Smart Structures Chenyang Lu Cyber-Physical Systems Laboratory
American Society for Civil Engineers 2017 Report Card for America's Infrastructure Ø Bridges C+ q Almost four in 10 are 50 years or older. q 56,007 (9.1%) bridges were structurally deficient. q Backlog of bridge rehabilitation needs: $123 billion. q https://youtu.be/JjN1FwzbJaY Ø Dams D Ø Levees D Ø Roads D Ø … … Ø America's Infrastructure GPA: D+ http://www.infrastructurereportcard.org 2
Structural Health Monitoring Current Practice Ø Bridges: inspected manually once every two years. Ø Costly and time consuming. Highway 40 Closing for Boone Bridge Inspection Monday August 10, 2009 If you're heading to St. Charles this weekend, Highway 40 is not your best option. Westbound 40 from Long Road in St. Louis County to Route 94 in St. Charles County will be closed (weather permitting) while work crews inspect the Daniel Boone Bridge across the Missouri River. The road will close at 5:30 a.m. on August 15 and won't reopen until sometime after 9 p.m. on August 16. 3
Smart Bridges Ø Bridges live for 50-100 years q Need to make sure they remain safe for a long time Ø Smart monitoring and control systems to prevent… Minneapolis Bridge Collapse Freeway after 1989 San Francisco Earthquake 4
Structural Health Monitoring Ø Monitor a bridge using a wireless sensor network q Detect and localize damages to structures Ø Smart q No human effort Ø Real-time q Every week q Right after an earthquake Ø Accurate q Technology instead of human eyes 5
Wireless Structural Health Monitoring Ø Detect and localize damages to structures q at high spatiotemporal granularities Ø Challenges q Computationally intensive q Resource constraints q Long-term monitoring 6
Existing (non-CPS) Approach Ø Centralized: stream all data to base station for processing. q Too energy-consuming for long-term monitoring Ø Example: Golden Gate Bridge project [Kim IPSN'07]. q Nearly 1 day to collect enough data. q Lifetime of 10 weeks w/4 x 6V lantern battery. Ø Separate designs of sensor networks (cyber) and damage detection (physical). Ø Sensor networks focus on data transport. Ø Not concerned with method for damage detection. 7
Edge Computing Ø Dilemma Raw Data q Too much sensor data to stream to the base station q Damage detection algorithms are too complex to run entirely on sensors Results ➪ Edge computing Partial q Perform part of computation on sensor nodes q Send (smaller) intermediate results to base station q Complete computation at base station 8
Cyber-Physical Co-design Ø Employ damage detection approach amenable for Raw Data distributed implementation in sensor networks. Ø Optimally map damage detection algorithm onto distributed architecture. Results Partial 9
Our Solution q Physical: Damage Localization Assurance Criterion (DLAC) [Messina96] q Identify structure’s natural frequencies based on vibration data. • “Signature” of structure’s health q “Match” natural frequencies to structural models with damages. q Cyber: optimally partition data flow between sensors and base station. q Minimize energy consumption q Subject to resource constraints 10
D Integers D: # of samples (1) FFT P: # of natural freq. (D » P) 2D Floats (3a) Coefficient (2) Power Spectrum Extraction 5*P D Floats Floats (3b) Equation (3) Curve Fitting Solving P Floats Healthy Model Damaged Location (4) DLAC Data Flow Analysis DLAC Algorithm 11
4096 bytes D: 2048 (1) FFT P: 5 Integer: 2 bytes 8192 bytes Float: 4 bytes (3a) Coefficient Effective compression (2) Power Spectrum Extraction ratio of 204:1 100 4096 bytes bytes (3b) Equation (3) Curve Fitting Solving 20 bytes Healthy Model Damaged Location (4) DLAC Data Flow Analysis DLAC Algorithm 12
Implementation Ø Platform: Imote2 + ITS400 sensor board q 13 – 416 MHz XScale CPU q 32 MB ROM, 32 MB SDRAM q CC2420 802.15.4-compliant radio q 3-axis accelerometer on sensor board Ø Data collection and processing application written with TinyOS 1.1 q 243 KB ROM, 71 KB RAM 13
Evaluation: Truss 5.6m steel truss structure at UIUC Ø Ø 14 0.4m long bays, on 4 rigid supports Ø 11 Imote2s attached to frontal pane DLAC WS #32 DLAC WS #45 DLAC WS #67 DLAC WS #28 DLAC WS #35 DLAC WS #75 1 1 1 1 1 1 X = 3 X = 3 X = 3 X = 3 X = 3 0.9 0.9 0.9 Y = 0.871 0.9 Y = 0.873 0.9 0.9 Y = 0.868 Y = 0.864 Y = 0.865 X = 3 Y = 0.825 0.8 0.8 0.8 0.8 0.8 0.8 0.7 0.7 0.7 0.7 0.7 0.7 0.6 0.6 0.6 0.6 0.6 0.6 0.5 0.5 0.5 0.5 0.5 0.5 0.4 0.4 0.4 0.4 0.4 0.4 Damage correctly Truss Frontal Panel 0.3 0.3 0.3 0.3 0.3 0.3 Wireless Sensor 0.2 0.2 0.2 0.2 0.2 0.2 localized to third bay � 0.1 0.1 0.1 0.1 0.1 0.1 0 0 0 0 0 0 123456789 10 11 12 123456789 10 11 12 123456789 10 11 12 123456789 10 11 12 123456789 10 11 12 123456789 10 11 12 Truss Central Bay Position Truss Central Bay Position Truss Central Bay Position Truss Central Bay Position Truss Central Bay Position Truss Central Bay Position 14
Centralized Sampling Decentralized Computation Communication 0 0.05 0.1 0.15 0.2 0.25 Energy consumption (mAh) Energy Consumption Evaluation 15 �
Sampling Equation Solving Computation Communication Coefficient Extraction Power Spectrum FFT Raw Data Collection 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 Energy Consumption (mAh) Energy Consumption Evaluation 16 �
ROM Raw Data Collection FFT RAM Power Spectrum Coefficient Extraction Equation Solving 0 50000 100000 150000 200000 250000 Size (bytes) Memory Consumption Evaluation 17 �
What we have learned so far Ø Cyber-physical co-design of a distributed SHM system. q Reduces energy consumption by 71% q Implemented on iMote2 using <1% of its memory Ø Effectively localized damage on two physical structures. Ø Demonstrated the promise of cyber-physical co-design. G. Hackmann, F. Sun, N. Castaneda, C. Lu and S. Dyke, A Holistic Approach to Decentralized Structural Damage Localization Using Wireless Sensor Networks, RTSS 2008. 18
Hierarchical Damage Localization Ø The DLAC method employs no collaboration among sensors à limitations in SHM capabilities. q For example, cannot detect multiple damages. Ø New hierarchical architecture for collaborative localization. q Embed processing into a hierarchical architecture q Send (smaller!) partial results between layers of hierarchy q Multi-level damage localization Ø Demonstrate the generality of cyber-physical co-design 19
Flexibility-based Methods Ø Flexibility acts as a “signature” of the structure’s health q Structures flex slightly when a force is applied q Structural weakening => decreased stiffness Ø Two flexibility-based methods of interest for our work q Beam-like structures: Angles-Between-String-and-Horizon flexibility method (ASHFM) [Duan, J. Structural Engineering and Mechanics 09] q Truss-like structures: Axial Strain flexibility method (ASFM) [Yan, J. Smart Structures and Systems 09] θ � 20
Hierarchical Architecture 1 Ø Sensors form groups Base Station Ø Group members q collect raw vibration data q à power spectrum Group Group Leader Leader Ø Group leaders q collect and correlate power Group Group Member spectrum from children Member Group q à modal parameters (natural Member frequencies + mode shapes) Group Group Member Member 21
Hierarchical Architecture 2 Ø Base station q collects modal parameters from group leaders q à structural flexibility q compared against “baseline” collected when structure was known to be healthy Ø Differences in flexibility à localize damage 22
Distributed Data Flow Group Member � Group Leader � Base Station � Sensing Flexibility Cross Spectral 2D ints Density FFT D matrices D: # of samples D floats P: # of natural freq. Singular Value (D » P) Power Spectrum Decomposition D floats P natural frequencies + mode shapes 23
Enhanced Distributed Data Flow Group Member � Group Leader � Base Station � Sensing Flexibility 2D ints Cross Spectral Density FFT P matrices D floats D: # of samples P: # of natural freq. Singular Value Power Spectrum (D » P) Decomposition D floats Peak Picking P natural frequencies + P floats mode shapes 24
Multi-Level Damage Localization Ø Only a handful of sensors are needed to detect damage Ø As more sensors are added, localization gets more precise Ø Save energy by exploiting localized nature of flexibility-based approach 25
Multi-Level Damage Localization Ø Only a handful of sensors are needed to detect damage Ø As more sensors are added, localization gets more precise Ø Save energy by exploiting localized nature of flexibility-based approach 26
Multi-Level Damage Localization Ø Only a handful of sensors are needed to detect damage Ø As more sensors are added, localization gets more precise Ø Save energy by exploiting localized nature of flexibility-based approach 27
Implementation Ø Hardware: Imote2 + ITS400 sensor board q 13 – 416 MHz PXA271 XScale CPU q 32 MB ROM, 32 MB SDRAM q CC2420 802.15.4-compliant radio q 3-axis accelerometer on sensor board Ø Software platform q TinyOS 1.1 operating system q UIUC’s ISHM toolsuite used for sensing, reliable communication, and time sync 28
Recommend
More recommend
