinou, M.C, (1993), 3D-BASIS-TABS: Computer Program for Nonlinear Dynamic Analysis of Three Dimensional Base Isolated Structures, Technical Report NCEER-93-0011, Multidisciplinary Center for Earthquake Engineering Research, University at Buffalo. Nagarajaiah, S., Reinhom,A.M. and Constantinou, M.C, (1991), 3D-BASIS - Nonlinear Dynamic Analysis of Three-Dimensional Base Isolated Structures: Part II, Technical Report NCEER-91-0005, Multidisciplinary Center for Earthquake Engineering Research, University at Buffalo. Nagarajaiah, 5., Reinhom, A.M. and Constantinou, M.C, (1989), Nonlinear Dynamic Analysis of Three-Dimensional Base Isolated Structures (3D-BASIS), Technical Report NCEER-89-0019, Multidisciplinary Center for Earthquake Engineering Research, University at Buffalo. Papageorgiou,A.S., (2000), "Ground Motion Prediction Methodologies for Eastem North America," Research Progress and Accomplishments, 1999-2000, MultidiSciplinary Center for Earthquake Engineering Research, University at Buffalo, pp. 63-67. Park, Y]., Reinhorn, A.M. and Kunnath, S.K., (1987), IDARC- Inelastic Damage Analysis of Reinforced Concrete Frame - Shear-Wall Structures, Technical Report NCEER-87- 0008, Multidisciplinary Center for Earthquake Engineering Research, University at Buffalo. Ramirez, 0., Constantinou, M.C, Kircher; CA., \'Vhittaka;A.S.,]ohnson, M.W. and Gomez, ].D., (2000), Development and Evaluation of Simplified Procedures for Analysis and Design of Buildings with Passive Energy Dissipation Devices, MCEER-OO-0010, Multidisciplinary Center for Earthquake Engineering Research, University at Buffalo. Reinhom, A.M., Simeonov, v., Mylonakis, G. and Reichman, Y., (1998), IDARC Bridge: A Computational Plat/orm for Seismic Damage Assessment of Bridge Structures, Technical Report MCEER-98-0011, Multidisciplinary Center for Earthquake Engineering Research, University at Buffalo. Reinhom, A.M., Nagarajaiah, 5., Constantinou, M.C, Tsopelas, P. and Ii, R., (1994), 3D BASIS-TABS Version 2.0: Computer Program for Nonlinear Dynamic Analysis of Three Dimensional Base Isolated Structures, Technical Report NCEER-94-0018, Multidisciplinary Center for Earthquake Engineering Research, University at Buffalo. Reinhom, A.M., Kunnath, S.K. and Panahshahl, N., (1988), Modeling of R/C Building Structures With Flexible Floor Diaphragms (IDARC2), Technical Report NCEER-8& 0035, Multidisdplinary Center for Earthquake Engineering Research, University at Buffalo. Analysis and Design ofBuildings with Added Energy Dissipation Systems 123 References, (Cont'd) Simeonov, Y.K., Sivaselvan M. and Reinhorn, A.M., (2000), "Nonlinear Analysis of Frame Structures by the State-Space Approach," Computer-Aided Civil and Infrastructure Engineering, (Special Volume in honor of Prof. Gear), Vol. 15, Jan 2000, pp. 76·89. Sivaselvan, M., and Reinhorn, A.M., (2000), "Hysteretic Models for Deteriorating Inelastic Structures," Journal of Engineering Mechanics, Vol. 126, No.6, Jun. 2000, pp.633-&iO. Soong, T.T. and Dargush, G.F., (1997), Passive Energy Dissipation Systems in Structural Engineering,]. Wiley, England. Tsopelas, P., Constantinou, M.C., Kircher, e.A. and Whittaker,A.S., (1997), Evaluation of Simplified Methods of Analyses for Yielding Structures, NCEER-97-0012, Multidisciplinary Center for Earthquake Engineering Research, University at Buffalo. Tsopelas, P.C., Constantinou, M.C., and Reinhorn, A.M., (1994), 3D-BASIS-ME: Computer Program for Nonlinear Dynamic Analysis of Seismically Isolated Single and Multiple Structures and Liquid Storage Tanks, Technical Report NCEER-94-001O, Multidisciplinary Center for Earthquake Engineering Research, University at Buffalo. Tsopelas, P., Nagarajaiah, S., Constantinou, M.e. and Reinhom,A.M., (1991), 3D-BASIS M' Nonlinear Dynamic Analysis of Multiple Building Base Isolated Structures, Multidisciplinary Center for Earthquake Engineering Research, University at Buffalo. Valles, R.E., Reinhorn, A.M., Kunnath, S.K., Ii, C. and Madan, A., (1996), IDARC2D, Version 4.0: A Computer Program for the Inelastic Damage Analysis of Buildings, Technical Report NCEER-96-001O, MultidiSCiplinary Center for Earthquake Engineering Research, University at Buffalo. Winklemann, K., (2001), "IDARC3D VERSION 2.0: Three Dimensional Inelastic Dynamic Analysis of Structures with Protective Systems," MS Thesis, University of Darmstadt (completed while Visiting Research Associate at the University at Buffalo). 124 Using Cost-Benefit Analysis to Evaluate Mitigation for Lifeline Systems by Howard Kun1'euther (Coordinating Author), Chris Cyr, Patricia Grossi and Wendy Tao Research Objectives The purpose of this research is to examine how cost-benefit analysis (CBA) National sCience Foundation, can be utilized to evaluate the attractiveness of mitigation for lifeline systems Earthquake Engineering subject to earthquake ground motion.We propose a framework for the CBA Research Centers Program that can be used in conjtmction with work being completed by other re searchers at MCEER (Shinozuka et al., 2000; Chang et al., 2000). With their development of fragility curves and insight into specific utility lifelines sys tetns, our fran1ework is useful for the next step in the analysis. In this paper; we use an example of a transportation system to show the CBA framework. Howard lom-reutber, Co TIlen, we consider two case studies to show the effects of the disruption of director, Chris Cyr, Undergraduate Student, utility lifeline service on two stakeholders in the analysis. First, the indirect Patricia Grossi, Post economic loss to business owners is studied, and then, the cost to public Doctoral Researcher; and agencies to shelter displaced residents is considered. Wendy Tao, Undergraduate The two lifelines serving as case studies for this work are the electric Student, Wharton Risk power system in Shelby County, Tennessee [building on previous work Management and Decision done at MCEER] and the water distribution system inAlameda and Contra Processes Center, The Costa Counties, California [working with the East Bay Municipal Utility Wharton School, Unh:ersity District (EBMUD)]. Our research provides a fran1ework to link data from a/Pennsylvania the physical and engineering sciences (i.e., seismology of the region and vulnerability of the lifeline) with the social sciences (Le., costs of natural disasters and public policy implications of mitigation). ost-benefit analysis (CBA) is a systematic procedure for evaluating de C cisions that have an impact on society. There are different ways to conduct a valid CBA, depending on the information one has and the na ture of the problem at hand.We chose a simplified five-step procedure to illustrate this approach (Figure 1). Amore comprehensive approach, which incorporates several additional steps, is discussed in Boardman et al. (200 1). The five-step procedure includes: defining the nature of the problem, in cluding the alternative options and interested parties; determining the direct cost of the mitigation alternatives; determining the benefits of miti gation, via the difference between the loss to the system with and with out mitigation; calculating the attractiveness of the mitigation alternatives; and, fInally, chOOSing the best alternative. 125 These steps were chosen keep ing in mind the complex process of estimating losses to lifeline sys Step 1 • David Lee and Tim tems and evaluating the benefits of Specify Nature of Problem Fuette, East Bay Municipal mitigation to the system. Previous - Alternative Options Utility District work performed at the Wharton - Interested Parties School analyzed the cost-effective • Water distribution systems ness of mitigation to residential I in Alameda and Contra structures (Kleindorfer and Costa Counties, California Ktmreuther, 1999). In this analysis, Step 2 (working with East Bay Determine Direct Costs the reduction in damage to the Municipal Utility District) of Mitigation structure was accomplished Alternatives through a shift in the fragility (Le., vulnerability) curve in the analysis and a recalculation of the expected I loss. If the expected reduction in Step 3 loss exceeds the cost of undertak Determine Loss to System ing mitigation (e.g., step 4 in Fig with and without ure 1), then one can justify Mitigation Alternatives investing in it. The analysis of the benefits of I mitigating lifeline systems is a more complicated process than for a resi Step 4 dential structure. Lifeline systems Calculate Attractiveness have unique characteristics, which of Mitigation Alternatives (NPV or BIC ratio) make the calculation of damage to the system difficult (Chung et al., 1995). First, the loss of fimction of I the lifeline is dependent on many Step 5 parts of the system, often buried un Choose Best Alternative derground, spread across a large geographic region rather than at one location (e.g., as in the case of • Figure 1. Simplified Cost-Benefit Analysis a residential structure). For ex for Lifeline Systems ample, in analyzing the functional f~·-C.~7Y;:~,i,'5£j,:H:0~~'.i?~;./ .'...... ". l ~;(~l:;:pl~:~g~ttSiijli!ig~!eis plimadlyundertaken by owners, but [;!.li~~~:==i:::!:!":: l\~~'l;\flJi,R!;~~~M~e;:~~n~~:~:t:;s.,;~!:!::~. t, ... t()~<>rg~~o>i'() fitti(•• tb~~ost of implementing mitiga- ~.. ;~()i}l~~~~t"~"'l.~~1?-~~;~~es can help determine if f .Yi~i~~c~~df~~~()~it11os~exceeds the cost of undertak LitigtWtig~tign~:TJie'thrllsto(iheresearch is to encourage and f~.JlJstifY:! ~plemeriUl.tion of i~b:ofit schemes for a variety of r lif'eliile 'SYs~lDS. .' ~\~ 126 reliability ofan electric power trans considered and the interested par mission network, one needs to con ties in the process. Normally; one sider the loss of connectivity of alternative is the status quo. In the different substations, as well as the case of the current analyses, the sta Program 1: Seismic vUlnerability of the various compo tus quo refers to the current vul Evalllation and Retrofit of nents of each substation, to the sys nerability of the lifeline system Lifeline Networks tem as a whole.This process can be without a mitigation measure in even more complex when one must place. The status quo is likely to be • Task 1.3 Loss Estimation Methodologies consider the collocation ofdifferent the reference point for evaluating lifeline systems. For example, if the how well other alternatives per • Task 1.41lfemphis Lifeline electric power system's cables are form. In general, if there is sufficient Systems Analysis adjacent to the water distribution political dissatisfaction with the system's pipelines underground, proposed mitigation options and! • Task 1.10 Rehabilitation damage to one can compound the or the perceived expected benefits Strategiesfor Lifelines damage to the other. (i.e., reduction in lifeline disrup Second, the damage to the sys tion) are considered to be less than Program 2: Seismic Retrofit of Hospitals tem, measured in outage or service the expected costs to mitigate the disruption, needs to be translated system, then the status quo will be into dollar loss to the interested maintained. parties in the cost-benefit analysis. For the utility lifeline problem we The benefits of mitigation are the are studying in Shelby County,Ten difference between the loss with nessee, the status quo is the vulner out mitigation and the loss with ability of the electric power system mitigation. The costs of mitigating currently in place. An alternative the system must also be estimated. option is to retrofit or replace some While these mitigation costs are or all of the substation equipment generally borne by the owner and components in the electric power operator of the lifeline, everyone system so they are still functional in the region benefits from the un after a severe earthquake. For ex interrupted or faster restoration ample, the transformers in the sub lifeline service after a disaster. I stations can be seismically Therefore, the attractiveness of retrofitted to withstand lateral load mitigation to a lifeline system ing. Alternatively, the high-voltage should be viewed as beneficial to transformer bushings can be re society as a whole, calculated via a placed before an earthquake occurs. societal benefit-cost ratio. The Each of the alternative options thrust of this research is to encour will impact a number of individu age and justify funding for earth als, groups and organizations in our quake hazard mitigation of lifeline society. It is important to indicate systems. who will benefit and who will pay the costs associated with different alternative options when undertak The Five-Step ing a CBA analysis. In the case of a Procedure lifeline system, one needs to con sider a broad set of interested par Step 1: Specify Nature ofthe Problem ties. These include residents and To initiate a CBA, one needs to business owners affected by the specify the options that are being earthquake, public sector agencies that must respond and fund the re- 'llsing Cost-'Benefit _'2lnafgsis to 'Eva[uate 9'vfitigation for LifeEine S!JStems 127 4r;;;;~''f!Z~C''''''CCC=='';;::'';;::'''' covery process, as well as the gen out mitigation. In the other alter eral taxpayer that will bear some natives, benefits will be estimated BttrBctiveness of the repair costs of the damaged from the change in damage to the of mitigBtion to lifeline system(s). system with the status quo and B lifeline damage to the system with mitiga system should Step 2: Determine Direct Costs of tion in place. Once the system reli be viewed BS Mitigation Alternatives ability with each mitigation beneficiBI to For each mitigation alternative alternative has been specified, it (i.e., all alternatives except the sta society BS B should be possible to quantify the whole." UIS quo), one needs to specify the effects of serviceability to the in direct cost to implement the miti terested parties by attaching a dol gation measure. For a lifeline sys lar value to them. The calculation tem, the owner and operator of the from damage to loss is a compli system incurs the costs of mitiga cated process and will differ from tion. In a large majority of the cases one interested party to another in the United States, it is a public (e.g., losses to industry from busi sector agency. Furthermore, the ness interruption differs from resi costs are most likely direct mon dentialloss due to relocation). etary costs to struculrally retrofit or replace some components of the Step 4: Calculate AttraCtiveness of lifeline system. In the case studies Mitigation Alternatives we will present, a government In order to calculate the attrac agency incurs the direct cost of tiveness of mitigation, the nature of mitigation. the benefits to each of the inter ested parties is estimated and com Step 3: Determine the Benefits of pared to the upfront costs of Mitigation Alternatives mitigation.With respect to lifelines, Once the costs are estimated for the alternatives involve a degree of each mitigation alternative, one outage or serviceability over a pre needs to specify the potential ben scribed time horizon (I). One char efits that impact each of the inter acterizes the impact on the key ested parties. In the case of seismic interested parties during the days risk, one considers either a scenario or weeks that the system will not earthquake event or a set of sce be fully functional. One then utilizes nario earthquakes of different mag a societal discount rate to convert nitudes, location, duration, and the benefits and costs of the alter attenuation that can affect the sys native over time into a net present tem. With the specification of the value (NPV). If the NPV is greater vulnerability of the lifeline system, than zero, then the alternative is the damage to the various compo considered attractive .Alternatively, nents of the system is then esti one could calculate the ratio of the mated for each alternative option. discounted benefits to the upfront Then, overall system reliability is costs to determine the attractive estimated. ness of the alternative. Whenever With the status quo, there will be this ratio exceeds 1 the alternative no benefits because no retrofit or is viewed as desirable. replacement scheme is character To illustrate, consider the case of ized. In other words, the status quo damage to a water distribution sys is the damage to the system with- tem from a scenario earthquake 128 event. There could be a period of lation or a ratio of the benefits to time where businesses may not be the costs, one can choose the alter able to operate in a normal man native with the highest NPV or ben ner due to loss of service. If mitiga efit-cost ratio.This criterion is based tion were implemented (e.g., on the principle of allocating re underground pipelines were re sources to its best possible use so placed or retrofitted), the time to that one behaves in an economi restore water to businesses may be cally efficient manner. shortened. Suppose that the miti gation measure reduced the resto ration time by 3 days follOwing a Applymg the Five-Step severe earthquake. The resulting Procedure to a savings in business interruption Transportation System costs during this three-day time in terval following the earthquake are We now illustrate how the above then discounted back to the five-step procedure of eBA can be present. These savings are multi utilized to evaluate mitigation fur a life plied by the annual probability of line system through a simple example. such an earthquake occurring to In the next section, we will indicate compute the expected benefits of the types of data that are required to mitigation. A similar calculation undertake different levels of analyses would be made for any earthquake ofmore realistic problems. that could occur in the area. These savings are then summed up to Step 1: Specify Nature ofthe Problem determine the total expected ben The following question has been efits which are then compared to posed to a public agency that owns the upfront costs of mitigation to and operates a transportation net determine the cost-effectiveness of work in an earthquake prone region the mitigation measure. in the United States: Should thetran& One must be careful, however, portation agency seismically upgrade when considering the time that ser their 2,200 highway bridges so that vice is unavailable to the users of they perfonn adequately in a major certain lifeline systems. Depending earthquake? on the type of system (e.g., trans There are only two alternatives portation, water distribution, elec for this problem: tric power), days without service At. Seismically retrofit the bridges are, on average, less with electric AZ. Do not retrofit the bridges power systems than other types of (i.e., maintain status quo) systems.This is primarily due to the There are a number of interested redundancy in these systems and parties who are affected by this the critical nature ofelectric power problem. These include the resi systems in emergency response dents who use the bridges to com and coordination following an mute to work, the businesses that earthquake (Chung et al., 1995). use the transportation network to move goods, the owner of the trans Step 5: Choose the BestAlternative portation network, police and fire Finally; once the attractiveness of teams that cannot move across the each alternative is calculated bridge, environmental groups con through a net present value calcu- cerned with the impact of a col- Using Cost-13eneftt :wEgsis to 'Ella[uate !kfitigation for LifeB.ne SgsteW.5 129 lapsed bridge on residents of the Step 4: Calculate Attractiveness of 1~;;;:,;;;;';~;?0?~ sea.We will focus on just the owner Mitigation Alternatives meaningful of the network to keep the analy To determine the impact of ret cost-benefit sis simple. rofitting the bridges through a net analysis present value calculation, three ad methodology Step 2: Determine Direct Costs of ditional pieces of data are required: Mitigation Alternatives requires (1) the probability of the scenario The direct cost associated with earthquake, p; (2) the life of the bringing seismically retrofitting the bridges bridge, T, in years; and (3) the an scientists, is the material and labor cost to nual societal discount rate, d. engineers and complete the retrofit scheme. In As each of these parameters is social scientists this case, we assume that, on aver varied, one will obtain a different together to age (i.e., over 2,200 bridges), this relationship between the net analyze a cost is approximately $30 per present value (NPV) of the. ex problem." square foot. pected benefits and costs ofA 1 and A2.To illustrate this point, suppose Step 3: Determine the Benefits of that one utilized the following pa Mitigation Alternatives rameters to determine whether or The benefits of retrofitting the not retrofitting the bridges is ben bridges depend on whether or not eficial to the transportation depart an earthquake occurs and the ment. First, the probability of the length of time (1) that the bridges earthquake occurring,p,is a 1 in a are impaired after the earthquake 1000year event (i.e.,p = 0.01). The with and without the retrofit life of the bridge, T, is 50 years, and scheme in place. For Simplicity, we the discount rate, d, is 4%. consider only one scenario event, Suppose an earthquake occurred with two possible outcomes: (1) and the bridges were retrofitted. the earthquake occurs with prob Then, the reduction in losses from ability,p, or (2) no earthquake oc this retrofit scheme is $140 per curs with probability, I-P. square foot, and the expected ben For this event, the benefits con efits from mitigation would be cal sist of several different impact cat culated (1/100)(140) = 1.4.With d egories. We will consider one = 4%, the expected discounted ben category for this analysis: the dam efits of retrofitting over the 50 year age to the bridges. Damage and loss life of the average bridge would be will be the same or lower if the $30 per square foot. As we vary bridges are retrofitted (AI) than if each of the parameters, the dis they are not (A2). In other words, counted expected benefits from the benefits from mitigation are the mitigation will change. In general, reduction in costs to repair the higher values of d and/or smaller damaged bridges. In this case, we values of p and T will cause ben assume that the cost to repair the efits to decrease. If the cost of ret damaged bridges without a retro rofitting the bridges were set at $ 30 fit scheme (Le.,replace the bridges) per square foot, then the net is $140 per square foot. With the present value would be exactly 1. retrofit scheme, the cost is zero. Any time that this cost was less than $ 30 it would be beneficial to retrofit the bridge. Of course, from a societal point of view, it would 130 be beneficial to retrofit the bridge The reasoning is simple. The ben if the cost per square foot exceeded efit-cost ratio for this problem $30. There would be additional would be (30/15), which is 2. Even benefits to residents who use the if the estimates of the benefits of bridges to commute to work or for mitigation were off by a factor of pleasure, businesses using the 2, one would still want to retrofit bridges to move goods, and avoid the bridges if the cost of this mea ance of business interruption due sure is $15 per square foot. to loss of the transportation net work. These additional benefits of mitigation are due to the reduction Mitigation of Lifelines in time to repair the damaged in Tennessee and bridges. California "Step 5: Choose the Best Alternative We now tum to the two case stud The criterion used by CBA is to ies oflifeline systems to illustrate how maxjmjze net present value (NPV) of one would compute the benefits and societal benefits or minimization of costs ofmitigation: an electric power the total societal costs. In the above system in Shelby County,Tennessee example, with the expected costs and a water distribution system in equal to the benefits of $ 30 per Alameda and Contra Costa counties, square foot, retrofitting the bridges California. The impacts of an earth (A2) would be preferable over main quake on these lifeline systems in taining the status quo (Al). clude a wide range of direct and In addition to the maximization of indirect losses. Our focus is on two social benefits, there may be equity types oflosses to two different stake considerations that play a role in the holders: (1) the impact of interrup evaluation of different alternatives. tion of service on business For example,ifthere were concerns operations and the resulting losses by taxpayers on how much extra in gross regional product (to Shelby they would have to pay for tolls over County) and (2) loss ofservice to resi the bridges to reflect the extra ex dential customers and the need to penditures of retrofitting the bridges, relocate individuals and entire fami then this may impact on the imple lies to temporary shelters (in mentation of A2 even if it was AIamedaand Contra Costa counties). deemed cost-effective using maximi In these analyses, we consider zation of NPV as a criterion. only these two limited impacts of In essence, the choice of an opti earthquake loss. Other impacts, mal alternative is based on a set of such as the costs associated with assumptions that need to be care fire following an earthquake, are fully examined. In particular, one not considered here. For the total will want to undertake a set of sen loss to the water distribution sys sitivity analyses to determine how tem in Alameda and Contra Costa robust the proposed solution is. For counties, we refer the reader to the example, if the expected cost of study completed for the East Bay retrofitting the bridges were only Municipal Utility District (Homer $15 per square foot, then there and Goettel, 1994). would be little doubt that mitiga tion would be a cost effective one. 'llsing Cost-'Beneftt .'Jl.nafysis to 'Eva[uate Mitigation for Life{{ne S!fstems 131 Mitigation of an Electric Power tucky to southern Illinois, and System in Shelby County, whose center is located to the Tennessee northwest of Memphis. The NMSZ has a history of earthquake activ As a continuation of an analysis ity, with the largest earthquakes done by researchers at MCEER recorded in 1811-1812, and thus, it (Shinozuka et aI., 1998), we devel is an interesting case study for po oped a framework to look at the tential earthquake damage. cost-effectiveness of mitigation to For this analysis, we used a M = the electric power system of the 7.5 earthquake with an epicenter Memphis light, Gas, and Water Di in Marked Tree, Arkansas, situated vision (MLGW) of the Memphis, fifty-five kilometers northwest of Tennessee area. Loss to the gross downtown Memphis. We looked regional product (GRP) of the dif specifically at mitigation that could ferent business sectors in the re be undertaken on electric power gion due to loss of power is systems, basing this study on a net considered. work of electric power serviced by Figure 2 depicts a map of the Memphis Light, Gas and Water Memphis/Shelby County area with (MLGW). We examine the loss of indicated Modified Mercalli Inten serviceability of the electric power sity (MMl) impacts by census tract system from the scenario M = 7.5 for a M = 7.5 earthquake. The earthquake, based on data devel Shelby County area is at risk from oped by Chang (See Shinozuka et. seismic activity in the New Madrid aI, 1998).We should note that dam Seismic Zone (NMSZ) of the Cen age to the overall system reliability tral United States.This seismic zone was estimated using Monte Carlo lies within the central Mississippi simulation techniques, and with Valley, extending from northeastAr this damage, loss of service to dif kansas, through southeast Missouri, ferent service zones was estimated. western Tennessee, western Ken- The focus of this research is how to use this output from the simula tion analysis to look at the cost-ef fectiveness of mitigation. MMl by Census Tract 1'1 VIII Electric power is absolutely cru • Vlll-ll2 cial in maintaining the social sys tems of a city, directly affecting businesses that cannot operate without it. The study of mitigation for the gas, power and water distri bution systems is not studied here. For an update on the damage and loss to the water distribution sys tem, see Chang et al. (2000). Mitigation in this case is a seis mic retrofit of a component of the electric power substation. Specifi • Figure 2. Shelby County, Tennessee (NCEER Bulletin 1996)2 cally, the transformers, used within the substations of the network to convert power, are retrofitted to 132 withstand lateral loading. Tying In this Simplified analysis, if the down the wheels on the track can transformers are seismically retro significantly improve the sliding fitted, then we assume the loss of movernent and reduce the chances serviceability will be reduced, since that these large, critical structures some or all of the transformers will overturn.3 would be functional after the earth The estimation ofloss to the busi quake. After knowing the differ nesses in the area, due to service ence in the number of functional interruption, is calculated using the transformers before seismic retro direct economic loss methodology fit and after the retrofit scheme, one from ATC-25 and modified by can then compute the savings to Chang to accoDlfllodate the impor the business sector. In other words, tance factors specific to the Mern the reduction in electric power out phis area. The initial loss, I., to J age ch~cterizes the benefits of business sector j in the region di- mitigation should the scenario rectly following earthquake is esti earthquake occur in the service mated by equation (1). area.The expected benefits of miti Lj=(l-a)·gj·Pj (1) gation are calculated by multiply ing these benefits by the In the above equation, a is the probability of the earthquake (p) initial availability to the area. In and discounting over the relevant other words, it is the percentage of time horizon (1) that the trans customers receiving service ifllme former is expected to be in use (i.e., diately after the earthquake. gj is step 4 in the cost-benefit analysis). the gross regional product (GRP) To illustrate the type of CBA that in the region for the business sec could be undertaken, consider the tors, j. In our case, j includes ten following illustrative example. As separate sectors. Examples include sume that the cost of retrofitting agriculture, construction, manufac each transformer is $10,000. If all turing, and services. The GRP was 60 transformers in the electric developed using data from the u.s. power system are retrofitted, there Census Bureau. Finally, Fj is the im is a one-time total cost of $600,000. portance factor of the business sec Therefore, there are two alterna tor, described as the percentage of tives: the status quo and retrofitting production in sector j that would all the transformers for an upfront be lost if electric power service cost of $600,000.The stakeholders were completely disrupted (ATC- in the analysis are the businesses 25,1991). in the area needing electric power For the scenario earthquake for production output.The benefits event, suppose restoration time is of mitigation are the reduction in T days. So, the loss of production restoration time due to more trans over T days, IT, is addressed by formers functional following the equation (2). earthquake event. Additionally, we assume that the electricity system will last for 50 4=f L/(T- t) (2) t=O T years and an 8% annual discount rate is used to compute the ex pected benefits over the 50-year horizon from retrofitting the trans- '[Ising Cost-'Beneftt ,0,n.afgsis to 'Eva[uate Mitigation for .£ijefine S!JStems 133 formers. We consider an earth percent of the users within one day. quake of M = 7.5 would occur in But, in the Chi-Chi, Taiwan earth the region once every 500 years, es quake in 1999, it took longer to re timated from a study done by the store power, with rolling blackouts USGS (Atkinson et aI., 2000). to the northern part of the island Figure 3 depicts the sensitivity of for a week or more. the CBA to restoration times of the This analysis should be viewed as electric power system would be illustrative rather than definitive, reduced by either 1, 3 or 5 days. since it is highly dependent on the The figure is designed to show how assumptions made regarding the CBA can be utilized to evaluate costs of retrofitting and the busi whether or not mitigation of cer ness interruption savings as well as tain parts of the system is cost-ef the discount rate, d. For example, if fective to certain stakeholders in d were four percent, then retrofit the analysis. In this example, retro ting the transformers would be fitting transformers can be seen to even more attractive than under be beneficial (i.e., the benefit-cost the current analysis. ratio is greater than 1) for a time horizon T greater than 13 years Mitigation of a Water when restoration of the power sys tem is reduced by five days and for Distribution System in Alameda T greater than 23 years when res and Contra Costa Counties, toration is reduced by three days. California Retrofitting the transformer is not The East Bay Municipal Utility cost-effective for any T when res District (EBMUD) is the public util toration of the power system is ity that provides potable water ser only one day. This is important to vice to Alameda and Contra Costa note because after the Northridge Counties in northern California.We earthquake in 1994,electric power utilized information from the (LAPWD) was restored to ninety- EBMUD Seismic Evaluation Pro gram Final Report (Homer and Goettel, 1994), an intensive study 1.8 of the East Bay water system that 1.6 developed mitigation initiatives --' .' 1.4 that will be completed during their 0 Seismic Improvement Program. :; 12 ----...... ~ ...... ~ 1~..,J' __ ~o->--"""'" cr -1 day Our analysis focuses on the mitiga u; 1.0 0 tion of 172 water storage tanks in U ? 7' ~-3day a;"" 0.8 /' 120 of the 122 EBMUD pressure .",,J'" 5 day ~ 0.6 .-' zones . !O Basel;" 0.4 To conduct this analysis, four sce nario earthquake events were cho 0.2 I I sen for analysis, based on the 0.0 I~ completed EBMUD report (Table 0 5 10 15 20 25 30 35 40 45 50 1).These events represent the maxi Tirre period (}ears) mum plausible earthquake event that could occur on any given fault, • Figure 3. illustrative Example of Effects of Electricity Lifeline Mitigation with the exception of the Hayward 134 M = 6.0. For the Hayward fault, II Table 1. Scenario Events an earthquake of magnitude M ------. , .... ',; ;'::',,: ... \': . 6.0 is the most probable event ;Fau!t Name , .• M I .....: iJeScriptInll: ..' '.' = Hayward 7.0 event ruptures 50 to 60 km long segment of fault and M =7.0 is the maximum plau Hayward 6.0 event ruptures an 8 to 13 km long segment of the fault sible event. Calaveras 6.75 event ruptures the northernmost part of the known active fault In this analysis, there are two Concord 6.5 event occurs along this fault east of the EBMUD service area alternatives: the status quo and the mitigation of the water stor age tanks.Those affected by the dis vice is disconnected from a place ruption of the water supply system of residence, the individuals who are the residents in the service area live in these buildings will be that may be forced to relocate to forced to find alternative shelter temporary shelters for a period of until water service is restored. In days or even weeks.The direct cost the event of an earthquake, the of retrofitting the water tanks is potable water system can be dam based on data supplied by the aged to an extent that water ser EBMUD and given in Table 2. vice may not be restored to an area until many weeks following the Loss of Serviceability and Effects event. Figure 4 details our method ology for estimating the costs that on Residential Displacement are associated with these displaced The benefits of mitigation are the individuals. losses resulting from loss of water First, we determined the likeli service to residential customers hood of tank failure in each pres with and without the seismic up sure zone. EBMUD performed a grade ofthe tanks. When water ser- detailed analysis of their system, us- II Table 2. Direct Cost to Retrofit Water Tanks ;:;.J< ::.'''''';.-'''= ·:::f .:., ISe~: :1;>; !~~~~s:.C~t,:: ". ip~~ rr.~;:~~;i5:t mat.SlEe!:, l;.~r_~£:~~)c< ~ '.' :7'; ~~.1~'~~.< '''o~k 1~1Ic':' ,,:::;, ~~~ifj. ::.,';::'." iF:::} ...: ..',.' '. i~ ..... ! :: ...... ri Irtl~2.:: ':".? 1'7?;r':~~}_ .• :': ':;',;f:.. .:: iCC;',:.. '. .S,.. ,. Concrete I 0.5 $ 500.000 $ 150.000 6 weeks 110 weeks 1$ 1.000.000 11 vear 1 $ 600 000 $ 150 000 7 weeks 10 weeks $ 1.500 000 11 vear 1.5 $ 750.000 $ 150 000 7 weeks 10 weeks $ 1 750.000 11 vear 21$ 900.000 $ 200 000 8 weeks 10 weeks $ 2.000.000 11 vear 31$ 1.200 000 $ 200.000 8 weeks 10 weeks $ 2.500.000 1 vear 51$ 1500.000 I $ 200,0001 9 weeks 10 weeks $ 3.000.000 11 vear I I 1 I Steel I I I 0.251 $ 100,000.00 $ 50,000 1 week 10 weeks 1$ 650,000 11 vear 0.51 $ 200,000.00 $ 50,000 1 week 10 weeks Is 1,000,000 11 year 11$ 250.000.00 $ 50.000 1 week 10 weeks 1$ 1.500,000 11 vear 31$ 300.000.00 I $ 70,000 2 weeks 10 weeks 1$ 2,500.000 11 vear 51$ 400.000.00 $ 70.000 2 weeks 10 weeks j$ 3.000.000 11 vear Footnotes: 1 Includes non-seismic upgrades, which are approximately 15% of seismic costs. 2 Partial loss for steel tank due to damaged valve pit piping. Time due to valve replacement. The tank structure assumes to either fully survive or fully fail. 3 The existing tanks may be out of service for this period of time; however, if there are multiple reservoirs in a pressure zone, the business interruption will likely be a much lesser time than indicated. Assume temporary replacement with either a .28 MG or 0.4 MG tank. 4 Cost includes: demolish existing tank, valve pit modification, metal appurtenances, miscellaneous site work, water quality piping, 15% construction contingencies. 'U5ing Cost-'Benefit .'7mafysis to 'Evafuate Mitigation fOT Lijeflne Sgstell1.5 135 overall loss of service in a pressure zone is a combination of the effects of individual tanks in the pressure zone failing. The loss to pressure zone one, LpZI' therefore, can by defined by equation (3). In the equation'~1 is the probability of Total Service Loss full loss for a tank in pressure zone within Pressure one and Nl is the number of tanks Zone in zone one. L --N ANI PZI - I Nl (3) t=O Although a pressure zone may experience loss of service due to failure of its internal tanks, it may also experience service interrup tion if pressure zones that pro vide water to its tanks experience failure.To fully tmderstand the to tal loss within a zone, we must • Figure 4. Cost to Shelter Displaced Residents from one Pressure Zone take this into consideration as in dicated by equation (4): ing their own system risk analysis P 1 =(X I ·X 2 ....X 11 )+ software, SERA. Ten simulations of each of the four above earthquakes (i.e., forty simulations in all) were In equation (4),P] is the percent run to determine the extent of age of customers in pressure zone damage that their system is ex one that will be without service pected to face following an earth following an earthquake, Xl quake. For each simulation, tanks through XNrepresent the probabil were said to fully fail, partially fail, ity of failure in pressure zones that or not fail. Our analysis of expected feed into pressure zone one, and loss consists only of two of these L PZI is the loss due to tank failure three states - full failure and no fail in pressure zone one. Ifevery pres- ure - since we have no means by sure zone that provides water to which to accurately analyze the the tanks in pressure zone one fails, complex network effects of partial then no customers in pressure loss. Therefore, if a tank was said zone one will have service. The to fully fail in 6 of 10 Hayward M probability that this occurs is rep 7.0 scenarios, then we say there is resented by the product of the a 60% chance of full tank failure probabilities that each individual following a Hayward M 7.0. precedent pressure zone will fail. Next,determination oflikelywa Ifany or all of these precedent pres ter service within individual pres sure zones has service,however, we sure zones was calculated. Some will assume that the district will be pressure zones are served by more able to provide enough water to its than one tank, and therefore, the 136 tanks in pressure zone one so that To determine the total cost of customers have their minimum sheltering residents displaced by water demands met. Then, the only loss of potable water service to residents without water service their home following the earth ' will be those represented by LpZ1 quake, we multiply Rs by $11.50. or those who do not have service This value is based on an American due to tank failure within pressure Red Cross estimate of between $10 zone one. and $13 to shelter a person for a Once the overall service to each single day (Red Cross, 2000). pressure zone is calculated, the Finally, to determine the total ag number of people displaced and gregate displacement costs follow seeking temporary shelter can be ing an earthquake, we must :f1rst detennined. Since we are analyzing consider whether the per-day costs 120 pressure zones in the system, will remain constant over time.We the number ofpeople seeking tem assume that if a tank is fully dam porary shelter following an earth aged it will not be able to supply quake is as follows: water to any of its customers until 120 it is fully repaired, or until a replace ment tank can be acquired. There Rs = L.,uz • Pz ' nhz ,sz (5) . z=l fore, there will be no incremental In equation (5), Rs represents the increase in service over time and number of residents seeking tem the per-day total costs of shelter porary public shelter following a ing residents will remain constant over that period (since people are disaster and u z is the number of occupied housing units in pressure not able to return home). Based on zone z. This is found by overlaying this assumption, the total cost to U.S. Census data on the service dis shelter displaced residents follow trict.Pz is the percent of service loss ing an earthquake is easily calcu in pressure zone z, or the percent lated by multiplying the total cost age of housing units without wa per day by the total time until res ter service, and ~ is the number toration of service. In this case, we of people per housing unit in zone assume ten weeks or seventy days, z. This is calculated by dividing the based on information provided by total population in zone z by the EBMUD. total number of occupied housing Note that the above analysis char units.Finally,szis percentage of dis acterizes the impact on shelter placed residents in zone zwho well costs under the assumption that seek shelter in a public facility. This the status quo was being main is determined by inputting U.S. Cen tained. The alternative option is to sus tract demographic information retrofit the water tanks in some of into a methodology developed for the pressure zones inAlameda and HAWS (NIBS, 1997)4. This analysis Contra Costa counties. By undertak is run for each of the districts 120 ing this protective measure, one pressure zones considered, then reduces the chances that one or aggregated to detennine the total more of the tanks will fail and number of residents across the dis hence, disrupt water service for trict who will seek shelter follow some period of time. The benefits ing a disaster. of this mitigation measure will be determined by the reduction in res- Vsing Cost-'Benefit Jillafgsis to 'EvaCuate :Jvfitigation. for Life8ne Sgstems 137 toration time for water to resi able for the analysis and keep the dences in the area affected by the interested parties' priorities in earthquake. mind. For this reason, we have been One also needs to take into ac working very closely with person count whether individuals were nel at EBMUD to ensure that our forced to evacuate their homes be analysis of their water supply is a cause there was severe structural meaningful one. damage due to the earthquake.Sup Our intention is to undertake a pose one is evaluating the benefits sufficiently rich analysis of how of retrofitting water tanks to resi mitigation can be utilized for par dents in the area. Then one should ticular lifeline systems, such as the eliminate any homes where evacu EBMUD water distribution system. ation would be required even if the Unless key decision makers can was no disruption of water.. Other appreciate the role CBA can play wise, one would be overstating the in determinillg whether to imple benefits from retrofitting the water ment specific mitigation measures, tanks in the East Bay area. this methodology may have some This analysis of the residential theoretical interest but no practi sector in the East Bay area illus cal importance. trates the type of calculations one The task of making CBA a useful would make to evaluate the ex methodology is a challenging one. pected benefits of mitigation.1\lrn It requires bringing together scien ing to business interruption losses tists and engineers with social sci from an earthquake,a similar analy entists to analyze a problem. It sis to the one in Shelby County is requires one to articulate the na being undertaken with EBMUD in ture of the uncertainties associated northern California to evaluate the with the recurrence interval of impact of mitigation on this sector. earthquakes of different magni Combining both the residential dis tudes, as well as the confidence in placement costs and business inter tervals surrounding the expected ruption, one could undertake a benefits and costs of different al more comprehensive CBA, deter ternative strategies. In a nutshell, it mining under what circumstances requires the integration of science the proposed mitigation will be with policy. most cost-effective. The data and techniques are now available to undertake this type of integration with respect to earth Making CBA Useful for quake mitigation. The challenge is Policy Analysis to present the analyses to decision makers so they are willing to de The above examples are only il fend the proposed recommenda lustrative as to how CBA can be tion because it makes economic used, rather than how it is actually sense to them while satisfying their applied to either Shelby County, political concerns. Cost-benefit Tennessee or Alameda and Contra analysis provides a framework for Costa counties, California. For CBA accomplishing this important task. to be a useful tool for policy analy sis in a specific region, one must have the most accurate data avail- 138 Endnotes ------ 1 Of course, the public utility or private sector organization operating the lifeline facility may raise its rates to reflect the additional cost of the mitigation measure. In this sense, all the users of the facility bear the costs of loss prevention. 2 http://mceer.buffalo .edu/publications/bul1etin/%/02/apr96nb.html. 3 This type of mitigation was chosen, based on discussions with Masanobu Shinozuka and insight from the annual MCEER conference in November 1999. 4 This methodology does consider that a proportion of displaced residents will stay with friends or family rather than seek publicly-funded shelters. References------ATC-25, 1991, Seismic Vulnerability and Impact of Disruption of Lifelines in the Conterminous United States, Applied Technology Council, Redwood City, California. Atkinson, G. et aI., 2000, ~Reassessing New Madrid." 2000 New Madrid Source Workshop, http:Uwww.ceri.memphis.eduJusl!s/oubs.shtml. Boardman, A., Greenberg, D., Vining, A., and Weimer, D., 2001. Cost-Benefit Analysis: Concepts and Practice (2"d Edition), Upper Saddle River; New Jersey: Prentice Hall. Chang, S., Rose, A., Shinozuka, M., Svekla, WD., and Tierney, K., 2000. Modeling Earthquake Impact on Urban Lifelines Systems: Advances and Integration. Research Report, Multidisciplinary Center for Earthquake Engineering Research, Buffulo, New York. Chung, R.M., Jason, N.H., Mohraz, B., Mowrer, Ew, and Walton, W.D., (eds.), 1995. Post Earthquake Fire and Lifelines Workshop: Long Beach California, January 30-31, 1995 Proceedings, NlST Special Publication 889, U.S. Department of Commerce. Homer, G. and Goettel, K., 1994. Economic Impacts of Scenario Earthquakes, Final Report for East Bay Municipal Utility District. Kleindorler, P. and Kunreuther, H., 1999, ~e Complementary Roles of Mitigation and Insurance in Managing Catastrophic Risks." Risk Analysis, 19:727-38. NrnS, 1997. HAZUS: Hazards US.: Earthquake Loss EstimationMethodo[ogy, National Institute of Building Sciences, NIBS Document Number 5200. Red Cross, 2000. Phone conversation with Garrett Nanninga, American Red Cross National Headquarters, 28 July 2000. Shinozuka, M., Rose, A. and Eguchi, R. (eds.), 1998, Engineering and Socioeconomic Impacts of Earthquakes: An Analysis of Electricity Lifeline Disruption in the New Madrid Area, Monograph No.2, Multidisciplinary Center for Earthquake Engineering Research, Buffalo, New York. Shinozuka, M., Cheng, T.-C., Feng, M., and Mau, S.-T., 2000. Seismic Peifbrmance Analysis of Electric Power Systems, Research Report, Multidisciplinary Center for Earthquake Engineering Research, Buffulo, New York. Shinozuka, M., Grigoriu, M., Ingraffea, A.R., Billington, S.L., Feenstra, P., Soong, T., Reinhom, A., and Maragakis, E., 2000. Development of Fragility Information for Structures and Nonstructural Components, Research Report, Multidisciplinary Center for Earthquake Engineering Research, Buffulo, New York. 'Using Cost-r.Benefit .9l.nafgsis to 'Eva{uate 9'vfitigationJor Liftan/! Sgstems 139 Page Intentionally Left Blank Retrofit Strategies For Hospitals in the Eastern United States by George C. Lee, Mai Tong and Yasuhide Okllyama Research Objectives National Science Foundation, This paper describes an approach used to develop retrofit strategies Earthquake Engineering for hospitals and other critical facilities in low to moderate seismic Reseach Centers Program hazard zones, where strong earthquakes are infrequent, but if they should occur, the consequences would be high. Hospitals in New York State and other urban centers in the eastern U.S. fall into tIlis category, George C. Lee, Director, where seismic retrofit requires information on the impact of losing Multidisciplinary Center medical services after a destructive earthquake. A team of MCEER re Jor Earthquake searchers is currently developing an approach to address this task. It is Engineering Research, and a truly multidisciplinary effort, with team members from a variety of Samuel P. Capen Professor disciplines including engineering, seismology, structural dynamics, risk ofEngineering, ApostoI4JS and reliability analysis, manufacturing process engineering, computer Papageorgiou, Professor, simulation, urban and regional planning, and economics. When this and.ilfai Tong, Senior Research Scientist, research task is completed, it will be united with MCEER's general Department ofChi I, hospital project to develop seismic retrofit strategies. Structural and Environmental Engineering, Unitersity at major MCEER research thrust is the development of retrofit strat Buffalo A egies for critical facilities. By fostering team efforts, the research is Li Lin, Associate Professor, focused on comprehensive protection of emergency medical service func Department ofIndustria! Engineering, University at tion of hospitals in the event of a destructive earthquake. This research BUffalo requires system integrated studies involving earthquake hazards, fragili Yasubide OlmyamtJ, ties of all structural and nonstructuraI components and systems, as well as Assistant Professor and human services provided by medical and support staff, and impact and Ernest Stemberg, benefit-cost analyses. Experiences and approaches developed from the Associate Professor, hospital projects will then be extended to seismic retrofit of other critical Department ofArchitecture facilities such as communication centers or manufacturing complexes. and Planning, University at Ruffolo MCEER's hospital retrofit research program addresses two specific types MasanDbu ShinDzuka, Fred of seismic hazard locations. The first is fur hospitals located in regions Champion Chair in Civil with frequent and/or high seismic hazard levels such as many communi Engineering, Department ties in California where retrofit is required by law. The second type of ofCivil Engineering, hospitals are those located in regions with low seismic hazard levels where University o/Southern earthquakes have a very long return period but the structures and con California tents are likely to be damaged when earthquakes do occur. 141 For the first type of hospitals (e.g., in the stage of writing the imple those in California) a considerable mentation procedures required by amount of engineering and social SB1953 for the nonstructural pro • State ofNew York science studies are being carried visions. Under such legal require Department ofHealtb, out by MCEER and other research ments, the challenges for California Bureau ofArcbitectural ers (for example,seeJohnson et aI., are largely focused on engineering and Engineering Facility 1999). Relatively little information tasks. Planning, Troy, New York is available on how to approach In areas of the eastern U.S. such • Columbia Presbyterian hospitals located in low seismic as New York City, hospital retrofit Medical Center, New York, hazard regions but having high risks decisions are made based on differ New York (e.g., those located in eastern U.S. ent considerations. In particular, urban centers). This article briefly given that only limited financial re • Erie County Medical summarizes MCEER's approach in sources are available for protection Center, BtifJalo, New York developing retrofit strategies for the from various natural hazards of ap latter, with an emphasis on estab proximately the same level of prob lishing an evaluation system for ret ability of occurrence, retrofit rofit strategies. decisions become an optimal risk management issue. For these two different conditions, we may thus Different Questions begin by asking two different ques Asked for California tions for the MCEER hospital and New York project. For California hospitals: Hospitals • How can the requirements to Because of California Law SB1953 retrofit be met cost-effectively? (Alquist Act), California hospital For New York hospitals: administrators and code writing • Should resources be allocated for authorities are required to consider the seismic retrofit of hospitals? the nature of functional design for MCEER's hospital project is di critical care facilities. OSHPD (Cali vided into two separate aspects in fornia Office of Statewide Health their initial phase. For the more Planning and Development), which general situation (represented by is directed by SB1953 to address California hospitals), major efforts and implement the legal require are devoted to engineering activi ments, now requires that by Janu ties to establish fragility information ary 1, 2030, all hospital buildings for the physical components and will meet the seismic standards of systems, and identify critical prob the Hospital Act. Also, OSHPD is lem areas in structures, nonstructural Hospitalarl:mfu.istrators, building owners and other stakeholder~i1:lregtons. ofminor to mode.-ate seismicity. can. usetb.~ ~~uation. system for retrofit strategies to lllakeoptim~lri~k management decisions. Resources for hazardmi~isati0tl0f all types are limited, and a decision making methOd based on solid cost-bend'it principles will be a valtlabletooL 142 components, equipment, etc. that For a free-will decision, the third require seismic retrofit. For the step is crucial because the retrofit t"o~ second situation (represented by benefit has to be evaluated in com Resetri'dJ hospitals in the eastern U.S.), we parison to those of other compet Program 2: Seismic Retrofit of concentrate on establishing a deci ing projects for limited resources. Hospitals sion-making method which can Being aware of this important provide information on the impact link,MCEER is concentrating on the to the community if medical ser third step by working with several vice function is lost after an earth hospitals in New York State. These quake due to different levels of hospitals are located in Seismic damage scenarios to the various Zone C (Z = 0.15), where earth required service functions of the quakes with magnitude ~45 or in hospital. Once a decision is made tensity ~ VI have been experienced to perform seismic retrofit, the pro historically. cess will be merged with that de For the purpose of evaluating veloped for the California hospitals. various natural hazard reduction At that point, we consider impact schemes, we start to develop a hos to the community when there are pital operation model based on pa multiple hospitals,followed by ben tient flow as shown in Figure 1. efit-cost analyses for different pos Ifhospital services are considered sible retrofit options. as a process, the key element in this patient-flow model is the center block that describes the process of System Evaluation for how patients receive their medical Hospitals in New York services.The services are supported by two types of resources: human We envision a five-step decision making process for the seismic ret rofit of hospitals. These steps are: 1. Establish earthquake hazard 2 Develop fragility informa- tion and identify critical problem areas in the physi SupportStaff I cal system 3. Establish an analysis tool (hospital operation model) q Patients Out to carry out evaluation of se lected seismic scenarios to determine their impacts to medical services of a given hospital 4. Carry out community im pact analyses (multiple hos pitals/health care facilities) 5. Perform benefit-cost analysis and determine retrofit op tions. III Figure 1. The Patient Flow Model of a Hospital 1?g.trofit Strategies !for %ospitafs in the 'E.astem 'Us. 143 and material. Since our target is to distribution, a critical disaster event evaluate the benefits of structural! may have a different impact to nonstructural retrofit, the emphasis these relationships. Also, to evalu is given to material resources, ate different retrofit schemes, the which typically include power sys level of detail ofthe model may vary. tems, water systems, information In Figure 2, only some typical ser systems, medical systems, transpor vice units are indicated for the pur tation systems, HVAC and others. pose of illustration.The arrows only Depending on the designated func provide examples of possible one tion of hospitals (trauma center, direction patient flow. The total sys general hospitals, special medical tem would be too complicated for care, etc.), the center block will in illustration of the concept. In gen volve different service units. Our eral, an all-purpose comprehensive current emphasis is on modeling model may not be a good approach, the emergency medical service, since too many factors introduce which is illustrated in Figure 2. too many uncertainties, which Each of the service units has in would eventually lead to an unreli ternal structures and are intercon able model. nected.Therefore, modeling of the Similar to modeling the necessary hospital operation has to consider components of providing emer two layers of relationships. gency medicine, the utility system With the emergency medical ser such as water supply, as expressed vice unit of a hospital configuration in Figure 3, can be modeled so that described in Figure 2, a key step of the relationship among the various modeling is the internal relation units can be examined (e.g., effect ship between various service units of damaged water pumps on the (departments) and the delivery of water supply). This utility model emergency medical services. In par can be linked to the emergency ticular, some factors such as season medical service model, where con ality, abnormality, and patient sumption ofwater is required. Here, rPrimary Care I ISpeCialty Carel • Figure 2. Hm;pital Medical Service Units 144 it is important to understand that the same utility system may be modeled differently for applica tion to various physical problems and retrofit treatments. .. Water Pumps Forrester Network Model As mentioned earlier, the pur pose of modeling hospital opera tion is to evaluate the benefit of II Figure 3. Water System Units seismic retrofit. For this reason, we need a quantitative modeL functions; or they could be repre This requirement can be satisfied sented by an empirical function or by a Forrester type of network some logical relations. model. The essential steps of a Figure 4 shows an example of Forrester model are to break down Forrester-type systems model for the physical units and their rela the emergency room in a large hos tions into standard input/output pital. Incoming patients are classi units and networked relations be fied into three categories: minor, tween these basic units. Then the moderate and major injuries/illness. relations are modeled by difference Patients classified as minor and functions including differentiation, moderate are attended to first, and integration, and other elementary complete an information sheet in II Figure 4. Forrester JYpe Systems Model for Emergency Room Operation j(g.trofit Strategies :For J{ospitafs in the 'Eastern ·U.s. 145 the reception area; they then pro time variant function (either con ceed to the triage rooms for treat tinuous or discrete). In general, life ment.The time spent in this process line facilities, medical gas, and (reception to triage) depends on medical supplies are modeled such the degree of injuries/illness (pa that under normal circumstances, tients classified as "moderate" spend these resources are consumed on more time than those classified as a per patient basis, and the inven "minor") and on the capacity of the tory is filled either when consumed triage rooms. Then, after the triage (water, electricity, medical gas) or room, the patients proceed to the when additional supplies are pur waiting room before receiving treat chased (medical supplies). Once an ment. The time spent in this pro internal disaster occurs, one or cess (waiting room to treatment more of these resources sustain room) also depends on the degree damage. For water, electricity, and of injuries/illness and the capacity medical gas, the piping may be dam of treatment rooms for minor and aged and they cannot be supplied moderately injured patients. On the as usual (reduced amount or total other hand, a severely injured pa cut off). Then, emergency transfer tient skips the triage process, due from the reserve tanks (for water to the urgency of their injuries/ill and medical gas) or from the hos ness, and the information on their pital power generators (for electric injuries is provided by paramedic ity) compensates for the reduced staff.The time spent from reception supply. The capacity of these re to treatment for a patient classified serve emergency measures can be as "major" depends on the capacity specified by resource. For medical of the treatment rooms. Because supplies, when an internal disaster this model represents only the op limits their capacity in the emer eration of an emergency room, the gency room or damages them di patients are moved out of the rectly, reserve supplies can be model after the treatment rooms. transferred from the hospital's in Once an internal disaster occurs ventory. In conjunction with dam (such as fire, water pipe broken, age to lifelines and supplies,related power outage, and so forth), the laboratories (such as X-ray, blood process times of the triage room test, and so on) and facilities (oper and treatment room are affected ating rooms, intensive care units, and become longer depending on and so forth) in the hospital are also the damage. In order to simulate an restricted by the internal disaster. internal disaster, the material re The impacts on these material sources of the hospital are modeled resources are then fed back to the as shown in Figure 5. Internal di operation model discussed above. saster in this model can be damage Due to the internal disaster and re to either water pipes, electric lines, lated damage, the process times of medical gas pipes,emergency room both triage and treatment rooms structure, structure in the other ar are affected, and become longer. eas, medical supplies in the emer Consequently, the number of pa gency room, the inventory of tients treated in this emergency medical supplies in the hospital, or room may be reduced in a complex any combination of these situations. manner. The degree of impact to The damage can be defllled as a each patient (classified as minor, 146 II Figure 5. Sub-Model of Material Resources for Internal Disaster moderate or major) may vary due Transfer Function and State to the different requirements for Space Models treatment (for example, a patient with major injuries requir~s more A Forrester type of network electricity; medical gas, and medi model is convenient for modelers cal supplies than less seriously clas but it is not easy to carry out sys~ sified patients). tern analysis and evaluation. Fortu This model can be tailored for a nately, we have several other specific hospital with information analytical models which are equiva regarding the capacity of the emer lent to the time domain network gency room and hospital, and the model. Two of the most useful size of its emergency reserve of re equivalent models are the transfer sources. The model can simulate function model and state-space patient flows through the emer model. For these models, many gency room under any scenario of available control theoretical analy internal disaster, external disaster:, ses can be applied. For instance, the or a combination of both. The zero-pole analysis for transfer func model can also provide statistics, tion analysis may help us to under such as average treatment time, av stand the frequency domain erage time spent in the waiting characteristics of the modeled sys room, and so forth. tem. Several impact response pa- 2?g.trofit Strategies :For J[ospitafs in the 'Eastern 'U.s. 147 rameters such as arising time, ad to determine how much service justing time, peak response time loss will be from the event. Conse and PO% (percentage overshoot) quently, the value of a retrofit will may provide a quantitative measure be evaluated against the chance of of the hospital performance lmder the risk and the associated poten a sudden hazard event. tialloss. For instance, consider an earth quake event which results in a sud den increase of in-flow patient rate Conclusion and simultaneous damage to some Based on information supplied to utility systems.With the shortage of us from the hospitals in New York, medical staff, water and power sup we are in the process of establish ply, the medical services of the hos ing these models. It is our inten pital will be reduced while patients tion to examine the dynamic are arriving at a much higher rate. behavior of the emergency medi The rising time indicates how cine unit of the hospitals under pre quickly the hospital capacity will be scribed hazard conditions and saturated. Adjusting time reveals damage scenarios, and eventually to how long the services delay prob develop a cost vs. risk evaluation lem will last. The peak response procedure leading to decision-mak time indicates when the worst case ing support for seismic retrofit. will be and PO% describes how Some pertinent references con bad the situation could be. With cerning retrofit strategies for hos these evaluations, we may be able pitals are provided. References------ Chong, W.H. and Soong, T.T., Sliding Fragility of Unrestrained Equipment in Critical Facilities, Technical Report MCEER-OO-O005, University at Buffalo, July 2000. Coyle, R.G., Management System Dynamics, John Wiley & Sons, 1977. Forrester, ].w., Industrial DynamiCS, The MIT Press, 19S0. Johnson, G.S., Sheppard, R., Quilici, M.D., Eder, S.]., and Scawthorn, C. R., Seismic Reliability Assessment of Critical Facilities: A Handbook, Supporting Documentation, and Model Code Provisions, Technical Report MCEER-99-000S, University at Buffulo, April 1999. Roth, c., "Logic Tree Analysis of Secondary Nonstructural Systems with Independent Components," Earthquake Spectra, Vol. 15, No.3, 1999. Shinozuka, M., Grigoriu, M. (Coordinating Author), Jngraffea,A.R., Billington, S.L., Feenstra, P., Soong,T.T., Reinhorn,A.M. and Maragakis, E., "Development of Fragility Information for Structures and Nonstructural Components," MCEER Research Progress and Accomplishments, May 2000. Soong, T.T. (Coordinating Author), Yao, G.c., and Lin, c.c., "Damage to Critical Facilities Following the 921 Chi-Chi, Taiwan Earthquake," MCEER Research Progress and Accomplishments, May 2000. Zhu, Z. and Soong, T.T., "Toppling Fragility of Unrestrained Equipment;' Earthquake Spectra, Vol. 14, No.4, 1995. 148 Passive Site Remediation for Mitigation of Liquefaction Risk by Patricia M. Ganagher and James K. iWtchel1 Research Objectiv~s Passive site remediation is a new concept proposed for non-dis National Science Foundation, ruptive mitigation of liquefaction risk at developed sites susceptible to Earthquake Engineering liquefaction. It is based on the concept of slow injection of stabilizing Research Centers Program materials at the edge of a site and delivery of the stabilizer to the target location using the natural groundwater flow. Stabilizer candidates need to have long controllable gel times and low viscosities so they can flow into a liquefiable formation slowly over a fairly long period of time. Colloidal silica is a potential stabilizer for passive site remediation because at low Patricia 11. Gallagher, concentrations it has a low viscosity and a wide range of controllable gel Assistant Professor, times of up to about 200 days. Loose sands treated with colloidal silica Depar.hnentoj'Cimland grout had significantly higher deformation resistance to cyclic loading than Architectural Engineering, untreated sands. Groundwater and stabilizer transport modeling were done Drexel Uniz:erSity, Former Graduate Assistant, to determine the range of conditions where passive site remediation might Department oj'Ciztil be feasible. For a 200-foot by 200-foot treatment area with a single line of Engineering, Virginia injection wells, it was found that passive site remediation could be feasible Polytechnic Institute and in formations with hydraulic conductivity values of 0.05 cm/s or more and State Unit;erstity hydraulic gradients of 0.005 and above. James K. lfitchell, Unil-ersity Distinguished Professor Emeritus, t many sites susceptible to liquefaction, the simplest way to mitigate Departm 149 natural or augmented groundwater The objective of this study was flow. The concept is illustrated in to establish the feasibility of passive Figure 1. site remediation.The work included • Mr. Chris Gause of The set time of the stabilizer identification of stabilizing materi Master Builders would be controlled so there would als, a study of how to adapt or de contributed materials for be adequate time for it to reach the sign groundwater flow patterns to laboratory testing of desired location beneath the site microfine cement grouts deliver the stabilizers to the right and expertise on grouting prior to gelling or setting. If the place at the right time, and an evalu technology. natural groundwater flow were in ation of potential time require adequate to deliver the stabilizer to ments and costs. • Dupont provided some of the right place at the right time, it the colloidal silica used in could be augmented by use of low the experimental program. head injection wells or Performance Criteria downgradient extraction wells. and Identification of • Dr. Ernst Ahrens, Sandia National Laboratory and Once the stabilizer reached the Potential Stabilizers Mr. Henry Pringer, Blue desired location beneath the site, it Circle, contributed cement would gel or set to stabilize the for For a stabilizer to work in this for testing. mation. application, it should have a low Passive site remediation tech viscosity and a long induction pe niques could have broad applica riod between mixing and the on tion for developed sites where set of gelation. Once gelation starts, more traditional methods of ground it should proceed rapidly. The sta improvement are difficult or impos bilizer should also be permanent, sible to implement. It would be less nontoxic and cost-effective. Mate disruptive to existing infrastructure rials evaluated as potential stabiliz and facilities than existing ground ers included colloidal silica, improvement methodsAdditionally, microfme cement grouts, chemical access to the entire site would be grouts, zero-valent iron, and unnecessary using this technology, ultramicrobacteria. Colloidal silica and normal site use activities would was selected as a potentially suit probably not need to be disrupted. able stabilizing material because it Finally, excessive deformation and has a wide range of gel times and a disturbance of the grOlmd around low viscosity. Colloidal silica is an and beneath existing structures aqueous suspension of tiny silica could be avoided. particles that can be made to gel pas"ivesitere~ed!:ltio~isa new technology fo.. niltiga- . . tion~f liqpef~cti~ii.Cl11d ground failure risk at developed ,and.!nacce"sib~e ~~t~swhere more tradition~ !Jletho4Sfor gt"oun 150 II Figure 1. Passive treatment for mitigation of liquefaction risk by adjusting the pH or the salt con times is questionableAcrylamide is centration of the solution. Gel times a neurotoxin in powdered form, so of more than 200 days have been it was eliminated due to environ measured in laboratory tests.Addi mental, safety, and handling con tionally, the initial viscosities of di cerns. Additionally, it is very lute solutions of colloidal silica are expensive.Acrylate was eliminated about 2 centipoise (water=l cP) due to durability concerns. Epoxy and the viscosities remain very low and polysiloxane were rejected for most of the induction period. because they are very expensive. Microfine cement grout was Zero-valent iron is extremely sensi eliminated because its viscosity is tive to oxidation and reduction, so too high to meet the necessary re it would be difficult to treat a large quirements for passive site area and the minerals precipitated remediation. Additionally, since ce would probably not be chemically ment grouts are particulate suspen durable. mtramicrobacteria might sions, the particles tend to settle in be able to clog the pores of a for the suspension and further increase mation with a biofiIm, but biofilms the viscosity. Numerous chemical can be dissolved by strong oxidants grouts were considered. All were such as bleach, so there are dura eliminated as potentially suitable bility concerns. stabilizers, but for different reasons. Sodium silicate was eliminated be cause gel time is not well controlled Feasibility at long gel times. Additionally, the The feasibility of passive site chemical durability of sodium sili remediation depends on the an cate formulations with long gel swers to the following questions: Passiz.1e Site 2?gmediation for fMitigation ofLiquefaction 'J?...isl( 151 1. Will the colloidal silica grout ad An example is shown in Figure 3 equately stabilize the soil? for two samples at a relative den 2 Can the stabilizer be delivered sity of 22 percent that were tested 0.27. Program 2: Seismic Retrofit of to the liquefiable formation and at a cyclic stress ratio of The Hospitals achieve adequate coverage cyclic stress ratio is defined as the within the induction period of ratio of the maximum cyclic shear • Task 2.], Geotechnical the grout? stress to the initial effective confm Rehabilitation Site and 3. How much will it cost? ing stress. The untreated sample Foundation Remediation Strength testing of stabilized strained 1 percent in 11 cycles and sands was done to address the first collapsed in 13 cycles.The sample • Task2.6,MEDAT-2 issue. Groundwater and stabilizer Workshop on Mitigation of treated with 10 weight percent Earthquake Disaster by transport modeling were done to colloidal silica was tested for 400 Advanced Technologies determine if the stabilizer could be cycles. It strained less than about delivered to the formation within half a percent in 11 cycles, about 8 the induction period of the grout. percent in 400 cycles, and never Finally, a preliminary cost analysis collapsed. Only the first 40 load was done to address the fmal issue. ing cycles are shown in Figure 3. These results are typical for samples treated with 10 percent Strength Testing of colloidal silica by weight. For com Stabilized Sands parison, a magnitude 7.5 earth quake would be expected to Cyclic triaxial tests were done on generate about 15 uniform stress Monterey No. 0/30 sand samples cycles. treated with colloidal silica grout to Samples stabilized with concen investigate the influence of colloi trations of 15 and 20 weight per dal silica grout on the deformation cent colloidal silica experienced properties Of loose sand (relative very little (less than two percent) density, Dr 22%). The grain size = strain during cyclic loading. Sands distribution of Monterey No. 0/30 stabilized with 10 weight percent sand is shown in Figure 2. Distinctly colloidal silica resisted cyclic load different deformation properties ing well, but experienced slightly were observed between grouted more (up to eight percent) strain. and ungrouted samples. Untreated Overall, treatment with colloidal samples developed very little axial silica grout significantly increased strain after a few cycles of loading the deformation resistance ofloose and prior to the onset of liquefac sand to cyclic loading. tion. However, once liquefaction was triggered, large strains oc curred rapidly and the samples col Groundwater and lapsed within a few additional Stabilizer Transport cycles. In contrast, grouted sand samples experienced very little Modeling strain during cyclic loading. What Stabilizer delivery is the main fea strain accumulated did so uniformly sibility issue with respect to passive throughout loading and the site remediation. Preliminary samples remained intact after cyclic groundwater and solute transport loading. modeling were done using the codes MODFLOW, MODPATH, and 152 MT3DMS for a generic liquefiable formation. A "numerical experi ment" was done to determine the ranges of hydraulic conductivity and hydraulic gradient where pas .... sive site remediation might be fea ~ 40+------H--+~H----~---+-----! sible. For a 200-foot by 200-foot u: _ 20+---4-~-~--~---+--~ treatment area, with single lines of c: injection and extraction wells, ...~ 0+---4-~~~~~~---+-----! CI) travel times through the treatment 0.. 10 1 0.1 0.01 0.001 0.0001 area will be about 100 days or less Particle Size (mm) if a formation has a hydraulic con ductivity greater than about 0.05 II Figure 2. Gradation curve for Monterey No. 0/30 sand. cmls and a hydraulic gradient higher than about 0.005. Based on the possible gel times, this time frame is considered feasible. Extrac Untreated Monterey Sand {Dr =22%} tion wells will increase the speed of delivery and help control the ..-.. 8.0 ~ down gradient extent of stabilizer ~ 6.0 movement. .~ 4.0 The results of solute transport Z 2.0 en_ 0.0 .. l"!,I,,!, I modeling indicate that stabilizer I'G T 'x -2.0 I delivery will vary throughout the I treatment area. A typical stabilizer « -4.0 I I I I contour plot for a hypothetical for o 10 20 30 mation with a uniform hydraulic Cycles conductivity of 0.05 cm/s and a hydraulic gradient of 0.005 is Monterey Sand with 10% Conoidal Silica shown in Figure 4.A stabilizer con (Dr = 22%} centration of 100 gil would be de ~ 8.0 livered through an infiltration ~ 6.0 trench for 100 days. The best cov c: 4.0 _~ 2.0 erage would be achieved close to (J') _ 0.0 ,-- ~-."' ...... I.A.a •• ~Il',. '.~'~lUl1!l~,!\~ the source of the stabilizer. Concen .!.'!! trations would decrease laterally x -2.0 -4.0 I I away from the source and down « gradient of the source. If the mini o 10 20 30 mum amount of stabilizer required Cycles for adequate stabilization could be delivered to the majority of the II Figure 3. Axial deformation during cyclic loading (CSR=O.27) for treated and treatment area, it is likely that the untreated sand. formation would be stable enough to withstand seismic loading.How ever, there could be some differen tial or variable response across the site. It may be necessary to deliver a higher concentration at the up gradient edge of the treatment area Passive Site 1(g;metfiation for Mitigation ofLiquefaction 'l{isl( 153 in order to get an adequate concen from point to point within the for tration at the down gradient edge. mation.An example stabilizer con Heterogeneity in the formation tour profile through a treatment will actually control how well the area with a variable hydraulic con stabilizer can be delivered. If the for ductivity is shown in Figure 5. In mation is highly variable, then the this case, the hydraulic conductiv stabilizer concentration will vary ity was varied slightly in each layer as shown for a total variation throughout the layer of about one order of magnitude.The remainder of the simulation is the same as the previous case. The layers with higher hydraulic conductivity have a higher concentration at the down gradient edge. These layers would probably be more stable than lay ers with lower hydraulic conduc tivity that receive a lower concentration of grout during the treatment period. However, even if the regions oflower hydraulic con ductivity liquefy, the presence of very stable seams will likely lessen o 100 the severity of the overall deforma tion. Accurate characteriZation of the hydraulic conductivity through • Figure 4. Stabilizer contours for 200 ft. by 200 ft. treatment area (outlined in black) after 100 days of treatment. Stabilizer delivered through inftltra out the treatment area will be es tion trench at concentration of 100 gil. 1\vo extraction wells at the down sential for successful treatment by gradient edge withdraw a total of 7500 cfd. Contour intervals are 10 gil. passive site remediation. Concentration at extraction wells is 60 gil. Travel paths for individual wa ter particles are superimposed over the treatment are-a in 10-day incre ments. Particle travel times are about 75 to 80 days. ;:-----...... '\:~""::=~;:::::::===:::::::;;,...... ,\JlW~~----___, k (cm/s) 0.010 0.075 0.050 0.100 0.025 o Horizontal and vertical 100 • Figure 5. Stabilizer profile through centerline of 200 ft. by 200 ft. treatment area after 100 days of treatment. Stabilizer de livered through infiltration trench at concentration of 100 gil. Extraction wells at the down gradient edge withdraw a total of 7500 cfd. Contour intervals are 10 gil. Concentration at extraction wells is about 70 gil in lower 30 ft. Travel paths for indi vidual water particles are superimposed over the treatment area in 10-day increments. Particle travel times range from about 40 to 420 days. 154 Cost Delivery of the stabilizer is the central feasibility issue with respect The cost of passive site to passive site remediation. For a remediation is expected to be com 200-foot by 200-foot treatment area parable to other methods of chemi with a single IiUe of injection wells, cal grouting. It is likely that a 10 it was found that passive site weight percent concentration of remediation could be feasible in colloidal silica will be adequate to formations with hydraulic conduc stabilize a liquefiable formation. It tivity values of 0.05 crn/s or more is possible that lower concentra and hydraulic gradients of 0.005 tions could be used. Based on a 10 and above. However, the actual con percent concentration, it is ex centration proflle across the site pected that materials costs would will depend on the variation in hy be in the range of $120 to $180 per draulic conductivity throughout cubic meter of treated soil. These the formation. costs are competitive with other The anticipated final outcome of methods of chemical grouting. this work is a new technology for mitigation ofliquefaction and ground Conclusion fuilure risk. Passive site remediation technology will be less disruptive to Based on the feasibility analysis, existing infrastructure and facilities passive site remediation appears to than existing methods.It is expected be a promising new concept for that passive site remediation will be mitigation of liquefaction risk. At cost-competitive with other methods this time, a minimum concentration of chemical grouting. Model testing of 10 percent colloidal silica ap of both the injection method. and the pears to be suitable for stabilizing performance of grouted ground is liquefiable sandsAdditiorial testing planned as the next step in the evalu is being done with concentrations ation of this new technology. It will of 5 weight percent to determine be done using a geotechnical centri if the level of strain during cyclic fuge equipped with a shake table. loading would be acceptable. Passi'lJe Site !f(g.meaiation for Mitigation ofLiquefaction :/(isl( 1;; Page Intentionally Left Blank Advanced GIS for Loss Estimation and Rapid Post-Earthquake Assessment of Building Damage by Thomas D. O'Rourke, Sang-Soo Jeon, Ronald T Eguchi and Cbarles K Huyck Research Objectives The objectives of the research are to: 1) develop regressions between National Science Foundation, building damage and various seismic parameters to improve loss estima Earthquake Engineering tion, 2) identify the most reliable seismic parameter for estimating build Researcb Centers Program ing losses, and 3) develop GIS-based pattern recognition algorithms for the identification of locations with the most intense post-earthquake build ing damage. This latter objective is coupled to the goal of improving emer gency response as part of the MCEER vision of creating earthquake resilient Thomas D. O'Rourke, communities. GIS-based technologies for visualizing damage patterns pro Thomas R. Briggs Professor vide a framework for rapidly screening remote sensing data and dispatch o/Engineering and Sang SooJe011, Graduate ing emergency services to the locations of greatest need. Research Assistant, Deparlmentoj"Cimland Environmental Lifeline Damage Engineering, Cornell Unimrsity CEER researchers working on the seismic retrofit and rehabilitation Ronald T. Egucbi, President M of lifelines (program 1) and on seismic response and recovery (pro and Charles K Huyck, gram 3) have collaborated on a project organized to apply advanced GIS Vice President, ImageCat, techniques for the rapid identification oflocations with most intense build Inc. ing damage.This collaboration has resulted in the development of regres sions between building damage and various seismic parameters, as well as the application of GIS-based recognition algorithms with the potential for screening remote sensing data to identify areas of highest post-earthquake damage intensity. Using GIS technology, Cornell researchers developed the largest U.S. database ever assembled for spatially distributed transient and permanent ground deformation in conjunction with earthquake damage to water sup ply lifelines (O'Rourke et aI., 1999).This research has helped to delineate local geotechnical and seismological hazards in the Los Angeles region that are shown by zones of concentrated pipeline damage after the Northridge earthquake.The research has resulted in regressions between repair rates for different types of trunk and distribution pipelines and various seismic parameters. The regressions are statistically reliable and have improved predictive capabilities compared with the default relation ships currently used in loss estimation programs.They will be referenced in the next version of HAZUS software that implements the National Loss 157 ~IJlI"bor"t.ite 0.6 Partners 0.5 .~ • California Office of r" Emergency Services 0.4 (OES) •. ~ Fit Curve 8 0.3 • Los Angeles Department All LA :J AIISFV ofWater and Power .'0 .~ 0.2 Section of SFV Federal Emergency J;:" Sherman Oaks • .~ Management Agency 0.1 * North of LA Basin • San Francisco 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Dimensionless Grid Size, DGS • Figure 1. Hyperbolic Fit for Threshold Area Coverage and Dimensionless Grid Size for Pipeline Damage Patterns (Toprak, et aI., 1999) Estimation Methodology spon With GIS, the spatial distribution of sored by FEMA.The regressions and damage can be analyzed by dividing statistical databases are being incor any map into squares, or cells, each of porated in pre-standards for esti which is n by n in planA repair rate is mating water supply losses defined as the number of repairs di developed by theAmerican Lifeline vided by the total distance ofpipelines Alliance through ASCE under con in each cell. Contours of equal repair tract with FEMA. rate, or damage rate, can then be The research has led to the discov drawn using the grid of equi-dimen ery of a relationship for visualizing sional, n-sized cells. If the contour in damage patterns by linking the two terval is chosen as the average repair dimensional representation of local rate for the entire system or portion damage with the grid size used in GIS of the system covered by the map, to analyze the spatial distribution of then the area in the contours repre data.This concept is illustrated in Fig sents the zones of highest (greater ure 1, which shows the relationship than average) earthquake intensity as developed for visualizing post-earth reflected in pipeline damage. quake pipeline damage (O'Rourke et al.,1999). The users of this research include two main groups: 1) water distribution companies, such as the Los Angeles De partment of Water and Power (LADWP) and the East Bay Municipal Utility District (EBMUD) and 2) those who ben efit from improved loss estimation, such as insurance com panies, the Federal Emergency Management Agency, state and municipal planners, and emergency service providers. 158 A hyperbolic relationship was principally single and multiple fum shown to exist between the thresh ily residences The inspection records 'oCu~ old area coverage (TAC) [in this include location and estimate ofdam Res~ case, the fraction of the total map age as a percentage of replacement Program 1: Seismic Et'aluation area with damage exceeding the cost. Figure 2 shows the types of tim and Retrofit ofLifeline overall average repair rate] and the ber frame structures included in the Networks dimensionless grid size, defined as database with examples of damage, 2 Program 3: Earthquake the square root of n , the area of an as presented by the NAHB Research Response and Recovery individual cell, divided by the total Center (1994).The database did not map area,~.This relationship is il include specific infonnation about lustrated in Figure 1, for which a the type ofdamage to each structure. schematic of the parameters is pro The number of one to two story vided by the inset diagram. The re timber frame structures affected by lationship was found to be valid the No~dge earthquake was esti over a wide range of different map mated from tax assessor records sup-' scales spanning 1200 km2 for the plied through OPS.The numbers of entire Los Angeles water distribu structures determined in this way tion system affected by the was 278,662As a check,an alternate Northridge earthquake to 1 km2 of number was estimated from 1990 the San Francisco water distribu census block data (Wessex, 1996) by tion system in the Marina affected evaluating the number of buildings by the Lorna Prieta earthquake associated with detached and at (foprak et at, 1999) tached single housing units in com As explained by O'Rourke et al. bination with the buildings needed (1999), the hyperbolic relationship to accommodate two and three to can be used for damage pattern rec four housingunits.The resulting num ognition, and for computer "zoom ber was 267,868, which is in excel ing" from the largest to smallest lent agreement with the 278,662 scales to identify zones of concen units estimated from tax assessor trated disruption. Ifsuch a relation records. ship can be shown to be valid for building damage, then remote sens ing data acquired to characterize building damage can be evaluated rapidly for the locations of highest loss intensity and thereupon tar geted for emergency response. Building Damage Inspection records available through the California Office of Emergency Services (OPS) were ob tained for 62,020 buildings thatwere investigated after the Northridge earthquake. Of these, 48,702 build ings were associated with one and two-story timber frame structures, III Figure 2. Typical Damage to 1-2 StoryTJillber Frame Buildings after the Northridge Earthquake (after NAHB, 1994) Advanced GIS for Loss E..,timation and Rapid Post-Eartbquake Assessnzent of Building Damage 1,""9 Loss Estimation Regressions A GIS grid composed of 2, 106 cells was created, and the num ber of relevant structures within each 0.42 km2 cell was calculated and geocoded at the cell center. Figure 3 shows the GIS grid developed with tax as sessor records superimposed on the spatial variation of peak ground velocity determined from more than 240 strong mo tion records geocoded as part of this study. By superimposing a grid in which each cell is char • Figure 3. GIS Grid with Ta.~ Assessor Data Superimposed on Peak Ground acterized by various damage ra Velocities tios linked with damage factors, linear regressions can be per- Two damage parameters, referred formed to quantify damage vs. seis to as damage ratio and damage fac mic excitation for loss estimation tor, were calculated with the data purposes. base. Damage ratio is the fraction Figures 4a and b show the linear or % of existing stnlctures with regression of damage ratio (DR) vs. damage equal to or exceeding a peak ground velocity (PGV) and particular damage factor. Damage Spectmm Intensity (SI) for various factor is damage expressed as a % damage factors (DF). Most regres ofbuHding replacement cost. sions for PGV show excellent fits ~ 1.00 'i 1.00 ~ . • o ~ ex:~ ex: (J) (J) g> g> E E ~ 0.10. ~ 0.10. eDF • Figure 4. Damage Ratio Regression with Peak Ground Velocity (PGV) and Spectrum Intensity (SI) for 1-2 Story Tunber Frame Buildings 160 as indicated by high r2, although the "goodness" of fit is relatively low fot" DF ~ 70%. In contrast, all regressions for SI have high r2 and excellent characteristics with re spect to residuals. The relation ships between damage and various seismic parameters were probed in this way to determine which parameters were statisti cally most significant as damage predictors. The seismic param 0.01 +.=-,..,..=.."". -,.,.,.",. "". -,.,..,...,-" .. = 13 2 4 66 Z oC- 68 2. 46a 2. 46a 2. 46"8 2. 4 sa :1 4 sa eters investigated include the mea 0.1 1.0 10.0 100.0 0.1 1.0 10.0 100.0 sured peak ground acceleration, PGVIDF1.16, ([cm/secJl%1.16) SIIDF1.37, ([cm/sec}/%1.37) velocity, and displacement; spec tral acceleration and velocity for periods of 0.3 and 1.0 s;Arias In II Figure 5. Damage Ratio Regression for ScaledPGV and SI for 1-2 Story Timber tensity; and SI. Frame Buildings SI is defined as the area under the pseudo-velocity, Sv, response spectra curve for a damping ratio, U in which (SP/DFi3/ ) is the scaled of 20 % between periods,T, of 0.1 ~ seismic parameter. and 2.5 s: Using (3), the data in Figure 4a 51= f2-5 SV(T,g) DT (1) and b are replotted in Figures 5a 0.1 and b as linear regressions that ac SI was first proposed by Housner count for DR, DF, and SP. A5 evinced (1959) as a measure of the maxi by high r2 and excellent character mum stresses that would be in istics with respect to residuals, the duced in elastic structures by relationships in Figures 5a and b are ground motion. Katayama et al. statistically significant. (1988) found that house damage Seismic parameters normalized correlated more strongly with SI with respect to DF combine the than with peak acceleration, and effects of strong motion and dam recommended that SI calculations age level in a convenient way that be performed for ~ = 20%. facilitates loss estimation. For ex An examination of Figure 4 re ampie, consider an area for which veals that DR, DF, and the seismic the predicted SI is 30 cm/sec. The parameter, Sp, are interrelated in a % of timber structures in that area consistent way. Using multiple lin that would have damage equal to ear regression techniques, this in or exceeding 10% of the building terrelationship can be expressed as replacement cost is calculated as DR =0.92(30/101.3,)1.05\ or 1.19%. LogDR= LogK +aLogSP -JHogDF For damage equal to or exceeding (2) 50% replacement cost, DR = 0.92(30/501.3')1.051, or 0.11 %. in which K, a, and ~ are con stants. (2) can be rewritten as DR = K(SP IDFWa.yl (3) Advanced GIS for Loss Estimation and Rapid Post-Eartbquake AsseSSlnent ofBuilding Damage 161 lif;';;;;;;=;;:;~~» Damage Pattern hyperbolic surface for which the initial slopes and asymptotes vary an advanced Recognition as a function of DE GIS with Using the same concepts that Experience has shown that the advanced were applied to the spatial distri ideal TAC for visualization is about remote sensing, bution of pipeline damage, investi 0.33. Figure 6 indicates that the di MCEERis gations were undertaken to define mensionless grid size to achieve developing the a relationship between TAC (de this TAC varies significantly, de enabling fined for buildings as the fraction pending on DE technologies of the total map area with DR ex Figure 7 illustrates the applica tion of Figure 6 for damage inten fora new ceeding the overall average DR) sity recognition and computer generation of and the dimensionless grid size.The grid containing tax assessor data "zooming. "The top map shows the emerge.ncy was sufficiently refined to evaluate entire San Fernando Valley and ad response and the effects of progressively larger jacent areas. The figure shows a rapid decision grid sizes by developing grids with cascade of different computer support cell sizes of n by n, 2n by 2n, 4n by screens that "zoom" on increasingly systems." 4n, and Sn by Sn. The data gener smaller areas from top to bottom. ated for these grids and various DFs Each screen shows as blue dots the are plotted in Figure 6. All relation actual locations of one to two story ships are well described by hyper timber frame structures with DF ~ bolic functions, similar to the one 70% after the Northridge earth for pipeline damage in Figure 1. Be quake. The red contour lines indi cause building damage is character cate areas of highest damage ized by DF, it has an additional intensity. dimension when compared to the The contour lines in the top map single hyperbolic curve for pipe were generated by analyzing the lines. In fact,Figure 6 is actually a data with a GIS grid size chosen for TAC =0.33 from Figure 6.The area 0.7 outlined in red in the top map was identified for more detailed assess 10 % 0.6 +-+--+-+--d;;;,.--+~t=::::t== ment. The grid size for this assess ment was again determined forTAC ~ I-' 0.5 20% = 0.33 from Figure 6. The map in oi ------01 ---- the center of the figure shows an e 70% Q) > 0.4 expanded view of the area outlined 0 0 30% above after computer analysis. The 11:1 Q) 0.3 zones of highest damage intensity 162 Contours Indicate Highest Intensity of Damage II Figure 7. GIS Evaluation of BuiJding Damage for DF ~ 70 % Increasing density of contour lines Because the relationships in Fig reflects increasing intensity of dam ure 6 are independent of scale, the age.The scale embodied in the bot same algorithm for a specific DF tom map allows the viewer to can be used for increasingly smaller discriminate damage patterns on portions of the Original map to virtually a block-by-block basis. "zoom" on areas of most intense damage.The visualization algorithm Advanced GIS for Loss Estimation and Rapid Post-Ea-rtbquake Assessment of Building Damage 163 allows personnel who are not spe Summary cifically knowledgeable about structures or trained in pattern rec GIS research on visualizing dam ognition to identify the locations of age patterns in pipeline networks most severe damage for allocation has been extended to buildings. of aid and emergency services.The Algorithms developed for pipelines entire process is easy to computer have been modified and validated ize, and personnel would be able to chose optimal GIS mesh dimen to outline any part of a map with a sions and contour intervals for vi "mouse" and click on the area so sualizing post-earthquake damage defmed. In each instance of defin patterns in buildings.This work has ing a smaller area for evaluation, the been performed by a collaboration average DR is recalculated for the with researchers working in Pro new,smaller-sized map. In this way, grams 1 and 3 on advanced tech the average is calibrated to each nologies for loss estimation and new map area. real-time decision support systems. When the damage pattern recog Robust and statistically significant nition algorithms are combined with regressions have been developed regional data rapidly acquired by ad between the fraction of existing vanced remote sensing technologies, timber frame buildings at any dam the potential exists for accelerated age state and the magnitudes of management of data and quick de various seismic parameters. Such ployment of life and property saving work improves loss estimation sig services. By combining an advanced nificantlyand also creates advanced GIS with advanced remote sensing, technology to visualize post-earth MCEER is developing the enabling quake damage patterns in buildings technologies for a new generation of for rapid decision support and de emergency response and rapid deci ployment of emergency services. sion support systems. References------Housner, G. w., (1959), "Behavior of Structures During Earthquakes," Journal of Engineering Mechanics Division, ASCE, Vol. 85, No. EM4, pp. 109-129. Katayama, T., Sato, N., and Saito, K., (1988), "SI-Sensor for the Identification of Destructive Earthquake Ground Motion," Proceedings, 9'b World Conference on Earthquake Engineering, Vol. VII, Tokyo-Kyoto, Japan, August, pp. 667.(,72. NAHB Research Center, (1994), Assessment of Damage to Residential Buildings Caused by the Northridge Earthquake, Reported Prepared for U.S. Department of Housing and Urban Development by NAHB Research Center, Upper Marlboro, MD. O'Rourke, Thomas D., Toprak, Selcuk, and Jeon, Sang-Soo, (1999), "GIS Characterization of the Los Angeles Water Supply, Earthquake Effects, and Pipeline Damage," Research Progress and Accomplishments 1997-1999, Multidisciplinary Center for Earthquake Engineering Research, University at Buffalo, July, pp. 45-54. Toprak, S., O'Rourke, T.D., and Tutuncu, I., (1999), "GIS Characterization of Spatially Distributed Lifeline Damage," Proceedings, 5 tb US Confet'ence on Lifeline Earthquake Engineering, Seattle, WA, ASCE, Reston, VA, August, pp. 110-119. Wessex (1996), US. Demographics, Wessex, Inc., Wmnetka, IL. 164 Seismic Evaluation and Retrofit of the Ataturk International Airport Terminal Building by Michael C. Constantinou, Andrew ~~ WlJittaker and Emmanuel Velivasakis Research Objectives MCEER was part of a team charged with the seismic upgrade of LZA Technology, Thornton Tomasetti Group Istanbul's Ataturk International Airport Terminal Building. The project involved analysis of post-earthquake damage to the Terminal Building National Science Foundation, following the 1999 Marmara,Turkey earthquake, and development of a Earthquake Engineering retrofit scheme. Nonlinear static and dynamic analyses were conducted Research Centers Program using procedures set forth in FEMA 273 (including numerous contri butions from MCEER researchers). Pushover analyses were conducted using IDARC (developed with MCEER support). Dynamic analyses of the roof-isolated structure considering inelastic frame behavior were Michael C. Constantinou, conducted using computer software based on modifications of pro Professor, AmlreUJ S. gram 3D-BASIS (developed with MCEER support)~ Several retrofit Whittaker,Assodate schemes were investigated, including conventional methods and ma Professor, Oscor Rnmirez, terials, and new technologies. The scheme selected included isolation Graduate Research of the roof trusses, addition of shock transmission units, and selected Assistant, Ani Natali retrofit and strengthening of reinforced concrete construction. Sigaber, Graduate TIle Turkish build-operate-transfer consortium,TEPE-AKFEN-VIE (TAV) Research Assistant, Eric Wolff, Graduate Research and its advisor, NewYork-based Turner International, approved the evalu Assistant, and Tony Yang, ation and retrofit scheme. Members of the team were LZA Technology, Undergraduate Student, a division of Thornton-Tomasetti Group (leader), Michael C. Department OfCi'liil, Constantinou and Andrew S.Whittaker (formerly of PEER), both of the Structural and University at Buffalo, and Tuncel Engineering and Fondsiyuon Envlronmental Muhendislik Insaatvetic Ltd., both of Istanbul. The project is an excel Engineering, University at lent example of bringing research results to implementation. Buffalo Emmanuel Velivt1sakis, Senior Vice President, LZA Technology, New York, New A t the time of the August 19, 1999,Izmit earthquake, the new Ataturk York fiInternationalAirportTerminal building was nearing completion.The airport, which is located 25 km from the center of Istanbul, was shaken and damaged by the earthquake. (The airport is located approximately 70 km from the fault rupture plane.) The new Terminal building is a three story reinforced concrete building with a space-fuune roof.The plan foot print is approximately 240 m by 168 m.A view of the terminal building is presented in Figure 1. 165 ing showing typical framing. The building is framed in reinforced • Eric Stovner and concrete. Above the John Abbrnzzo, LZ4 third floor level, cantile Teclmo/()gy ver columns on 24 m centers supported a • Anoop Mokha, Vice President, Earthquake three-dimensional steel Protection Systems space-frame roof struc ture. The space-frame • Douglas Taylor, Taylor roof was equipped with Devices • Fif,rure 1. New Maturk International Airport Terminal sleeved movement Building joints to permit thermal • Farnk Tuncel, Engineer expansion and contrac ofRecord, Tunce! Engineering and The lowest story of the building tion of the roof. These movement Construction Co., Ltd., provided mechanical and baggage joints did not align with the expan Turkey handling services. The second and sion joints in the reinforced con third stories of the building housed crete framing.At the third floor and • ThollUlSJ. McCool, the Arrivals and Departures Halls, below, gravity loads were sup Project Engineer and respectively. A plan view of the ported by reinforced concrete Construction Consultant, building is shown in Figure 2a. As waffle slabs and columns at 12 m Turner Steiner International, New York shown in Figure 2a, the 240 m by on center. Lateral loads were re 168 m building is composed of 20 sisted by waffle-slab moment-frame • Sani Sener, CEO, and pods (or independent frames); the construction in each direction. Gokhan Ozber, Deputy typical pod dimension is 48 m by Solid beams of a depth equal to that GeneralManager, Tepe 48 m. A typical pod is shown of the waffle slab spanned between Akfen Vie (1ftv), shaded in this figure.The pods were the columns.Around the perimeter Investment Construction separated by 50 mm wide expan of each pod, the column sizes were and Operation Inc., Turkey (Airport Owner) sion joints (EJs). Figure 2b shows substantially reduced from those in the framing in the third and second the interior of the pod.At the cor stories of a typical pod; the fram ners of each pod, the four columns ing in the first story was similar to were approximately square but that in the second story. Figure 2c with dimensions one-half of those is a cross-section through the build- of the interior columns (and called It'i~ anticipate(j;that this i>t(}leC,t wiilJ)e of interc.;st to bullditlgoWfl itiseis~Q.regJ.<>,~t:bfQ1igJtout the world. Three" differen(ers retrofit aPJ:lroaclj~~,~eretaken, to . strengthen vari()u~ secti0ns,C)~!lt~Jjl1ikting.The' perfor~ . mance of theseismicisola~~n~d.l~~:~p deviC~s in fu ture seis~c events wll111rO\rideadc;U~ional data on their reliability and effectivenessforeaftllquake hazard miti- gation. The entire evaluation,and subsequent installation of the retrofit scheme was accomplished in less than four months. ..' 166 240m ~ • The seismic ez;aluation 24 m procedures used in this t project were tk/;-eloped with full or partial.MCEER including!DARe 168 m support, 48 •m 4.0, 3D-BASIS, FEJfA273/274 , ~ and NEHRP 2000. \\ ~ 48m L ~ 50 mm EJ (typ.) a. Plan view of the Terminal building Quarter-col Fu1tc( 3rdFioor ~iip::J 2nd Floor ~~ Second story 1 b. Plan view of framing in one pod c. Cross-section showing typical framing II Figure 2. Construction of the Terminal Building quarter columns in this paper). acceleration was less than O.05g. In Along each edge of each pod, and vestigations by LZA engineers im between the corner quarter col mediately following the earthquake umns, the columns were approxi identified damage to parts of the mately rectangular and half the area building, including spalling of the ofthe typical interior columns (and cover concrete and buckling of the called half columns in this paper.) longitudinal rebar at the base ofthe During the August 19,1999 earth third story columns (showing also quake, the Terminal building was lap splices and little transverse subjected to modest earthquake rebar in the hinge zones), loss of shaking. The maximum recorded concrete at the underside of the horizontal ground acceleration re roof truss-cantilever column con corded at theAirport was approxi nections and slippage of the roof mately OJ g; the maximum vertical truss baseplates atop the columns, Seismic 'Evafuation aM1(etrofit of the 5itaturf( Intemationaf.'il.irport 'Tennina{ '13uiUfing 16 j ·stt.sOverall Project Information: LZA Technology: http://www.lzagroup.com Hardware Information: Earthquake Protection Systems: httpJlearthquakejJrotection.com Taylor Devices: http://wu!w.taylordelJices.com • Figure 3. Obsmed Damage to the Terminal Building; base of third story column (left) and base of roof truss connection to column (right) and splitting cracks and spalled m wide above the 3rd floor level concrete in the beam-column joints and 12 m wide at that level and at the third floor level (indicating below. no transverse reinforcement in the The nonlinear static (or push joints). Photographs of damage to over) analysis was conducted using the building are shown in Figure 3. the procedures set forth in FEMA 273 (FEMA 1997).The existing fram ing was modeled using the as-built Seismic Evaluation of construction drawings. The tribu Tenninal Building tary widths of the waffle-slab or beam framing for stiffness calcula One frame in one of the typical tions was set equal to 12 m; the 48 m by 48 m pods was selected strength of the beam framing was for evaluation by nonlinear static based on the reinforcement in the analysis.The objectives of the analy solid segments of the waffle slabs sis of the existing building were to between the column. The third (a) correlate the locations of the story columns were linked at the observed damage and that pre roof level by rigid axial elements to dicted by analysis, (b) to estimate simulate the effect of the space the displacement capacity of the frame roof structure.The deforma existing framing system, and (c) to tion capacities of the reinforced provide guidance to the design concrete components were set team on plausible retrofit schemes. equal to the values listed in Chap The lightly shaded zones in Figure ter 6 of FEMA 273 for reinforced 2b indicate the location and width concrete columns, beams, and of the sample frame.The central bay beam-column joints. Because little of framing was selected because it transverse reinforcement was pro included the third story cantilever vided hI the critical or hinging re columns that were damaged during gions, the deformation capacities of the earthquake. This frame was 24 all components were established 168 assuming non-conforming trans verse reinforcement.A sample cal 4000 ~------~ culation for the third-story column 3500 showed a maximum plastic rota 2 Unifonn Pattern tion of 0.01 radian for the perfor mance level of collapse prevention ModalPattcm because the axial load ratio was less Point I: Center 3.rd story column forms a plastic runge. than 0.1, the transverse reinforce Point 2: End 3rd story columns form plastic hinges. ment was non-conforming, and the o 100 200 300 400 shear force ratio was less than 3.A Displacement at top of3rd story column (mm) target displacement for the push over analysis was established based on revised criteria established by II Figure 4. Pushover Analysis of a Frame in the Original Building the owner aftertheAugust 19 earth quake, namely, a spectrum with or to the reinforced framing with one dinates equal to 150 percent of the exception: the beam-column joint elastic spectrum set forth in the damage was not predicted because 1997 Turkish seismic code for the these joints were assumed to be site of the airport.The reSUlting tar rigid for the analysis. For informa get displacement at the roof level tion, the third-story drift corre was 230 mID for an elastic period sponding to a plastic rotation of of 1.25 seconds. 0.01 radian in the third-story col The pushover analysis was ac umn was 80 mm. Clearly; the defor complished using IDARC 4.0 (Valles mation capacity of the third-story et aI., 1996).The existing building columns would be exhausted well was analyzed using two lateral-load before the target displacement at profiles; second-order effects were the roof level was achieved. Further automatically included. For this evaluation of the existing building frame, both a uniform pattern and showed that the moment-resisting a modal pattern were used. The frames were undesirable weak col collapse mechanism involved hing umn-strong beam frames. ing of the third story cantilever col umns for both loading proftles. Figure 4 presents the base shear Conventional and roof displacement relationships for Protective Systems the two lateral-load proftles.The sig Retrofit Concepts nificant difference between the two curves is due to the large dif Retrofit schemes for theTerminal ferences between the weights at building were developed using con the roof and lower levels and the ventional methods and materials resulting differences between the and new technologies.The conven loading profiles. Nearly all of the tional retrofit options considered building deformation occurs in the by the design team all made use of third story of the building with the a new ductile lateral-force-resisting modal load pattern and the fram system, including steel braced ing below the third story does not frames, special reinforced concrete yield_ Such a distribution of pre shear walls, and special reinforced dicted damage is completely con concrete moment frames. All of the sistent with the observed damage conventional retrofit schemes in- Seismic 'Evaluation and!R§trofit of tfie _'4taturl( Internationaf_'4irport fJ"enninaf 'Buifding 169 eluded new foundations under the Retrofit (Upgrade) of new lateral-foree-resisting compo nents, repair and reconstruction of the Tenninal Building the third-story column to roof-truss This scheme selected for the ret framing, elimination of the expan rofit (or upgrade) of the Terminal sion joints between the pods, join building involved the isolation of ing and jacketing adjacent quarter the roof trusses to reduce the de and half columns, and jacketing of mand on the third story columns the interior third-story columns. and the framing at the lower lev The addition of a new lateral-force els, the addition of shock transmis resisting system was not consid sion units to the roof trusses to lock ered feasible because substantial the space-frame truss pods together new vertical elements (walls or during earthquake shaking so that braced frames) could not be in the space frame would act as a dia stalled in or below the first story. phragm, and selected retrofit and New technologies in the form of strengthening of the reinforced seismic isolation and supplemental concrete construction as summa damping were evaluated for the rized below. retrofit of the Terminal building. Because of architectural con Because the addition of supplemen straints, the size of the third story tal damping devices would have columns could not be increased involved the addition of braces or substantially,so the existing flexural wall panels in the lower two sto strength of these columns as canti ries of the building, a detailed ret~ lever elements dictated the inertial rofit design using supplemental force that could be developed at dampers was not prepared. Two the roof level. Fuses in the form of seismic isolation options were con Friction Pendulum (FP) isolation sidered: base isolation of the entire bearings were used to limit the lat building, and isolation of the roof eral forces that could be imposed trusses. The building isolation op on the third-story columns. (FP tion was the preferred option of the bearings were used because such two, but was rejected by the owner bearings can isolate light compo because of the advanced state of nents and structures.) Preliminary construction of the building. The calculations called for an isolated installation of an isolation system period of 3.00 seconds (based on immediately above the foundation the radius of the sliding surface), a would have required the demoli design friction coefficient of 0.09, tion and reconstruction of the and a displacement capacity of ap ground floor of the Terminal build proximately 260 mm. ing, and the removal and reinstalla The quarter and half columns tion of the mechanical and baggage around the perimeter of each pod handling systems located in the first were joined to the adjacent quar story of the building. The second ter and half columns respectively isolation option was studied in de using reinforced concrete. Addi tail, and was selected for the retro tional vertical reinforcement was fit of the building by the owner. placed in the joints between the Information on this retrofit scheme part columns and around the pe is presented below. rimeter of the joined columns to 170 further strengthen the columns. The objective of such strengthen ing was to eliminate the weak col I 17 19 20 umn-strong beam framing. These 5 columns and the modestly strength /""-/""/""/""-/""/"" L , /' ened interior square columns were " " , " , "I " "I " then jacketed with circular steel , 11 ! . , -- casings to substantially increase the " l" "1:' " shear strength of all the columns o 100 200 300 400 500 600 700 800 900 1000 and to provide confinement in po Displacement at top of3rd story corumn (mm) tential hinge regions. Such column strengthening was implemented in the second and third stories. In ad • Figure 5. Pushover Analysis of the Retrofitted Frame dition, the expansion joints be tween the pods at the second and third floor levels were eliminated the roof level of 190 mm is indi by tieing the pods together with cated by the open-ended arrow on reinforced concrete components to the pushover curve of Figure 5.The substantially increase the redun deformation demands on the dancy of the lateral-force-resisting beams and columns in the retrofit system. No beams at either the sec ted building frame at this roof dis ond or third floors were strength placement were considered ened. acceptable for the performance The performance of the retrofit level of collapse prevention. ted building was checked by non Two photographs of the retrofit linear static analysis of the frame work showing an installed FP bear described above. The reSUlting ing (before release) and a jacketed pushover curve for the modal load column (covered by an architec pattern is presented below in Fig tural treatment) are presented in ure 5 together with a sketch show ing the sequence of plastic hinge formation (1 through 20). Note that some hinges form simulta neously. For a roof displace ment less than 500 mm, hinges form in the beams, at the base of the columns above the foundation, and in the two-story columns at the underside of the third story only: The retrofit design was fur II Figure 6. Isolation and strengthen ther evaluated by nonlinear ing of the Terminal building; FP dynamic analysis using 20 isolators atop jacketed columns (above) and jacketed third-story ground motion records that column (left) matched on average the re vised design spectrum de scribed above. The mean maximum displacement at Seismic 'Evafuation and g(gtrofit of the . References------ FEMA, (1997), NEHRP Guidelines for the Seismic Rehabilitation of Buildings, Report No. FEMA 273, Federal Emergency Management Agency, Washington D.C. October. Valles, R.E., Reinhom, A. M., Kunnath, S. K., Li, C and Madan, A., (19%), IDARC2D Version 4.0: A Computer Program for the Inelastic Damage Analysis of Buildings, Report No. NCEER-96-0010, National Center for Earthquake Engineering Research, University at Buffalo, June. 172 Estimating Earthquake Losses for the Greater New York City Area by Andrea S. Dargush (CoordinatingAutbor), jrJichaelAugustyniak, George Deodatis, Klaus HJacob, Lau1Yl iJ1cGinty,r George Mylonakis, Guy].P. Nordenson, Daniel O'Brien, Scott Stanford, Bruce Swire-il, Michael W. Tantala and Smn Wear Research Objectives The goal of the project is to use the federally sponsored loss estimation National Science Foundation, software, HAZUS (Hazards-U.S.), to project the magnitude of potential Earthquake Engineering losses that might be experienced by the metropolitan New York City are-a Researcb Centers Program as a consequence of a damaging earthquake. After modification of the Federal Emergency Management Agency defuult HAZUS data sets for soil and building characteristics, a more cred Federal Emergency ible estimation of losses will emerge and be useful to metropolitan emer Management Agency gency personnel, as well as to other public and private stakeholders. It is Region II hoped that the results will contribute to improved disaster mitigation and New Jersey Office of emergency response plans throughout the area. Emergency Management New York State Emergency Management Organization he metropolitan NewYork-New Jersey region is vulnerable to several Tpotential natural disasters. Ice storms, snow, and other severe weather events such as hurricanes, associated storm surges, and flooding,foot-and mouth, and the West Nile virus are among those of current concern and prominent public attention. After adding such man-made disruptions as George Mylonakis, Walter terrorism and hazardous materials releases, and the hazards posed by de Fish and Paul Spiteri, caying infrastructure and aging buildings, earthquakes seem a dim pros The City CoUege ofthe City pect. But however unrecognized or unacknowledged the threat from Unil;ersity ofNew York earthquakes may be in New York City, seismic events have occurred and, Bruce Swiren, Federal until recently; no codes have existed to mandate earthquake strengthen Emergency Management Agency, Region II ing of structures. Statistics indicate that potential losses to a large urban KlausJacob, Noah area such as New York would be considerable. According to Scawthorn Ellelblum andJonatban and Harris (1989), the economic impact of a damaging earthquake (e.g., A.17wld, LollWntDoherty M6.0) in New York would be in the billions of dollars due to direct struc Earth Observatory tural and architectural damage, and does not reflect additional impacts on Andrea Dargush, Michael building contents, business continuity, fire suppression and human safety. K11kla and Montree An earthquake in the greater New York area would thus be considered a Polyium, MCEER Scott Stanford, New JersfJ)1 low-probability, yet high consequence event. Credible estimates of future Geological Sun;ey loss can be effective tools to encourage area stakeholders to mitigate against MichaelAllgnstyniak, New the possible future damaging consequences of earthquakes. Jersey State Police Office of Hazards U.S. (HAZUS) is a standardized, nationally-applicable loss esti Emergency Management mation tool, developed by the Federal Emergency Management Agency (FENIA.) in cooperation with the National Institute of Building Sciences. Research Team The HAWS software utilizes geographic information systems, such as continued on page 2 173 ArcView and MapInfo, to produce utilize the loss estimation program detailed maps and analytical re HAZUS to identify potential eco ports that describe a community's nomic loss to the greater New York Dan O'Brien, New York State potential losses. The current ver City area as a consequence of a Emergency Management sion applies a uniform engineering damaging earthquake. In the latter Office based loss estimation approach to stages ofYear 2,increased attention Dan Cadwell, New York State quantify damages, economic losses Geological Survey was given to inclusion of informa George Deodatis, Guy and casualties resulting from earth tion on critical facilities, to assess Nordenson, Michael quakes. Future adaptations are in post-event functionality. The con Tantala and Amanda tended to carry out similar analyses sortium consists of public agency Kumpf Princeton for hazards such as flooding and officials, business owners, emer University high wind. gency managers, engineers and ar Sam Wear, Laura McGinty Similar loss modeling studies chitects, utility owners and other and Andrea Alberl, have been conducted in the New area stakeholders. Researchers at Westchester COllnty Geographic Infomltltion York area in past years to examine Lamont-Doherty Earth Observa Systems, Department of the impact of flooding and hurri tory, Princeton University, and the Information Technology canes, and have been used to guide City College of New York work to emergency plans. Recognizing the gether to develop soils and build potential collective benefits of a re ing stock inventories which can be gional study, mitigation specialists used to refine the default data con at FEMA Region II, the New York tained within HAZUS. The data will State Emergency Management be used in future executions of Agency (NYSEMO), the New Jersey HAZUS and will assist researchers State Police Office of Emergency in refining their ultimate loss esti Management (NYOEM) and the mations for New York. New Jersey Geological Survey These studies are being comple joined together to develop a simi mented by information generated lar loss estimate for earthquakes. by similar studies for Westchester With NYSEMO and NJOEM, a County, New York and for a num FEMA-supported project was initi ber of counties in New Jersey. A ated in 1998 to apply HAZUS to the generalized HAZUS analysis will greater New York City area. A re provide a regional picture of poten gional consortium has been formed tial earthquake damage and loss. - the New York City-area Consor These assessments will ultimately tium for Earthquake-loss Mitigation be used to encourage and promote (NYCEM). MCEER was selected to earthquake mitigation action at the provide general coordination of the local, regional and statewide levels. activities and to conduct outreach activities to promote the outcomes of the project. The purpose of the Technical Summary consortium is to help develop the After preliminary explorations of necessary databases to effectively available data and its transferrability The users ofthis research include emergency responders, city planners, utility managers, building owners, design andconstruc;. tion engineers, critical facility managers, school and hospital administrators, the. public at large and other researchers. 174 to a HAZUS format, the NYCEM vidual Manhattan census tract level team prioritized its efforts, focus (Jacob, 1999).Pre-existing geologi ing on the development of a data cal studies were geographically lim base of geologic and building ited and not specific enough to • Metropolitan information for Manhattan below derive important information on Transportation Authority; 59th Street. As the project pro depth to bedrock. Other subsurface ltlew York TransitAttthority gressed, additional stores of data data collected in the form of bore • New furk City Department became available that allowed ex hole logs, were collected by private o/Design and Constmcrion pansion of the study to the entire developers or other entities, hold • New York City Department borough. In the current Year 3 of ing the data as proprietary. o/Finance and OPerations the project, results of the Manhat Using information provided by Research tan study will be merged with the New York City Department of • New York State Geological those of parallel studies being car Design and Construction, data from Survey ried out in N~w Jersey and 150 geotechnical borings for lower Westchester County, NewYork, for Manhattan were studied in the first NYCEffI Teclmical.A.tJvisory Committee extrapolation to a larger region cov year of the project. Using casing Dr. Carl]. Costantino, ering about 31 counties in New information and results from the Professor Emeritus, York and New Jersey. The respec standard penetration tests (SP1) Department ofCivil tive activities of the team members conducted, it was possible to de Engineering, City College and the parallel studies are de rive shear wave velocity profiles as o/Newfurk scribed in the following section. a function of depth at each loca Mr. Ronald 1'. Egucbi] CEO, Priorities, methodologies and tion, then translated into the appro ImngeCat progress of the NYCEM team have priate NEHRP site class. Some 200 Mr. Ramon Gilsanz, Gilsanz, Murray and benefited from periodic technical other older borings were available, Steficek advisement from a panel of experts but only offered information on Dr.Jeremy Isenberg, CEO] in earthquake engineering and loss depth to bedrock. Weidlinger Associates, Inc. modeling. To classify the overlying soils, a Dr. Stephanie A. King] shear wave velocityvs. depth func Stanford Unh'BrSity Mr. CbarlesA.. Kircher, Lamont Doherty Earth tion was derived from a subset of the initial 150 borings and applied Principal, Charles Kircher andAssociates Observatory (LDEO) to generate point values at the 200+ Dr.Joamw A-I. Nigg] Based on extensive prior re sites. The resultant classillcations Professor o/Sociology and search, the IDEO team, led by Klaus were primarily J\TEHRP soil types D Co-Director, Disaster Jacob, realized from the outset that and C. Surprisingly, no classA (very Research Center, University the default soils data within HAWS hard rock) sites emerged, although o/DelatlJare did not accurately reflect the sub numerous B (fum rock) sites were Dr. Thomas O'Rourke, identified. Several individual Professor, Schoolo/Civil surface conditions of the Manhat and Environmental borings indicated class E condi tan area. Near surface geologic Engineering, CorneD differences can introduce local tions. However, the conditions Unh'BrSity variations in shaking that can influ were not sufficiently continuous Mr. Maurice S. Power, Vice ence both the amplitude and spec throughout the census tract to President, Geomatrix tral composition of ground merit assignment of E to the tract. Consultants, Inc. motions, thus impacting structures This may be a pervasive problem Mr. Thomas Z. Scarange!lo, Vice President, Thomton and lifelines. associated with the census tract Tomasetti Engineers It would be a challenge to the approach to geological classifica team to upgrade the data to pro tion. These differences in soil type duce the needed NEHRP from the default assignments re geotechnical site classes at an indi- sulted in apparent reductions ofthe total loss estimations for buildings Estinzating Earthquake Losses for tbe Greater New }ork City Area 1_/," by factors of 0.7,0.72 and 0.8 for ing, coastal areas. An anticipated scenario events of Mw 5.0,6.0, and product of this effort will be the 7.0, respectively. Additional refine- development of a contoured soils • The NYCEM project is ment of the shear wave velocity map for Manhattan.This will allow primarily linked to MCEER functions made possible with more accurate extrapolation of cen outreach efforts to added calibration borings may be sus-tract specific geotechnical char increase awareness and reflected in future loss estimates. acteristics. acknowledgment of earthquake risk and the In Year 2, additional refinement An important Year 2 activity was value ofloss estimation of the lower Manhattan study area the coordination of approach to lnethodologies as was made possible by the inclusion classification of soils among the mitigation tools. of additional data provided by the New York City, New Jersey and Metropolitan Transportation Au Westchester County study areas. • Data collected from the thority/New York City Transit Au With input from members of the project may benefit thority. The study area was also NEHRP Seismic Provisions ongoing projects focusing on earthquake risk to expanded to include the entire bor Geotechnical Subcommittee and critical facilities in areas ough of Manhattan. In spite of the the Building Seismic Safety Coun ofIow- to-moderate hazard additional information, data density cil, a uniform methodology was de but with high collateral was greatest for midtown and rived to assure that assumptions damage. lower Manhattan and sparse for used to assign NEHRP soil class areas north of 137th St .. types are, to the extent possible in • Future phases ofthe Modifications were made to the agreement with the objectives of project witllikely benefit from MCEER lifeline methodology developed in Year 1 the NEHRP guidelines. There was research which has led to to reflect consideration of the improvements in HAZUS NEHRP 10-footrule, which assigns treatment ofburied site category E to any soil profile pipeline systems. that contains a continuous ten foot thick layer of very soft soils Oacob, 2000). Additional study also led to the increase in the cal culated shear wave velOCity for bedrock at the bedrock/soil inter face, better reflecting the prevail ing rock formations in the area. Site classes for all DDC borings in Year 1 were recalculated under these assumptions. A primary product in Year 2 is a census-tract based soil map for the entire borough, which can be used in validating results, gener Site Classes ated byHAZUS.Figure 1 describes A:!I:~~ the distribution of soil/rock types B:_ across the island, with higher el c:_ D: evations in uptown Manhattan on E:_ stiff rock/soil (Class B), interme diate type C in Midtown, and soft soil (Class D) in lower Manhattan, along the identified fault zones in • Figure I. NEHRP Site Classes for Manhattan Upper Manhattan and in low-Iy- by Census Tracts 176 general consensus that the informa istic. It is therefore even more im tion needed to accurately assign portant to accurately identify areas NEHRP classes is often unavailable of highest potential vulnerability to New York City-area and that in view of other inherent earthquake ground shaking so that Consortium for uncertainties within the HAZUS al mitigation, emergency response Earthquake Hazard gorithm, it should not unduly im and recovery approaches can be iJfitigation: pact results. strengthened. http://www.nycem.org The HAZUS algorithm is de In HAWS, the default building in Federal Emergency signed to execute loss estimations ventory is categorized by occupancy Management Agency using both deterministic and (commercial, residential, industrial, llAZf]S: probabiHstic earthquake inputs. governmental, educational, religious, http://J1l!wjema.gmVhazus Using historical local seismic agricultural), model building type events, the LDEO team also pro (e.g.,steel,reinfurced concrete,wood, Westchester County GIS: vided the epicentral locations for masonry), structural configurat!on http://giswullIJ.co.tL'estcr.e;ter. ny.us scenario earthquake events to be and height. HAWS also makes other used in the deterministic HAZUS assumptions about typical percent executions. age distribution of buildings within a census tract by age, quality (infe rior, built to code, superior) seismic Princeton University design level (low;moderate and high) New York City has one of the and height. The NYCfu\1 team at highest concentrations of high rise Princeton University (Guy J. P. buildings per square foot than any Nordenson, George Deodatis, other dty. Unreinforced masonry MichaelTantala andAmanda Kumpff) structures housing businesses, ar recognized the obvious disparities chitectural treasures, masonry fire between the typical suburban build stations, theaters and art galleries, ing inventory presented in HAWS apartment buildings, and more, and the actual structural characteris stand shoulder to shoulder with tics of New York. Such disparities these skyscrapers. would ultimately affect the accuracy Many New York structures hold of damage and loss scenarios to be historical status and many more generated by the program. were constructed without the guid One of the objectives of the stud ance ofa seismic building code. Be ies being carried out at Princeton fore 1996, earthquakes were not a is to develop a credible building design consideration although as inventory for the New York City construction progressed in the area, which will be more represen 20th century, wind load factors tative of its wide mix of structures. were being incorporated, thus ad Various damage and loss scenarios dressing (albeit coincidentally) would then be carried out, using some of the horizontal displace both the improved building data ments that might be experienced and the soils data provided by the during an earthquake. LDEOteam. Even today, seismic statutes ap In Year 1 of the study, the team ply to new construction only. Con used Sanborn fire insurance map sensus opinion is that retrofitting data (building height, size, location, thousands of New York buildings occupancy and type) and visual in to meet seismic standards is im spection to modify building inven practical and economically unreal- tories for two representative Estimating Earthquake Losses for the Greater New York City Area 1_/-/ 4000 ~3000 E 2000 4000 Washington Heights, ~ 1000 .8 3000 Inwood• E 2000 o Morningside Heights, ~1000 Wood Steel URM Concrete Hamilton• Heights, Building Type o Manhattanville Wood Steel URM Concrete 4000 ~ 3000 • ~2ooo 4000 ~ 1000 Central Harlem, ~30oo o Polo Grounds ~2000 Wood Steet URM Concrete Lincoln Square, ~1000 Building Type Upper West• Side, o Manhattan Valley East Harlem Flatiron, Wood Steel URM Concrete Midtown• Building Type 4000 ~3OOO •Upper East Side, •• ~2000 Yorkville, Chelsea, ::> 1000 Wood Steel URM Concrete Carnegie Hill, Clinton Z 0 Building Type Lenox Hill Gramercy, Murray• Hill, 4000 Wood Steel URM Concrete Turtle Bay, ~3000 ~2000 Damage State Tudor City ~1000 o 4000 Wood Steel URM Concrete ~ 3000 Building Type ~ 2000 ~1000 Wood Steel URM Concrete • Soho, Greenwich Village, o Wood Steel URM Concrete Chinatown, Little Italy, Building Type Building Type Noho • Financial District, Tribeca, Seaport, • East Village, Lower East Side, Battery Park City Tompkins Square • Figure 2. Building Count by Structural Type and Neighborhood census tracts - Wall Street (prima mates emerged, particularly when rily commercial) and Kips Bay (pri using smaller magnitude earth marily residential) (Nordenson et quakes. Total loss estimates in the al., 2000). Broader regional studies modified runs differed significantly examined New York below 59th from those of the default (in some Street,and the 31 county, NewYork cases more than a factor of ten). New Jersey area. HAZUS runs were Based on the runs, it appears that then carried out to assess the sen the most sensitive building informa sitivity of the methodology to tion for refined loss estimations changes in building inventory data. might be square footage and height, Dramatic differences in loss esti- and not structural type, as previ- 178 ously assumed.This will require fur dominant building type.This is not, ther validation. however, reflected in the amount With assistance from Dr. Jack of square footage occupied by Eichenbaum, New York City Asses unreinforced masonry, which is sor, the Princeton team acquired perhaps a more important Variable. valuable building data from the In the end, the Princeton team New York City Department of Fi was able to establish the building nance. The Mass Appraisal System inventory for the entire island of (MAS) data consists of over 2 mil Manhattan, with HAZUS-required lion property entries for the entire information at the individual build city, including such information as ing level. This is a unique accom parcel square footage, address, plishment for HAZUS applications. height, use, building footprint di With the assistance of Princeton mensions, construction year, qual student Evan Schwimmer, the da ity; and architectural style (Tantala tabase has been further enhanced etal., 2001). with the addition of critical facili The year 2 studies used this in ties data (hospitals, schools, and formation to further refine the police and fire stations). Several building inventory, although preliminary probabilistic and deter geocoding of entries was needed ministic HAWS runs have been to match information to the appro carried out using the critical facili priate census tract. In addition the ties data to generate projected richness of the MAS data required functionalities of schools,hospitais, simplification to fit into the HAWS and police and fire stations, to as categorization scheme. sess their relative capabilities to In contrast, the MAS construction provide shelter, provide emergency type data required addition ofmore care and control conflagration. specific information. Field surveys Casualties, population displace were carried out with the City Col ment and short- and long-term eco lege team and were used to validate nomic changes have also been some of the existing data as well projected. These are described in as some of the associated assump detail in the Year 2 report, which tions being made. Additional data may be found at the NYCEM sources from local engineers, Dun website. and Bradstreet business databases, These subsequent HAZUS (1999 aerial photography, existing struc and SR-16 versions) runs used the tural drawings and other sources improved building inventory and helped to further validate the new further verified the need for im database. proved data sources for buildings After thorough modification, the and soils to derive credible damage, resultant MAS database included loss and casualty estimates. records for more than 37,000 build ings, or over 2.2 billion square feet Westchester County, New York of area. Figure 2 shows the relative distribution of the number of build Under sponsorship of the New ings in each of four category types York State Emergency Management for twelve Manhattan neighbor Organization,a collaborative study hood districts, illustrating that was carried out in Westchester unreinforced masonry is the pre- County, NewYork.The study evalu- Estimating Eartbquake Losses for tbe Greater New Thrk City An,a 1-;9 ated the applicability of the HAZUS New Jersey Geological Survey to is important default datasets, and made appro define areas of seismic hazard. to accurately priate modifications to the extent Building inventory data was col identify areas possible. lected for selected areas to upgrade of high Considerable effort was dedi HAZUS inventory data. potentia' cated to data collection and to the Geologic data were acquired and vulnerability to updating of the HAZUS datasets, in analyzed in order to compile maps earthquakes so particular, additions of information of seismic soil class and liquefac that mitigation, on essential facilities (McGinty and tion susceptibility for Hudson emergency Wear, 2001). Important new cover County, New Jersey (Stanford, response and ages were included for such facili 1999). The soil class and liquefac recovery ties as day care centers and senior tion susceptibility data were en housing.Additional efforts were fo tered into the HAZUS model for approaches can cused on the inclusion of informa each census tract in the county. be strengthened. tion on essential facilities such as The HAZUS model was run with hospitals, schools, police and fire the upgraded geologic data and stations, and emergency medical with the default geologic data for services. earthquake magnitudes of 5, 5.5,6, This led to the identification of 6.5, and 7. The upgraded building shortfalls in county data availabil inventory data has not yet been in ity and development of a strategy cluded in the runs but is presently for future data creation. In the ab being processed by the Princeton sence of borehole data, scientists team. at the New York State Geological The upgraded geologic informa Survey assisted by conducting seis tion produced significant changes mic shear wave velocity tests on in both the spatial distribution of surficial soils within the county. damage and the total damage esti Tests were conducted in locations mates. This increased building dam of different NRC-defined soil types age was particularly notable in the so that a larger scale surficial soils Hudson waterfront and Hackensack map could be developed. Data was Meadowlands areas of the county, then extrapolated to a census tract where soils are softer and more liq level for necessary seismic soil clas uefiable than HAZUS default soil. sifications. Building damage was less on the The Westchester study culmi Palisades Ridge and on uplands in nated in a HAZUS execution based Kearny and Secaucus, where soils on three earthquake scenarios.The are stronger than the default. Be goal of the study is to provide loss cause most building construction estimation data to the sponsor to in the county is concentrated on enable development of a statewide these ridges, the total estimated mitigation approach for potential building damage is somewhat less future earthquakes. with the upgraded geologic data than with the default data at all magnitudes. Additional mapping New Jersey and scenario executions have also Under the oversight of Michael been carried out in Bergen county, Augustyniak at the New Jersey Of with Essex County underway. fice of Emergency Management, In addition to the HAZUS data up studies were carried out by the grades and scenarios executions, 180 shear-wave velocities were mea more than 600 buildings were sur sured on the three softest soil types veyed by the CUNY and Princeton at nine locations. These were con teams. ducted to validate the soil class as Survey data was input into anAC signments used in the scenario CESS database and compared executions, which use test-drilling against the default data. The differ data as a proxy for shear-wave ve ences between the default and ac locities. The measured velocities tual data were quite evident for the confirmed the assignments.An ad number of buildings and for the ditional 13 shear wave velocity percentage of residential buildings. measurements completed in There is also a significant difference Bergen County on four soil types between the actual data and default indicate once again that the SPT data for the number of commercial data accurately proxy for shear buildings. The HAZUS dataset de wave velocities, except in gravel picts a greater variety of building rich areas. These have fasterveloci types within a census tract than ties than SPT values suggest. Soil actually exists. These led to further classes were adjusted accordingly differences in loss estimations car in these areas. ried out for these areas. To further validate the data sen City College of New York sitivity, the CUNY team generated HAWS loss estimates for Kips Bay, To support the ongoing activities Rockefeller Center and Wall Street and to add some additional valida census tracts (Mylonakis,2000).The tion checks, MCEER provided sup Wall Street tract data were col port to George Mylonakis, CUNY, lected by the Princeton team. U s and his team of undergraduate en ing a fixed location (1884 gineering students to conduct field Rockaway Beach) scenario earth surveys of building inventory data. quake at varying magnitudes, the Their objectives were to assess the default and modified datasets were accuracy of both the default used to generate costs of structural HAWS building inventory data and damage. The 1999 version of the to verify and benchmark the MAS HAZUS software was used. The data for two of three representative Princeton group conducted similar census tracts. analyses usingHAZUS version '97. After initial training in HAWS, Good agreement existed between the team carried out walking sur the results generated by different veys of the Rockefeller Plaza cen versions of HAZUS. sus tract and a portion of the Kips Bay census tract. Buildings over a lO-block area were photographed, Outreach with data collected on frame type, Several years after the passage of exterior wall type, basement grade, seatbelt legislation, drivers and pas construction quality; exterior con sengers still disregard the need for dition, existing damage, HAZUS simple protective actions. Convinc building code, and occupancy. ing the public, the government and These categories correspond to the the private stakeholders in the required data fields within HAWS northeastern United States of the (Mylonakis, et al., 1999).A total of Estitnating Earthquake Losses for tbe Greater New York City Area 181 need to prepare for an earthquake agement, has helped to open dia e goal oftbe is a challenge not easily met. It has logues with City and metropolitan Westcbester been an important objective of this area emergency managers about County study project to convey that a low-prob the importance of earthquake haz was to provide ability event is a potential reality, ard mitigation. loss estimation· carrying with it consequences for In addition, conventional out data to enable which the metropolitan area may reach activities have been carried development of be ill prepared. out, targeted toward regional stake " state-wide The NYCEM project has taken ad holders who can both assist in re "pproacb to vantage of numerous media oppor fining the model through data potential future tunities to disseminate this proviSion, and can benefit from the information and to explain the im enhanced loss estimation model. e"rtbquakes." portance of accurate inventories Workshops and briefings have in and loss characterizations to the de creased their interest in the project velopment of mitigation strategies. and have led to sharing of informa The work has been featured in sev tion and data necessary to refine eral articles in the New York Times the default HAZUS data.Metropoli and other area newspapers, has tan transportation engineers are been the subject of a news focus interested in being part of possible piece on WNBC-News and has re future studies that would include cently been prominently broadcast transportation lifelines in the in the Discovery Channel program, model. "Earthquakes in New York?" The Project results are made widely awareness campaign was further available. A NYCEM website has augmented by a minor tremor that been established by MCEER, which struck midtown Manhattan, Janu serves as a repository for technical ary 17,2001 (see Figure 3).This at reports, data and maps generated tention, combined with the by the team. NYCEM results are re leadership of Bruce Swiren of ported regularly through the FEMA Region II and Dan O'Brien MCEER Bulletin and on the web, of NewYork State Emergency Man- and several additional information briefings will be featured through out the fmal year of the program. At the completion of the program, all data will be publicly accessible through a central repository. Conclusions and Future Research The study has revealed variances in HAWS default data that would, left uncorrected, compromise its ability to accurately project poten tiallosses due to a damaging earth quake. Loss quantifications generated by HAZUS have been • Figure 3. Epicenter of January 17, 200 1 Manhattan Earthquake found to be sensitive to variations (courtesy of Dan O'Brien) 182 in building inventory data, earth both the methodology, its inherent T.heNYCEM quake magnitude and local soil uncertainties, and its associated conditions. datasets have come under question, p1rojeah~s In some cases, the damage and loss HAWS is being utilized in other tliken values generated were notably large urban areas with higher lev adv.rJntn.ge of higher using default data than with els of seismic hazard as a tool to nume?ous the upgraded building and soils in assist area planners, responders, media formation. The data collection and and other stakeholders. opporl-unities to modification activities have substan The studies conducted have es dissemingte tially enhanced the program's ability tablished a foundation for more info-rmn.tiDn to become a credible forecast instru complete loss estimation studies Rnd explRin tbe ment of economic impact. for the greater NewYork metropoli impt:nnance of These activities have also signifi tan area.Additional data collection Rccurate cantly ~proved the general knowl and study of regional lifeline sys edge base about New York area tems, including water, gas and high inventories Rnd geology and have led to the devel way systems, will significantly loss ChR7N~cte7i opment of a more comprehensive enhance the risk characterizations %Rt-ions." building inventory that can be used that HAWS can provide for the by others. In faet,it is notably the first NewYork City area. database where the building inven Further involvement by emer tory has been described in a build gencyresponders,planners, builders, ing-by-building manner, which is not and health and hU1ill!l1 services offi only unique, but quite valuable to cials will help improve the effective future regional studies.The ability of ness of these studies. Accurate and the NYCEM team to build this infor complete fingerprints ofrisk, enabled mation base is notable,since available by more widespread participation by data for such a complex and large other stakeholders, will ideally lead urban area was difficult to identify to the development of an acceptable and obtain. Surmounting this initial damage and loss indicator that can obstacle has been an important be embraced by responders, build achievement of the project. Recon ers and the public. ciling this information to the census This goal has been encouraged as tract based approach of HAZUS has this studyhas progressed into its third also been an intensive and challeng year; as several presentations of re ingeffiJrt. sults to municipal agencies has led In an area of such vast and vital to greater coordination,in particular economic activity, it is paramount with the New York City Office of that any estimation of potential Emergency Management and with economic impact be based in well officials in Westchester County. founded data and analysis. While Acknowledgements The author would like to recognize and acknowledge the efforts of the following individuals: Bruce Swiren, FEMA Region II, and Dan O'Brien, New York State Emergency Management Office, fur their sponsorship and leadership; Elizabeth lemersal, FEMA, fur overall support of the program; and Dr. Jack Echenbaum, City Assessor, New York City Department of Fmance, and Dr.A. N. Shah, Metropolitan Transportation Authority/New York City Transit Agency, for their generous assistance. Estimating Eartbquake Losses for tbe Greater New lork City Area 183 Refurences------ Cadwell, D. and Nottis, G., (2000), "Seismic Hazard Assessment Data, Westchester County, New York," New York State Museum, Albany, New York. HAZUS 99: Estimated Annualized Earthquake Losses for the United States, Federal Emergency Management Agency Mitigation Directorate, FEMA 366, February 2001, 33 pages. Jacob, K. H., (1999), "Site Conditions Affecting Earthquake Loss Estimates for New York City," Lamont-Doherty Earth Observatory of Columbia University, Palisades, NY, New York City Consortium for Earthquake Loss Mitigation (NYCEM): Earth Science Tasks, First Year Technical Report May 6, 1999, 35 pages. Jacob, Klaus H., Edelblum, Noah, Arnold, Jonathan, (2000), "NEHRP Site Classes for Census Tracts in Manhattan, New York City," Lamont-Doherty Earth Observatory of Columbia University, Palisades, NY, New York City Consortium for Earthquake Loss Mitigation (NYCEM): Earth Science Tasks, Second Year Technical Report, December 31, 2000, 23 pages. Jarvis, Hugh, (1999), "Disaster Damage: Cost Issues for the New York City Area," Annotated Bibliography, MCEER Information Service. McGinty, Laura and Wear, Sam, (2001), "Initial Earthquake Loss Estimation Analysis for Westchester County, New York," Report to New York Emergency Management Organization, Westchester Counter Department of Information Technology, Geographic Information Systems, February, 2001, 53 pages. Mylonakis, George et al., (999), "Earthquake Loss Estimation for the New/York City Area;' NYCEM-MCEER Progress Report. Mylonakis, George, Fish, Walter, and Spiteri, Paul, (2000), "Development of a Building Inventory for Manhattan Region," NYCEM Preliminary Technical Report. Nordenson, G., Deodatis, G., Tantala, M. and Kumpf, Amanda, (2000), "Earthquake Loss Estimation for the New York City Area," NYCEM First Year Technical Report. Scawthorn, C. and Harris, S. K., (1989), "Estimation of Earthquake Losses for a Large Ea!>tern Urban Center: Scenario Events for New York City," Annals of the New York Academy of Sciences, Volume 558, June 16, 1989, pages 435-451. Stanford, S. D., Pristas, R. S., Hall, D. W, and Waldner, J. S., (2000), "Earthquake Loss Estimation Study for Bergen County, New Jersey: Geologic Component," prepared for the New Jersey State Police Office of Emergency Management by the New Jersey Geological Survey, 8 p. plus appendices and plates. Stanford, Scott, Pristas, Ronald S., Hall, David W. and Waldner, Jeffrey S., (1999), "Geologic Component of the Eatthquake Loss Estimation Snldy for Hudson County, New Jersey," Final Report. Stanford, S. D., Pristas, R. S., (1998), "Earthquake Loss Estimation Study for Newark, New Jersey: Geologic Component," prepared for the New Jersey State Police Office of Emergency Management by the New Jersey Geological Survey, 2 p. plus appendices and plates. Tantala, Michael W, Nordenson, Guy J.P., and Deodatis, George, (2001), "Earthquake Loss Estimation Study for the New York City Area," NYCEM Year 2 Technical Report, 1999- 2000. 184