Structural Health Monitoring and Diagnostic Technology

Transcription

1 1/5/211 Structural Health Monitoring and Diagnostic Technology Maria Q. Feng Chancellor s Professor Department of Civil & Environmental Engineering Director, Center for Advanced Monitoring and Damage Inspection University of California, Irvine Courtesy of Business Week Vulnerable Highway Bridges More than a quarter of the bridges nationwide are structurally deficient. How Can Structural Health Monitoring Help? Current Practice - Visual, Periodic Inspection Cannot timely detect problems Difficult to detect invisible damage Future Practice - Incorporating Structural Health Monitoring Continuous monitoring to identify hot spots in real time Condition-based inspection focusing on hot spots 3 4 Courtesy of LA Times, New York Times,, FHWA.ASCE Integration of Real-Time Global Monitoring with Targeted Local Inspection Where Are We Now? Continuous Monitoring for Real-Time Assessment of Global Structural Integrity and Identification of Hot Spots Condition-Based NDE Inspection of Targeted Hot Spots 1

2 1/5/211 We Need Better Sensors We Need Tools to Interpret and Use the Data Piano wire Solar Penal Accelerometer Bello ws Displacement Meter Strain Gauge Wall L.V.D. T + DC- - 5V To recorder Signal + Conditi oner - TSA- 2 Soil Pressure Sensor Bridge Site Wireless Real-Time Data Acquisition System UCI Campus Contents Recent Advances in Sensors Moiré fringe-based fiber optic accelerometer Wireless sensor network Vision-based remote displacement sensor Distributed fiber optic strain sensor Piezoelectric Impedance Sensor Microwave imaging system Integrated scour monitoring system Diagnostics and Prognostics Tools Case Study #1: Post-event damage assessment Case Study #2: Long-term structural health monitoring Case Study #3: Integrated NDE of concrete bridge decks 9 Moiré Fringe-Based Fiber Optic Accelerometer Uniquely Suited for Monitoring Large-Size Civil Infrastructure System in Harsh Environment Sensing Principle: moiré fringe Unique Features: High resolution and large measurementrange, uniquely suited for accurate measurement of both ambient microvibration and strong motion. Immunity to electromagnetic interference and lightning strikes, safe to use in explosion-prone environments. Excellent low-frequency performance suited for monitoring long-period structures. 1 fixed grating Intensity Continuous Monitoring of CalIT2 Building x(t) Intensity D/4 D y moving grating Changing the direction 2x( t) s1( t) sin d 2x( t) s2( t) cos d time Conventional Accelerometer Sokushinco. Y- direction Acceleration x(t) : displacement Z- direction Optical Fiber fiber Optic accelerometer Accelerometer 2

3 1/5/211 Seismic Shaking Table Test FOA Ch5 Beam Measurement of Bridge Vibration Ch4 Ch3 Columns Ch2 Conventional Accelerometer Ch1 Acc. FOA Base (g) (on Hwamyung Bridge in Busan, Korea) Wireless Sensor Network US : M. Shinozuka (UCI) Korea : J. T. Kim, Hwamyung S. W. Shin (PKNU); Cable stayed C. B. Yun Bridge: (KAIST) under Construction Wirelss TRX Sensing Unit Base Station Gimhae BLC BRC5 BRC12 BRC17 Busan Cable: 4 gophers Deck: 7 gophers Pylon: 2 gophers Total: 13 gophers Hpwamyung Bridge, a prestressed concrete girder cable bridge Main Span: 5m, Side Span: 115m x 2, including ramps: 1.39km To be completed in March, 213 Sensing Unit (Gopher) Wireless Communication Unit (Roocas) Daisy chain connection up to 12 nodes Low noise MEMS accelerometers Support to connect commercial accelerometers Built in tilt/humidity/temperature sensors Multi Wireless Technologies: WiFi/Xstream/XBee/Eco Support various network topologies Support real time monitoring system Establishment Server and Web based SHM method Sensor Deployment Real Time Remote Monitoring: Temporary Connection Gopher Roocas Anemometer Data Logging at UCI Weblink: Measured data are saved/sent to UCI server Web based Monitoring Web login using PC and Mobile Gopher Remote Monitoring Framework Gopher Roocas Roocas Cable node *HSDPA: High Speed Downlink Packet Assess (3.5G) 3

4 Time (s) Frequency, Hz Frequency, Hz Frequency, Hz Frequency, Hz 1/5/211 1 st Vertical Direction 1 st Lateral Direction Vision-Based Remote Displacement Sensor.279Hz.4134Hz Acceleration (g) FE Model and Modal Extraction Busan BLC2 Gimhae 115 m 27 m Cable Monitoring (BLC2) 115 m Remote non-contact measurement by using low-cost video camera 2 Measurement wt and w/o Target Panel Robust Image Processing tan c 1 I y I x N i f I x I otherwise y Dark Light Lighting Change Dark Light Robust Against Obstacles and Lighting Change OC Matching Measured Displacement - Excellent performance w/o target panel in the dark Time: 11: AM, 12/15/29 Seismic Shaking Table Test Displacement, mm Displacement, mm Great Hanshin-Awaji 5 Western Tottori Vision-based system Vision-based earthquake system earthquake 25 Displacement, mm Time, sec Time, sec 5 LVDT LVDT Displacement, mm Time, sec Time, sec Time: 5:15 PM, 12/15/29 PSD, mm2/hz Vision-based system PSD, mm2/hz Vision-based system PSD, mm2/hz LVDT PSD, mm2/hz LVDT Displacements by Vision-Based Sensor and LVDT 4

5 1/5/211 Remote Monitoring of Bridge Deformation Distributed Fiber Optic Strain Sensor Principle: Brillouin scattering Innovation: Stimulated Brillouin scattering - Increasing spatial resolution by 1 times Applications: Crack detection Long-distance distributed strain measurement Modulated Continuous (pump) Stimulated Brillouin Scattering Modulated Continuous (probe) 25 [Hz] - B 26 Detection of Cracks in Fiber Reinforced Concrete Optical fiber Strain [1-6 ] 12 Measurement 1 5 kn Analysis Position [m] P/2 P/2 Long-Term Monitoring of Bridge Optical fiber Strain [1-6 ] 12 6 kn Measurement Analysis Position [m] Strain [1-6 ] 12 7 kn 1 7 kn Measurement Analysis Position [m] Level Survey -2 Measurement by OFS -4 Level survey -6 OFS Position [m] 28 Deflections by optical fiber sensor vs. level survey Measurement Region Deflection [mm] Piezoelectric Impedance Sensor (C.B. Yun, KAIST, Korea) Reference-Free Damage Detection Using Dual PZT s (H. Sohn, KAIST, Korea) PZT (5x5x.5) Loosened 1 Bolt all dimensions in cm Real Impedance (Ohm) Samseung Bridge intact loose 1/2 rotation loose 1 rotation retighten All dimensions in cm Freqeuncy (Hz) x 1 4 5

6 1/5/211 Welded zones between stiffener and web/flange Signal Ab B B Signal Ab S A /S S /A A Microwave Imaging Microwave Image of E f (x f,y f,z f ) at r f (x f,y f,z f ) E f ( rf ) e E E L E f (rf ) I I L s, T1R1 s, T1R1 Es, T1R1 Es, T1R1 L ITt M M O T1 T 2 E E L s, T1R1 s, T1R1 N n Nm jke rrmn r Nl Nk jke rtkl r I Rmn( rf ) Iobj ITkl ( rf ) V n1 m1 4 rrmn r l1 k 1 4 rtkl r e Es, T1R1 I R1 E s, T1R1 I R2 M M Es, T1R1 I Rr Use Antenna Array Focusing dv Fast Measurement PZT A damage PZT b B Signal Ba B Signal Ba S S /A A /S A PZT a PZT B Difference (Signal Ab- Ba) 32 Detection of Concrete Voids Experimental vs. Numerical Results.1.1 y.8 m.8.8 Network Analyzer Concrete Block m.2 m Reconstructed Area.2 m.2 m Antenna Array x y (m) y (m) m Switch Box Antenna Array x (m) Experimental x (m) Numerical Multi-Frequency Technique Handheld Real-Time Microwave Imaging Device Focusing 2 2 Y (mm) -2 Y (mm) -2 Concrete Column FRP Jacket Steel Rebar X (mm) Using Single-Frequency (5.2GHz) X (mm) Using Multi-Frequency (4.6, 4.8, 5., 5.2GHz) Locate, in real time, invisible defects Portable & lightweight Easy operation requiring no training 36 6

7 1/5/211 Inspection of FRP-Wrapped Concrete Non-Destructive Detection of Corrosion Uncorroded rebar Corroded rebar Integrated Scour Monitoring System (KC Chang, NCEER, Taiwan) Field Scour Monitoring System IP Camera Server Wireless Sensing Network (WSN) Nodes (for pier to pier transmission) Rechargeable battery 39 Scour monitoring pipe 4 Field Bridge Scour Monitoring The measured data were transmitted through Wi-Fi communication system which is end connected with wired network, then the data was sent to the remote server database in the office. Freeway No.3 Bridge WiFi HouLi Toll Station Freeway No.1 Bridge Laboratory Experiments Freeway No.1 Bridge Freeway No.3 Bridge WiFi Experiment flume HouLi Toll Station 41 Taichung, Taiwan Verification on concepts of the field instrumentations Simulation on varies sour conditions to verify the monitoring system Experimental study on bridge collapse due to scouring and the associated numerical simulations. 42 7

8 1/5/211 Experiment Measurements Field Bridge Scour Monitoring Website Accelerometer Information from Central Weather Bureau Real-time measured information Predictive information Video Information Vibration sensors Tilt meter Location of instrumented bridges Numerical Simulation Information Update messages 43 Graphic information (water level, flow speed and scour depth) 44 Contents Recent Advances in Sensors Moiré fringe-based fiber optic accelerometer Wireless sensor network Vision-based remote displacement sensor Distributed fiber optic strain sensor Piezoelectric Impedance Sensor Microwave imaging system Integrated scour monitoring system Diagnostics and Prognostics Tools Case Study #1: Post-event damage assessment Case Study #2: Long-term structural health monitoring Case Study #3: Integrated NDE of concrete bridge decks 45 Use of Sensor Data for Decision Support Sensor Data from Instrumented Structure Post-Event Damage Assessment Residual Capacity Estimation Disaster Mitigation Speedy Recovery Long-Term Structural Health Monitoring Intelligent Maintenance 46 Case Study #1 Post-Event Damage Assessment Goals: To Detect, Locate, and Quantify Structural Damage To Predict Remaining Capacity of Bridges To Assist Decisions for Post-Event Disaster Mitigation and Speedy Recovering Damage Index: Structural Stiffness and Damping Identified from Structural Vibration (Ambient, Seismic) Measurement Approaches: A Variety of System Identification Techniques for Damage Assessment Link Damage to Remaining Capacity Realistic Experimental Validation 47 Damage Assessment Methods and Experimental Validation Seismic shaking table tests of a 3-bent concrete bridge Progressive damage caused by increasing ground motion intensity Damage assessment by different system identification techniques: 1. Bayesian Updating Extended Kalman Filter Central Difference Filter 2. Optimization-Based Approaches Quasi-Newton Evolutionary Algorithm 3. Nonlinear Damping 4. Neural Networks 48 8

9 1/5/211 Bayesian Upadating X k f ( X k1,u k1,w k1 ;) Z k h( X k,u k,v k ;) f: state transfer function h: observation function X is the states, Z the observations, U the deterministic input, W the process noise, and V the measurement noise T i K A i 1 2 L l By Bayesian Theorem, the recursive Bayesian filtering at time step k p(z p( X k Z 1:k ) k X k ) p( X k Z 1:k1 ) p(z k x k ) p(x k Z 1:k1 ) dx k a posteriori knowledge p( X k Z 1:k ) Deduction by h K i D Deduction by f a priori knowledge p(z k X k ) p( X k x k1 ) p(x k1 Z 1:k1 ) dx k1 p(z k x k ) p(x k x k1 ) p(x k1 Z 1:k1 ) dx k1 dx k 49 Extended State Parameter Estimation X k, k T X k f ( X k1,u k1,w k1 ;) Z k h( X k,u k,v k ;) f and h are augmented by parameter θ, element stiffness (Structural Condition) For SHM purpose, we concerns about the changes of θ and its distribution X F: state transfer function k, k T F( X k1, k1 T,U k1,w k1 ) Z H: observation function k H( X k, k T,U k,v k ) X f( X, U, W ; ) k k1 k1 k1 F: k f ( k 1, Uk 1, W k k1 W k1 ; I k 1) Propagate a distribution H: Z h( Xk, Uk, Vk; ) through a nonlinear system k 5 Extended Kalman Filter & Central Difference Filter Seismic Shaking Table Test /o 33/o Extended Kalman Filter projection X k1, k1 T /o Z k X k, k T Central Difference Filter projection Parameter Evolution (by Central Difference Filter) Identified Stiffness Reduction (by Central Difference Filter) WN-1 53 Identification results based on vibration data are consistent with observed damage sequence based on embedded strain sensors: 54 Bent 1 yields Bent 3 yields Bent 2 yields Bent 3 buckles 9

10 1/5/211 Identified Instantaneous Stiffness Reduction (by Extended Kalman Filter) T-13 Hysteretic Loops in Low-Amplitude Vibration Secant Stiffness T-14 Largest Excursion Point T-15 T-19 As PGA increases, the bridge column stiffness decreases. 55 Pre- and Post-Event Equivalent Linear-Time-Invariant System 56 Identification of Post-Event Stiffness Nonlinear Damping Analysis As PGA increases, viscous damping decreases and friction damping increases. As PGA increases, the bridge column stiffness decreases Results by Different Identification Methods Use of Identified Stiffness for Capacity Estimation 59 Idealized elasto-plastic pushover curve of Bent-1 6 1

11 1/5/211 Use of Identified Stiffness for Capacity Estimation Use of Sensor Data for Decision Support Sensor Data from Instrumented Structure Post-Event Damage Assessment Residual Capacity Estimation Long-Term Structural Health Monitoring DECISION: Open Partially Open Closed Disaster Mitigation Speedy Recovery Intelligent Maintenance Case Study #2: Long-Term Structural Health Monitoring Issues with Field Implementation Soil-Structure Interaction Amplitude Dependency Modeling of Traffic Excitation Long-Term Monitoring Data Vehicle- Structure Interaction Long-Term Monitoring Data Bridge Doctor for Decision Support Scaled Power Spectral Density 43 accelerometers Acceleration (cm/sec 2 ) Acceleration (cm/sec 2 ) Acceleration (cm/sec 2 ) Time (sec) Time (sec) Time (sec) 65 Earthquake Records, Mw=4.9 Yucaipa Earthquake, June Recorded ground motion at Calit2 building site\ Earthquake Date Magnitude Distance(km) PGA(g) Yucaipa Jun 16, S. Clemente Oct 16, Chino Hills Jul 29, Inglewood May 17, Identified modal frequencies and damping ratios Yucaipa San Clemente Chino Hills Inglewood Mode f (Hz) ξ (%) f (Hz) ξ (%) f (Hz) ξ (%) f (Hz) ξ (%)

12 Q 1 Q i 1/5/ f Frequency (Hz) Damping ratio (%) Identification of Stiffness by Neural Networks Experimental Modal Analysis Frequencies & Mode Shapes f Bridge Traffic Vibration Data f PGA (cm/sec 2 ) The variations of modal parameters with the peak ground accelerations PGA (cm/sec 2 ) 67 Assessment of structural Health Change of Stiffness from Baseline Neural Network Input layer Hidden layer Hidden layer Output layer 68 Traffic Excitation Modeling by Integration of Vibration and Traffic Monitoring Bayesian Updating cov[ F ( t) F ( s)] Excitation Covariance Five-Year Continuous Monitoring of Jamboree Bridge Stiffness Degradation Temperature Effects N Traffic Video S x i-x 1 (m) s-t (sec) 69 Image processing - Vehicle type, arrival time & speed e 97 r summer u t winter c u 96 r t s r 95 e p u S Time (year) 7 Superstructure Stiffness (%) Eight-Year Continuous Monitoring of West St. On-Ramp 4. Vibration Tests at West St. Bridge 3.5 Frequency [Hz] f 3=-.262(# of years) f 2=-.174(# of years) f 1=-.22(# of years) Jan 22 Jan 23 Jan 24 Jan 25 Jan 26 Jan 27 Jan 28 Jan 29 Winter 2 Spring 2 Summer 2 Fall 2 Winter 3 Spring 3 Summer 3 Fall 3 Winter 4 Spring 4 Summer 4 Fall 4 Winter 5 Spring 5 Summer 5 Fall 5 Timeline Winter 6 Spring 6 Summer 6 Fall 6 Winter 7 Spring 7 Summer 7 Fall 7 Winter 8 Spring 8 Summer 8 Fall 8 Winter 9 Spring 9 1st Mode Identified w/o Vehicle Interference 2nd Mode Identified w/o Vehicle Interference 3rd Mode Identified w/o Vehicle Interference 1st Mode Identified with Vehicle Interference 2nd Mode Identified with Vehicle Interference 3rd Mode Identified with Vehicle Interference 71 * Vehicle-bridge interaction causes notable variation of natural frequencies 72 12

13 1/5/211 Acceleration [g] TEST: 15mph, L1, BB-EB 5 x 1-3 CH1 5 x 1-3 CH3 5 x 1-3 CH5 5 x CH Time [sec] 5 x CH x CH x 1-3 CH9 5 x CH Time [sec] Amplitude 1.5 Natural Frequencies from AVT f 1 = 1.94Hz f 2 = 2.344Hz f 3 = 2.637Hz 1.5 x FDD using truck induced vibration 1 X: Y: 1.245e X: Y: 4.84e-6 X: Y: e Frequency in Hz 73 Acceleration [g] TEST: 15mph, L1, BB-EB 5 x 1-3 CH1 5 x 1-3 CH3 5 x 1-3 CH5 5 x CH Time [sec] 5 x CH x CH x 1-3 CH9 5 x CH Time [sec] Amplitude Amplitude 2.5 x X: 1.94 Y: 1.366e-7 FDD using AVT data X: Y: 5.895e X: Y: 2.11e Frequency in Hz 1.5 x FDD using truck induced vibration 1 X: Y: 1.245e X: Y: 4.84e-6 X: Y: e Frequency in Hz Identified Truck Modal Shapes Truck Modal Shapes corresponding to bridge natural frequencies Modeling of Vehicle-Bridge Dynamic Interaction f = 1.94Hz f = 2.344Hz f = 2.637Hz Truck can now be modeled as a SDOF To take vehicle-bridge dynamic interaction into account f = 2.93Hz Truck Modal Shapes corresponding to chassis natural frequencies f = 6.55Hz 75 Compression Tension 76 Bridge Doctor Software Sensor Data from Instrumented Bridges Post-Event Rapid Damage Screening Detailed Damage Assessment Post-Event Disaster Mitigation & Speedy Recovery Remaining Capacity Estimation Long-Term Condition Assessment Structure Intelligent Instrumented Maintenance with Sensors Case Study #3: Integrated NDE of Concrete Bridge Decks (N. Gucunski, Rutgers University) Electrochemical methods for corrosion assessment Impact echo for delamination detection Ground Penetrating Radar (GPR) inspection Ultrasonic Surface Wave (USW) testing of concrete modulus 78 13

14 1/5/211 Electrochemical Methods for Corrosion Detection Principle of Half-Cell Potential Measurement Electrical Resistivity Half-cell probe (reference electrode) -414 mv Voltmeter Half-Cell Potential Iso-potential lines -6-7 Current flow Cathode Anode Cathode Half-Cell Potential Measurement and Map Principle of Electrical Resistivity Measurement Current I Voltage U Resistivity= 2dU/I Wenner probe d Iso-potential lines Current flow Electrical Resistivity Measurement and Map Delamination Detection by Impact Echo DISTANCE (FEET) Electrical resistivity (kohm*cm) DISTANCE FROM WEST ABUTMENT (FEET) very high high moderate - low low

15 1/5/211 Impact Echo Validation with Cores GPR Ground Coupled System Distance from west abutment, ft Air-Coupled (Horn) Antenna GPR System GPR Scan 2.6 GHz, Municipal Drive Bridge, Warren County, NJ GPR Raw Scan and Condition Map Debonded Overlay Ultrasonic Surface Waves (USW) Method Deteriorated Zone Coherence Phase Wavelength DISPERSION CURVE Phase velocity Depth YOUNG'S MODULUS PROFILE Shear modulus Top Rebar Amplitude (Normalized db) - DEPTH CORRECTED IMPACT SOURCE 1 2 RECEIVERS Wavelength considered less than layer thickness S S DISTANCE FROM EAST ABUTMENT (FEET) 15

16 1/5/211 USW Testing Using PSPA Concrete Modulus from USW (Bridge R1, Iowa) Distance from west abutment (ft) Young s Modulus (ksi) Sustainable Transportation Infrastructure Paradigm shift Periodic Visual Inspection Continuous Monitoring + Condition- Based Inspection 93 Courtesy of Business Week 16