Encyclopedia of Structural Health Monitoring Onli

Chapter 123 Modular Architecture of SHM System for Cable-supported Bridges

Kai-Yuen Wong1 and Yi-Qing Ni2 1 Highways Department, Government of Hong Kong, China 2 Department of Civil and Structural Engineering, Hong Kong Polytechnic University, Kowloon, Hong Kong, China

reasoning and experience so that the current and 1 Introduction 1 expected future performance of the bridge, for at least 2 System Architecture of WASHMS 2 the most critical limit events, can be predicted or evaluated. Bridge health monitoring system has been 3 Operation of WASHMS 15 adopted in the past decade to monitor and evaluate 4 Conclusions 16 the structural health conditions of cable-supported References 16 bridges in Hong Kong [1–5]. The bridge health monitoring system in Hong Kong is referred to as wind and structural health monitoring system (WASHMS ). Bridge health monitoring system is currently consid- ered as an integral part of bridge operation, bridge inspection, and bridge maintenance and has been 1 INTRODUCTION included as a standard mechatronic system in the design and construction of most large-scale and Bridge health monitoring is the tracing of the multidisciplinary bridge projects such as Stonecut- structural health conditions of the bridge in terms ters Bridge (SCB) in Hong Kong [6], and Sutong of the physical parameters categorized as environ- Bridge [7] and Donghai Bridge in mainland China mental loads and status, traffic loads, bridge features, [8]. The experience gained in the design, installa- and bridge responses by reliably measured data and tion, operation, maintenance, and development of the evaluation techniques, in conjunction with inductive WASHMS for Tsing Ma Bridge (TMB), Kap Shui Mun Bridge (KSMB), Ting Kau Bridge (TKB), and Encyclopedia of Structural Health Monitoring. Edited by the cable-stayed bridge in Hong Kong side of the Christian Boller, Fu-Kuo Chang and Yozo Fujino  2009 Hong Kong-Shenzhen Western Corridor (HSWC) has John Wiley & Sons, Ltd. ISBN: 978-0-470-05822-0. a significant influence on the design of new structural 2 Civil Engineering Applications health monitoring systems (SHMSs) in Hong Kong the modular architecture and input/output block and in mainland China. The WASHMS in Hong diagram of WASHMS. Of these six modules, Kong, which is based on modular design concept, modules 1–3, which are the key modular systems has been improved particularly in the aspect of data for the execution of real-time structural health interpretation, health evaluation, and data manage- monitoring, are composed of different types of SSs, ment and such improvements have been incorporated data/video logging systems, cabling network systems, in the design and installation of the WASHMS for servers and software facilities for data acquisition, SCB (SCB-WASHMS) [6]. transmission, and processing. The schematic layout of typical connections among modules 1–3 is shown in Figure 2. Modules 4 and 5, which are key modular 2 SYSTEM ARCHITECTURE systems for the execution of offtime structural health OF WASHMS evaluation, are composed of servers, workstations, and software facilities for data interpretation, health The modular concept of WASHMS [9–11], which evaluation, data management, and generation of is devised to monitor structural condition and reports. Module 6 is a set of portable computers (that evaluate structural degradation as it occurs rather store the system design information and operation than to detect structural failures, is composed of and maintenance manual of WASHMS) and tools for six integrated modules, namely, module 1—sensory carrying out inspection and minor maintenance of the system (SS); module 2—data acquisition and WASHMS itself. transmission system (DATS); module 3—data processing and control system (DPCS); module 4—structural health evaluation system (SHES); 2.1 Module 1—sensory system (SS) module 5—structural health data management system (SHDMS); and module 6—inspection and The SS that refers to the sensors and their corre- maintenance system (IMS). Figure 1 illustrates sponding interfacing units for input signals gathered

Wind and structural health monitoring system (WASHMS)

Module 2 Module 3 Data acquisition and Data processing and transmission system control system Module 5 Structural health data management system Module 6 Module 4 Inspection and Module 1 Structural health maintenance system Sensory system evaluation system

Analytical models: Bridge data and information: Finite element models Bridge design criteria Surrogate models Bridge monitoring criteria As-built bridge records Bridge inspection and Measurement quantities: Monitoring reports: Maintenance records Environmental status Instant display reports Traffic loads Detailed reports Bridge characteristics Special reports Bridge responses Bridge rating reports

System inspection Data and information transmission Updating of data and information Figure 1. Modular architecture and input/output block diagrams of WASHMS. Modular Architecture of SHM for Cable-supported Bridges 3

Sensors Sensors Sensors Sensors building Administration DAU or DAU or DAU or DAU or DVC DVC DVC DVC

DPCS

DAU or DAU or DAU or DAU or DVC DVC DVC DVC

Sensors Sensors Sensors Sensors

DAU Data-acquisition unit Bridge site DVC Digital video converter

DPCS Data processing and control system

Global cabling network system—ring-type single mode optic fiber Figure 2. Schematic layout of module 2—data acquisition and transmission system. from various monitoring equipment and sensors is correlated or compared with the design values; (ii) the categorized into four groups: (i) sensors for moni- SS is required to integrate the predictive modeling toring of environmental loads and/or status, which and data interrogation processes with the sensing include anemometers (three-dimensional ultrasonic system design process; (iii) the SS should have the type and two-dimensional propeller type), tempera- function to acquire data in a consistent and retriev- ture sensors (for the measurement of temperatures in able manner for long-term statistical data processing respective air, asphalt pavement sections, structural and analysis; (iv) all sensors should be chosen from concrete sections, structural steel sections, suspension the contemporary commercially available sensors that cables, and stay cables), corrosion cells, hygrome- best match the defined sensing performance require- ters, barometers, and rainfall gauges; (ii) sensors for ments; (v) the SS should include additional measure- monitoring of traffic loads, which include dynamic ments by removable/portable sensors to quantify weigh-in-motion stations, digital video cameras, and changing operational and environmental conditions; dynamic (weldable foil type) strain gauges; (iii) and (vi) at key locations, different types of sensors sensors for monitoring of bridge characteristics, should be deployed so that cross calibration of sensors which include fixed and removable/portable servo- could be carried out. type accelerometers, global positioning systems, level The major parameters monitored by each type sensing stations, and dynamic strain gauges; and of sensors are listed in Table 1. Figures 3–7 illus- (iv) sensors for monitoring of bridge responses, trate the layouts of the SSs in TMB, KSMB, which include dynamic strain gauges, static (vibrating TKB, HSWC, and SCB on their respective full wire type) strain gauges, displacement transducers, three-dimensional finite element models, which are global positioning systems, tiltmeters, fixed servo- built by MSC-PATRAN. The layouts of the SSs type accelerometers, buffer sensors, bearing sensors, as shown in Figures 3–7 are deployed or arranged and elastomagnetic sensors. in such a manner that the measured raw data can In the design/selection of the types and locations be used to derive the information or parameters of SSs, the following six basic criteria should be as listed in the fourth column of Table 1. These fulfilled: (i) the SS should have the ability to capture derived information and parameters are then used to the local- and system-level responses, which could be compare/correlate with (i) the corresponding bridge 4 Civil Engineering Applications

Table 1. List of required sensory systems and physical parameters for processing and derivation Monitoring Physical Required types of Physical parameters for category quantity sensory systems processing and derivation Environments Wind load • Ultrasonic-type anemometers • Wind speed and wind direction plots monitoring • Propeller-type anemometers (time-series data) • Barometers(a,b) • Wind speeds (mean and gust) and • Rainfall gauges(a,b) directions (histograms) • Hygrometers(a,b) • Terrain factors and wind speed proﬁle plots • Wind rose diagrams • Wind incidences at deck level • Wind turbulence intensities and intensity proﬁle plots • Wind turbulent time- and length-scale plots • Wind turbulent spectrum and cospectrum plots • Wind turbulent horizontal and vertical coherence plots • Wind response and wind load transfer function • Wind-induced accumulated fatigue damage • Histograms of air pressure, rainfall, and humidity

Temperature load • Platinum resistance • Effective temperatures in towers, deck, and monitoring temperature detector (RTD) cables type for temperature • Differential temperatures in deck and tower measurements in structural • Air temperatures and asphalt pavement steel, concrete, asphalt temperatures pavement, and air • Temperature response • Thermocouplers for cables • Temperature load transfer function

Seismic load • Fixed servo-type • Acceleration spectra near tower and monitoring accelerometers anchorage • Deck and tower response spectra • Seismic response and seismic load transfer function

Corrosion status • Corrosion sensors(a,b) • Potential risk of rebar corrosion in concrete monitoring(a,b) • Hygrometers(a,b) towers, concrete piers in side spans, and concrete deck in side spans

Traffic loads Highway traffic • Dynamic weigh-in-motion • GVW spectrum in each traffic lane load monitoring stations (bending-plate type) • AW spectrum in each traffic lane • Dynamic strain gauges • Equivalent number of SFV spectrum in • closed circuit television each traffic lane (CCTV) cameras(c–e) • Equivalent number of SFA spectrum in • Digital video cameras(a,b) each traffic lane • Highway-induced accumulated fatigue damage—SFV • Highway-induced accumulated fatigue damage—SFA • Overload vehicles detection • Traffic composition in each traffic lane • Traffic load response and traffic load transfer function Modular Architecture of SHM for Cable-supported Bridges 5

Table 1. (continued) Monitoring Physical Required types of Physical parameters for category quantity sensory systems processing and derivation Railway trafﬁc load • Dynamic strain gauges(c,d) • Bogie loads in each line of train monitoring(c,d) • CCTV video cameras(c,d) • Train loading spectrum • Equivalent standard load (train) spectrum • Train-induced accumulated fatigue damage • Train load response and train load transfer function

Bridge features Static influence • Level sensing stations(c,d) • Lane stress history in each traffic lane—each coefficient • global positioning system vehicular type monitoring (GPS)(b–e) • Stress range of each type of vehicle in each • Dynamic strain gauges traffic lane • Influence surfaces for combined deck plates and troughs • Line stress history of each type of train • Stress range of each type of train • Influence coefficients at tower tops and deck midspan

Global dynamic • Fixed and portable servo-type • Global bridge modal frequencies characteristics accelerometers • Global bridge vibration modes monitoring • Global bridge modal damping ratios (derived) • Global bridge modal mass participation factors (derived)

Bridge Cable forces • Portable servo-type • Cable frequencies and hence cable forces responses monitoring accelerometers • Cable damping ratios

Geometry • GPS(b–e) • Thermal movements of cables, deck, and monitoring • Level sensing stations(c,d) towers • Displacement transducers • Wind movements in cables, deck, and towers • Servo-type accelerometers • Seismic movements in deck and towers • Static strain gauges(c) • Highway load movement in deck and cables • Railway load movement in deck and cables(c,d) • Creep and shrinkage effects in concrete towers(a,b)

Stress monitoring • Dynamic strain gauges • Stress historical plots of instrumented • Static strain gauges(a,b) components • Elastomagnetic sensors(b) • Stress demand ratio plots of instrumented components • Principal stress plots of instrumented components • Force demand ratios plots of concrete–steel interfaces

Fatigue life • Dynamic strain gauges • Total accumulated fatigue damage and hence, monitoring remaining fatigue life due to combined load effects • Accumulated fatigue damage and hence remaining fatigue life estimation due to individual load effects (continued overleaf ) 6 Civil Engineering Applications

Table 1. (continued) Monitoring Physical Required types of Physical parameters for category quantity sensory systems processing and derivation Articulation monitoring • Dynamic strain gauges • Stress histories in bearings(c,e) • Displacement transducers • Stress demand ratios in bearing • Bearing sensors(b) • Motion status in movement joints • Buffer sensors(b) • Stress and motion status in buffers(b) GVW, gross vehicular weight; AW, axle weight; SFV, standard fatigue vehicle; SFA, standard fatigue axle. (a) HSWC. (b) SCB. (c) TMB. (d) KSMB. (e) TKB.

Tsing Yi Anchorage Man Wan Tower

Tsing Yi Tower

Anemometers (6) Accelerometers (19) Temperature sensors (115) Weldable strain gauges (110) Global positioning systems (14) Displacement transducers (2) Level sensing stations (10) Lantau Anchorage Data-acquisition outstations (3)

Figure 3. Layout of sensory system and data-acquisition system in Tsing Ma Bridge. design information and parameters for detection of works. The deployment of the measurement system any signiﬁcant deviation from design values; (ii) should also be able to quantify the changing opera- previous similar measured/derived values for detec- tional and environmental conditions and to provide tion of any abnormal or adverse structural perfor- information for developing future loading prediction mance; and (iii) analytical results from numerical or models. physical models for estimating the extent of damage, if any. Portable or removable servo-type accelerometers are also deployed with the main purposes of (i) 2.2 Module 2—data acquisition calibration of the full three-dimensional ﬁnite element and transmission system (DATS) model of the bridge and (ii) extraction of high-order frequencies and mode shapes from time history accel- The DATS is composed of four subsystems, namely, eration data for facilitating future damage detection data-acquisition system (DAS), local cabling network Modular Architecture of SHM for Cable-supported Bridges 7

Lantau Pylon

Ma Wan Pylon

Anemometers (2) Accelerometers (3) Temperature sensors (224) Weldable strain gauges (30) Global positioning systems (6) Displacement transducers (2) Level sensing stations (5) Data-acquisition outstations (2) Figure 4. Layout of sensory system and data-acquisition system in Kap Shui Mun Bridge.

Central Tower

Tsing Yi Tower

Anemometers (7) Accelerometers (45) Temperature sensors (83) Weldable strain gauges (88) Global positioning systems (7) Displacement transducers (2) Ting Kau Tower Data-acquisition outstations (3) Figure 5. Layout of sensory system and data-acquisition system in Ting Kau Bridge. system (LCNS), global cabling network system random and digital video signals. All DAUs and (GCNS), and commercial cabling network system DVCs are PC-based equipment. The ﬁxed DAUs and (CCNS). The DAS is composed of ﬁxed data- DVCs are permanently installed in the bridge deck acquisition units (DAUs), portable DAUs, and digital and bridge towers for collection and processing of the video converters (DVCs) for collection of respective signals received from SS (excluding corrosion cells). 8 Civil Engineering Applications

Hong Kong

Shenzhen

Anemometers (8) Accelerometers (44) Temperature sensors (118) Weldable strain gauges (212) Displacement transducers (4) Barometers (3), rainfall gauges (3) and Hygrometers (3) Corrosion cells (24) Digital video cameras (6) Data-acquisition outstations (3) Figure 6. Layout of sensory system and data-acquisition system in Hong Kong–Shenzhen Western Corridor (Hong Kong side cable-stayed bridge).

East Tower

WIM

West Tower

WIM Anemometers (24) Fixed and removable accelerometers (58) Temperature sensors (388) Dynamic strain gauges (678) Static strain gauges (158) Global positioning systems (20) Displacement transducers (34) Buffer sensor (18) Bearing sensor (12) Tensile magnetic gauge (32) Barometers, rainfall gauges and hygrometers (28) Corrosion cells (33) Digital video cameras (18) WIM Dynamic weigh-in-motion sensors (4) Figure 7. Layout of sensory system and data-acquisition system in Stonecutters Bridge. Modular Architecture of SHM for Cable-supported Bridges 9

Sensor-signals Typical data-acquisition unit Accelerometers Dynamic Sensor-based for air strain gauges Analog/digital converter conditioning control Static (simultaneously or strain gauges sequentially) Analog signals Analog signals

Signal conditioning Hard disk Signal conditioning Hard disk Propeller-type anemometers Fault report Ultrasonic-type PCI-controller anemometers card card

Global positioning Network

systems Digital input/output Digital input/output card

Inclinometers Interfacing card

Digital signals Digital signals Air temperature Interfacing media Interfacing media Displacement controller transducers

Light/current Copper Interfacing Copper Light/current Fiber-optic cable electronic Fiber-optic cable converter wire wire converter Fiber-optic network interfacing Figure 8. Schematic layout of typical connection among DAU, SS, and GCNS.

The portable DAUs are used to collect signals from DAUs to DPCS-1 and the digital video signal trans- portable servo-type accelerometers and corrosion mission cabling network for transmission of digital cells during ambient vibration measurements and/or video signals from individual DVCs to DPCS-2. Both specified field measurement works. backbone cabling networks are ring-shaped single- Figure 8 shows the schematic layout of a typical mode fiber-optic cabling networks with a data trans- connection among SS, DAU, and GCNS (fiber-optic mission capacity of 1 Gbps. cable). The major components in the DAU are the The CCNS is the leased high-speed line with a peripheral components interconnect (PCI)-controller, data transmission rate of not less than 40 Mbps for the signal conditioning device, and the analog-to- data communication (i) between the data-acquisition digital converter, and the proper selection/design outstations in bridge site of HSWC and the bridge of these components are the key steps to obtain monitoring room in West Control Building and (ii) between the bridge monitoring room in the Tsing measurement data with high quality. The LCNS is Yi Administration Building at North Tsing Yi and the composed of two local cabling networks, namely, the bridge monitoring room in West Control Building in copper cabling network for transmission of the signals South Tsing Yi (or Tsing Ma Control Area). from SS (excluding global positioning systems and digital video cameras, which are transmitted by fiber- optic cables) to DAUs for random signals and DVCs 2.3 Module 3—data processing for digital video signals, as shown in Figure 2. and control system (DPCS) The GCNS is composed of two backbone cabling networks, namely, the random signal transmission Figure 2 shows that the measured data collected cabling network for transmission of digitized signals from module 1 are preprocessed and transmitted by (excluding digital video cameras) from individual module 2 to module 3 or DPCS, which is composed 10 Civil Engineering Applications of two high-performance servers, namely, DPCS-1 software tools) equipped with appropriate software and DPCS-2 for data processing and control of tools to carry out the following operational functions: random (digital) signals and video (digital) signals, respectively. The DPCS is devised to carry out four • Finite element software interfacing capability operational functions, namely, system control, system That is, finite element models built by MSC/PATRAN operation display, bridge operation display, and post can be transferred to and executed in ANSYS soft- processing and analysis of data. Figures 9 and 10 ware without any manual modifications, and vice illustrate the respective functions of DPCS-1 and versa. DPCS-2 in block diagrams. • Integration of finite element models and measured data 2.4 Module 4—structural health The automatic input of measured data into relevant evaluation system (SHES) finite element models and under relevant (predefined) solvers for execution is therefore required. Figure 2 shows the comparison part on evaluation criteria or the offtime structural health evaluation that • Analytical and experimental modal analyses is devised to be taken up by the SHES or module 4, The automatic extractions and plots of the global which is composed of two high-performance servers dynamic features of global bridge structural system (one mainly for MSC and MATALB software tools and local bridge components from the measured and the other mainly for ANSYS and MATALB time history acceleration data are required. Figure 11

Data-acquisition System control DPCS-1 system for Sensor operation modes random signals Sampling rate adjustment Channel mapping Data transfer Data-acquisition units Temporary data storage

DAU1 to DAU8 System operation display Dynamic Sensor operation status weigh-in-motion stations DAUs operation status Network operation status DWIMS1 to DWIMS4 Data preprocessing activities Data temporary storage status Failure reports Databases in SHDMS Bridge operation display MPDDB Environmental loads and status Global cabling network system for electromagnetic signal transmission Traffic flows and loads status SADDB Kinematic quantities NADDB Monitoring criteria comparison CMDDB HEDDB Postprocessing and analysis of data Environment loads and status derivation DPCS-2 Traffic loads derivation Office ethernet Bridge features extraction Data interfacing with: Bridge responses derivation • Analyzed video data

Figure 9. Functional block diagram of DPCS-1—module 3 for random signals. Modular Architecture of SHM for Cable-supported Bridges 11

System control DPCS-2 Data-acquisition Control of cameras (P/T/Z) system for digital Channel mapping video signals Video transfer Temporary video storage

Digital video converters System operation display Cameras operation status DVC1 to DVC8 Video converter operation status Network operation status Video preprocessing activities Databases in Video temporary storage status SHDMS Failure reports

MPDDB Bridge operation display SADDB Environmental loads and status NADDB Traffic flows and load status Global cabling network system for electromagnetic signal transmission CMDDB Over-weighed vehicles Speeding vehicles HEDDB

DPCS-1 Postprocessing and analysis of data Data interfacing with: Image analysis of traffic characteristics DWIM stations Statistical analysis of traffic flows and loads Dynamic strain gauges Office ethernet Influence analyses of traffic load-effects GPS Derivations of potential traffic load-effects

Figure 10. Functional block diagram of DPCS-2—module 3 for video signals.

Measurement Modal analysis and model Finite element analysis data in SHDMS updating in SHES data in SHDMS

Physical Transformation/mapping Finite element bridge model model Solver Static parameters correlation Measured static data Update stiffness (K) Analyzed static data Solver

Undamped modal parameters Measured correlation Analyzed vibration vibration mode mode shapes Update stiffness (K), mass (M) shapes Synthesis Synthesis Damped modal parameters Frequency correlation response Frequency Update stiffness (K), mass (M), functions response damping ratio (C) functions

Measured External forces correlation Modal superposition external Derived forces Identify external forces (F ) external forces

Figure 11. Modal analysis and model updating in SHES by interfacing with SHDMS. 12 Civil Engineering Applications shows the typical modal analysis and model updating The execution of the above operational func- works in SHES by interfacing with SHDMS. tions requires the development of a customized finite element interfacing software system (FEISS) • Structural diagnosis modeling analyses to manipulate the operation of different software For any key structural component or location with tools (including both finite element analysis tools a measured and/or derived stress and/or displace- and random data processing and analysis tools) ment that exceeds 60–75% of its predefined moni- under different hardware and software operating plat- toring criteria, structural diagnosis or assessment forms. The FEISS is composed of five modules, of the current state of the bridge structure and i.e., modules A, B, C, D, and E, as shown in its history will be carried out in order to assess Figure 12. Module A includes the prebuilt finite any adverse effects on the global structural system element models of (i) full 3D global bridge model, based on the current and historical environmental, (ii) full 3D local foundation bridge model, (iii) 3D and operational loading conditions. For the sake global spine-beam or gird-beam bridge model, and of facilitating such assessments, it is required to (iv) full 3D local segmental bridge models. Module B perform automatic execution of finite element anal- includes the preconfigured finite element analysis yses on those prebuilt models and under predefined types such as normal mode analysis, linear static anal- typical types of structural analysis such as geometric ysis (influence coefficients’ determination), nonlinear nonlinear analysis of traffic and temperature loads, static analysis, random response analysis (buffeting random response analysis of dynamic wind seismic and seismic responses determination), fatigue anal- loads, etc. ysis, and impacting analysis. Module C includes the finite element and statistics solvers such as MSC- • Structural prognosis modeling analyses NASTRAN, ANSYS-Vertical Physics, MATLAB data The aim of structural prognosis is to assess the analysis suite, etc. Module D includes post-posting ability of the bridge structural system to carry and display software for generation of analyzed out future loading conditions (derived on the basis results and reports. Module E includes the inter- of past and current measured loads) or extreme facing and control software for the execution of events. The applications of structural prognosis in (i) automatic/manual retrieval of measured/analyzed WASHMS are (i) to predict/assess the remaining data from the relevant database in module 5 to module fatigue life of structural steel components based on 4 for processing/analysis; (ii) automatic/manual inp- a combination of finite element analyzed results, utting of the retrieved data into relevant prebuilt BS5400: Part 10: fatigue assessment rules, and finite element or statistical model/models for prede- measured strain results from WASHMS [12]; (ii) to fined types of finite element or statistics analysis; investigate the different types of potential failure and (iii) automatic/manual display and storage of the modes and the associated predictable and unpre- analyzed results. dictable loading conditions that cause damage and subsequent failure; (iii) to determine the potential consequences of each failure event or multiple events acting simultaneously; and (iv) to facilitate the 2.5 Module 5—structural health data planning of scheduled inspection and maintenance management system (SHDMS) activities. The SHDMS is composed of a high-performance • Visualization of analyzed results server equipped with data management software, and In order to increase the efficiency and accuracy in is the interfacing platform for the interoperability of identification and quantification of abnormal features data and information so that the efficiency of fusion or defects, all the analyzed, measured, and derived of data and information for decision making can be results are presented in comparative plots (with significantly enhanced. animation, where necessary) and tabulated in the The following five major databases are devised to matrix form. be executed in the SHDMS for interfacing works: Modular Architecture of SHM for Cable-supported Bridges 13

SHDMS Finite element interfacing software system or FEISS (under the control of a RDBMS in SHES (i.e., SHES-1 and SHES-2) and MSC simmanager)

Measured and processed JAVA environment data database (MPDDB)

Module A Pre-built finite Module C Statistical analyzed element models Automatic setting data database (DADDB) and execution of appropriate analysis Solver Module B Computer modeling Preconfigured data database (CMDDB) finite element analysis types

Module E Numerical analyzed Finite element data database (NADDB) interface and control

Module D Automatic processing, Health evaluation analysis and display of data database (HEDDB) analyzed results

Figure 12. Layout of FEISS in SHES and its interfacing with the databases in SHDMS.

• Measured and processed data database • Health evaluation data database (HEDDB) (MPDDB) All updated structural health monitoring and evalu- All time-series data obtained from measuring sensors. ation criteria and concise monitoring and evaluation results of environmental loads and status, traffic loads, • Statistical analyzed data database (SADDB) bridge features, and bridge responses. All data generated from signal/data processing and analysis software tools such as MATLAB—data anal- These five databases are manipulated and managed ysis suite, NI—data processing and reporting, SD by a data warehouse system, which is customized tools, etc. basing on “IBM DB2 UDB Warehouse Enterprise”, and is equipped with data management and data anal- • Numerical analyzed data database (NADDB) ysis tools for integrating enterprise-wide corporate data into a single repository from which users or All finite element analyzed/output data generated engineers can easily run queries, perform analysis, from finite element analysis software tools such as and produce reports. The data warehousing system MSC-NASTRAN, ANSYS-Vertical Physics, MIDAS, in SHDMS is devised to carry out the following LUSAS, etc. functions: • Computational modeling data database 1. Systematic cleansing, reconciliation, derivation, (CMDDB) matching, standardization, transformation, and All finite element modeling/input data generated from conformity of data and information from all data finite element analysis tools such as MSC-PATRAN, source systems such as DPCS servers and SHES ANSYS-Preprocessor, ANSYS-Workbench, etc. servers as shown in Figure 13. 14 Civil Engineering Applications

Data source End-user Data staging area systems Data and metadata presentation (operational data storage area tools store) DPCS-1 Extract (LFC) Enterprise Ad hoc query tools Data storage data warehouse DPCS-1 • Relational Report writers Extract (TKB) • Fast Load Measured and processed End users data database (MPDDB) applications Statistical DPCS-1 Processing Extract Statistical analyzed Feed and load modeling tools (SWC) • Clean data database (SADDB) • Reconcile Numerical modeling tools • Derive Feed and load DPCS-1 Numerical analyzed FEM Extract • Match (SCB) data database (NADDB) interfacing - Combine tools - Remove dups Computational models Feed and load DPCS-2 • Standardize data database (CMDDB) OLAP tools Extract (SWC) • Transform Data mining tools • Conform Health evaluation Feed and load data database (HEDDB) - Dimensions Visualization DPCS-2 Extract tools (SCB) • Export to data warehouse End-users

Load SHDMS Internal Cleaned dimension workstations SHES servers External data Summarized data SHES (relational database) workstations SHDMS server Summarized data Other servers and Others Extract (multidimensional database) workstations

Figure 13. Architectural layout of SHDMS and its interfaces.

2. Manipulation of all types of correlation analyses (d) planning of scheduled bridge inspection and and features extraction plots, by online analytical maintenance activities with bridge mainte- processing tools and appropriate data mining nance team; and tools, based on all data and information generated (e) updating/calibrating the bridge rating sys- from the software tools in DPCS-servers and tem and computational (numerical and stat- End-users’ servers and workstations as shown in istical) models for processing and analysis Figure 13. of data and information, where necessary. 3. Creation of data marts (summarized data in Figure 13), based on the results of afore- 4. Forming the center of data interrogation and met- mentioned correlation analyses and plots, for amodeling for bridge health diagnosis and prog- the execution of the following monitoring and nosis through the integration of data and informa- evaluation works: tion from both measurement and computational systems. (Metamodels refer to the functional (a) reporting the current and future loading forms of statistical-based models, finite element conditions (such as wind, temperature, models, neural networks, etc.) seismic, and traffic) acting on the bridge; (b) reporting the current and future corrosion status on specified bridge components; 2.6 Module 6—inspection and (c) reporting the current and future structural maintenance system (IMS) health conditions of the bridge in terms of the physical parameters as listed in the The IMS is composed of two notebook computers fourth column of Table 1; (IMS-1 and IMS-2) and a tool-box (IMS-3). Its Modular Architecture of SHM for Cable-supported Bridges 15 function is devised to carry out inspection and main- refers to the comparison of the measured results tenance works on SSs, DAUs, display facilities, with the predetermined monitoring criteria (i.e., at LCNSs, and GCNSs. All information (drawings and about 60–75% of the design values at serviceability records) regarding system design, system installation, limit state), whereas the latter refers to the compar- system operation, and system maintenance is stored ison of the derived results with the evaluated criteria and operated in IMS-1 and IMS-2. The IMS-3 is a (i.e., at 100% of the design values at serviceability tool-box for carrying out inspection and minor reme- limit state). The criteria for monitoring and evalu- dial works. ation are defined and calibrated in accordance with the updated requirements of the damage types as defined in the criticality and vulnerability ratings, i.e., 3 OPERATION OF WASHMS structural damage (due to structural actions), environmental damage (corrosion), accidental damage, The system operation block diagram of WASHMS is and wearing damage [13]. In Figure 14, it is shown shown in Figure 14, where the monitoring of kine- that if the measured results exceed the monitoring matic quantities refers to the monitoring of bridge criteria, structural diagnosis and prognosis works will features and bridge responses. The figure shows be carried out. Both the structural diagnosis models two levels of monitoring, namely, the sensor-based and structural prognosis models are finite-element- comparison of the measured results and the moni- based and/or empirical/statistical-based models, of toring criteria, and the model-based comparison of the which the former is used to assess the current and/or derived results and the evaluation criteria. The former historical state of the global bridge structural system

System information: Real-time structural Structural characterization health monitoring Bridge performance records Bridge maintenance records

Kinematic quantities Traffic loads Environmental loads monitoring monitoring and status monitoring Offtime structural health evaluation

Measured Initiating/checking Updating, if any No exceedance results—monitoring criteria analytical and/or comparison empirical models Exceedance

Updated/calibrated Structural diagnosis Future traffic Future environmental analytical and/or and prognosis models loads models loads and status models empirical models

Updating, if any Derived No exceedance Estimations of: results—evaluation criteria Remaining service life comparison Time to failure Take action, update system and continue the process Exceedance Time to maintenance Decision for actions: Carry out detailed inspection of suspected components Bridge inspection and Investigate causes of defects and estimate time to maintenance maintenance team Determine scope of maintenance works, if required

Figure 14. System operation block diagram of WASHMS. 16 Civil Engineering Applications and/or local structural components whereas the latter planning of bridge inspection activities, and be able is used to predict the consequences under the assumed to determine not only the cause of the damage but future operational and environmental loads and status. also the extent of remedial work, once the damage is identiﬁed.

4 CONCLUSIONS REFERENCES This article has described a modular architecture that [1] Highways Department, Wind and Structural Health has been used in the design of SHMS for cable- Monitoring System, Particular Specification for the supported bridges in Hong Kong. The architecture Electrical and Mechanical Services in Lantau Fixed of the SHMS described consists of six integrated Crossing, Highway Contract No. HY/93/09, 1993. modules, namely, SS, DATS, DPCS, SHES, SHDMS, [2] Highways Department, Wind and Structural Health and IMS. Each module is well defined and encapsu- Monitoring System, Particular Specification for the lated. In order to ensure the reliability of measured Construction of Ting Kau Bridge and Approach results, different types of sensors should be deployed Viaduct, Highway Contract No. HY/93/38, 1993. at the same key locations and/or components where [3] Highways Department, Wind and Structural Health large displacements and stresses are expected to occur Monitoring System for Lantau Fixed Crossing and so that the measured results can be validated through Ting Kau Bridge, Consultancy Agreement No. CE correlation among themselves. 75/94, 1997. Since the hardware configuration of the DAS [4] Wong KY, The wind and structural health monitoring has a significant influence on data quality, the system (WASHMS) for cable-supported bridges performance requirements on the linearity, temper- in Tsing Ma Control Area. an invited paper, ature drift, accuracy, direct current resolution, band- Proceedings of the IFAC Conference on New width, etc., of the signal conditioning and the data- Technology for Computer Control. Hong Kong, 2001. acquisition device should be identified and quantified. [5] Highways Department, Wind and Structural Health Appropriate customized software systems should also Monitoring System, Appendix W of the Partic- be developed and configured to process or derive the ular Specification for the Construction of Hong measured data in the data formats applicable to bridge Kong—Shenzhen Western Corridor, Highway Contract No. HY/2002/21, 2002. health monitoring and evaluation. Structural health monitoring and evaluation works [6] Highways Department, Wind and Structural Health should be executed through the correlation analyses Monitoring System, Section 33 of the Particular Specification for the Construction of Stonecutters and features extractions of (i) measured and analyzed Bridge, Highway Contract No. HY/2002/26, 2002. results, (ii) current and previous measured results, (iii) previous and updated analyzed results, (iv) and [7] Dong X, Zhang Y, Xu H, Ni YQ. Research and design of structural health monitoring system derived results (from analysis and/or measurements) for the Sutong Bridge. In Structural Health and assigned bridge performance criteria. As the Monitoring 2005: Advancements and Challenges correlation analyses and features extractions involve for Implementation, Chang F-K (ed). DEStech the synchronized processing of two or more data Publications: Lancaster, PA, 2005, pp. 1736–1742. files/sets, the use of data warehouse system equipped [8] Sun L, Dan D, Sun Z, Health monitoring system with online analytical processing tools and appro- for Donghai Bridge in Shanghai, Proceedings of priate data mining tools will facilitate the automatic the Asia-Pacific Workshop on Structural Health execution of such synchronized data processing and Monitoring. Yokohama, Japan, 2006. analysis works. [9] Wong KY. Instrumentation and health monitoring It is concluded that the WASHMS for cable- of cable-supported bridges. Journal of Structural supported bridges should at least be able to monitor Control and Health Monitoring 2004 11:91–124. the loading and structural parameters set by the bridge [10] Wong KY, Recent development of structural health designer so that the bridge performance under current monitoring system. a keynote paper, Proceedings and future loading conditions can be evaluated, and of the International Workshop on Integrated Life- such evaluated results should be able to facilitate the Cycle Management of Infrastructure, The Hong Kong Modular Architecture of SHM for Cable-supported Bridges 17

University of Science and Technology: Hong Kong, 28–30 March. Chongqing, 2007, organized by September 2004. Merisis. [11] Wong KY. Design of a structural health monitoring [13] Wong KY, Criticality and vulnerability analysis of system for long-span bridges. Structure and Tsing Ma Bridge, Proceedings of the International Infrastructure Engineering 2007 3:169–185. Bridge Conference on Bridge Engineering.The [12] Wong KY, Stress and trafﬁc loads monitoring of Hong Kong Institution of Engineers: Hong Kong, Tsing Ma Bridge. China Bridge Congress 2007, November 2006.