Structural Health Monitoring of the Golden Gate Bridge

Structural Health Monitoring of the Golden Gate Bridge Sukun Kim, David Culler James Demmel, Gregory Fenves, Steve Glaser Thomas Oberheim, Shamim Pakzad UC Berkeley NEST Retreat Jan 15, 2004

Structure Monitoring Data Acquisition Data Collection Processing & Feedback

At the Golden Gate Bridge

Overview n Low cost structure monitoring - Monitor structure, and analyze the health of structure based on sensed data at low cost n n For Golden Gate Bridge, monitor vibration of bridge, and detect unusual behavior by wind, earthquake, or local damage Extend reach of Wireless Sensor Network in a different direction – high fidelity sampling n High accuracy, high frequency with low jitter, large amount of data

Table of Contents Overview n Data Acquisition n Accelerometer Board n High Frequency Sampling & Jitter n n Data Collection n Large-scale Reliable Data Transfer Signal processing & System Identification n Conclusion n Challenges & Future Work n

Accelerometer Board Crossbow donates manufacture and test n On board 16 bit ADC n Low-pass filter n Type Number of axis Range ADXL 202 E MEMS 2 -2 G ~ 2 G Silicon Designs 1221 L MEMS 1 -0. 1 G ~ 0. 1 G System noise floor Price 200(μG/√Hz) $10 30(μG/√Hz) $150

Table of Contents Overview n Data Acquisition n Accelerometer Board n High Frequency Sampling & Jitter n n Data Collection n Large-scale Reliable Data Transfer Signal processing & System Identification n Conclusion n Challenges & Future Work n

High. Frequency. Sampling Made by David Gay n Up to 6. 67 KHz with 4 bytes sample n Micro. Timer – Supports one timer, micro second level granularity n Buffer. Log – Has two buffers. One is filled up by upper layer application while the other buffer is written to flash memory as a background task n

Jitter Test (1 KHz, 5 KHz, 6. 67 KHz) n n n Peak to Peak is time to fill up buffer Spiky portion is time to write buffer to flash Can sample as long as the former is larger than the latter

Table of Contents Overview n Data Acquisition n Accelerometer Board n High Frequency Sampling & Jitter n n Data Collection n Large-scale Reliable Data Transfer Signal processing & System Identification n Conclusion n Challenges & Future Work n

Large-scale Reliable Data Transfer 4 Byte of data and 4 Byte of time stamp at 100 Hz in 100 nodes, transfer 40 pkt/s – Sample data for 5 minutes, and collect data for more than 5 hours!!! n Efficient and reliable data transfer is crucial n RAM to RAM one-hop transfer is implemented as a building block - LRX n

Table of Contents Overview n Data Acquisition n Accelerometer Board n High Frequency Sampling & Jitter n n Data Collection n Large-scale Reliable Data Transfer Signal processing & System Identification n Conclusion n Challenges & Future Work n

Signal Processing As an analog signal processing low-pass filter is used, which filters high frequency noise n For accelerometer board, low-pass filter with threshold frequency 25 Hz is used. Then ADC should sample at frequency much higher than 50 Hz by Nyquist theorem, and imperfect lowpass filter n As a digital signal processing, averaging is used. If noise follows Gaussian distribution, by averaging N numbers, noise decreases by a factor of sqrt(N) n

System Identification Identifying model of target system n By matching input to system and output from system, we can construct a mathematical system model. n Usual process is (1) fitting a general Box. Jenkins multi-input multi-output model to sampled data. (2) And natural frequencies, damping ratios and mode shape are then estimated using the estimated Box-Jenkins model. n Most part of system identification is under n

Table of Contents Overview n Data Acquisition n Accelerometer Board n High Frequency Sampling & Jitter n n Data Collection n Large-scale Reliable Data Transfer Signal processing & System Identification n Conclusion n Challenges & Future Work n

Conclusion n New challenges are analyzed which are brought by structure monitoring to wireless sensor network n High accuracy accelerometer, high frequency sampling with low jitter, low-pass filter, averaging, large-scale reliable data collection G Accelerometer variation. Acoustic m. G Temperature μG n. G Noise Gravity Variation Challenges versus Accuracy G m. G μG Large Scale Earthquake Traffic Local Damage Identification Detection Nuclear Test Detection n. G Possible Applications versus Accuracy

Table of Contents Overview n Data Acquisition n Accelerometer Board n High Frequency Sampling & Jitter n n Data Collection n Large-scale Reliable Data Transfer Signal processing & System Identification n Conclusion n Challenges & Future Work n

Challenges & Future Work Calibrating acceleration value to temperature n Time synchronization – RBS, TPSN n To maximize utility of channel, we need to monitor channel quality (loss rate), and throttle packet injection rate accordingly n Using LRX as a building block, multi-hop data collection need be implemented n TASK n

Jitter Test Histogram (1 KHz, 5 KHz, 6. 67 KHz) n n Jitter is within 10µs Peak at 625 ns – Wakeup time from sleep mode

Jitter Analysis Sampling Other job Non-preemptible portion (atomic section) Preemptible task portion Sample F(k 2) F(k 3) C C+T(k 1) C+T(k 2) . . . Jitter

LRX component (continued) Sender Receiver Open n Ack for Open Data Block 1 Data Block 2 Data Block 3 Data Block 4 DONE n Ack for Data DONE n Explicit open handshake - Data description and size of cluster is sent as a transfer request Data transfer is composed of multiple rounds. In each round, sender sends packets missing in the previous round Tear-down is implicit

Throttle for data packet is fixed at 10 pkt/s n Optimal case: window size is infinite n For the case with window size 16, throughput is 88% of optimal case. Considering loss rate of 3%, actual relative throughput is 91%, which is higher than 85% of channel utilization ratio. This is because control packets do not follow 10 packets/s. n n

n As loss rate increases, retransmission increases, and throughput decreases

Channel Utilization TOS_Msg LRX (only data) Total Data (bytes) 36 Meta Data (bytes) 7 Real Data (bytes) 29 Channel Utilization (%) 78. 38 Comparison to TOS_Msg 100 (%) 36 10 26 72. 22 89. 66 LRX (Window Size 16) 613 197 416 67. 86 84. 24 LRX (data only) is theoretical limit of LRX (when window size is infinite) n Usage LRX lowers channel utilization by 15% n

Signal Processing As an analog signal processing low-pass filter is used, which filters high frequency noise n For accelerometer board, low-pass filter with threshold frequency 25 Hz is used. Then ADC should sample at frequency much higher than 50 Hz by Nyquist theorem, and imperfect lowpass filter n As a digital signal processing, averaging is used. If noise follows Gaussian distribution, by averaging N numbers, noise decreases by a factor of sqrt(N) n

System Identification Identifying model of target system n By matching input to system and output from system, we can construct a mathematical system model. n Usual process is (1) fitting a general Box. Jenkins multi-input multi-output model to sampled data. (2) And natural frequencies, damping ratios and mode shape are then estimated using the estimated Box-Jenkins model. n Most part of system identification is under n

Table of Contents Overview n Data Acquisition n Accelerometer Board n High Frequency Sampling & Jitter n n Data Collection n Large-scale Reliable Data Transfer Signal processing & System Identification n Conclusion n Challenges & Future Work n

Conclusion n New challenges are analyzed which are brought by structure monitoring to wireless sensor network n High accuracy accelerometer, high frequency sampling with low jitter, low-pass filter, averaging, large-scale reliable data collection G Accelerometer variation. Acoustic m. G Temperature μG n. G Noise Gravity Variation Challenges versus Accuracy G m. G μG Large Scale Earthquake Traffic Local Damage Identification Detection Nuclear Test Detection n. G Possible Applications versus Accuracy

Table of Contents Overview n Data Acquisition n Accelerometer Board n High Frequency Sampling & Jitter n n Data Collection n Large-scale Reliable Data Transfer Signal processing & System Identification n Conclusion n Challenges & Future Work n

Challenges & Future Work Calibrating acceleration value to temperature n Time synchronization – RBS, TPSN n To maximize utility of channel, we need to monitor channel quality (loss rate), and throttle packet injection rate accordingly n Using LRX as a building block, multi-hop data collection need be implemented n TASK n

Questions n http: //www. cs. berkeley. edu/~binetude/ggb

Backup Slides

Cost Comparison Conventional piezoelectric accelerometer with PC system costs $40, 000 n Budget for structure monitoring budget is $1, 000 level n Wireless sensor network with MEM accelerometer costs $500 n n Cheaper by a factor of 100

Shaking Table Test n Silicon Design 1221 L is more quite, but less sensitive to dynamic movement

Noise Floor Test Blue – Seismic Vault n Red – Mc. Cone Hall n

Jitter Analysis (continued) T(i): execution time of atomic section i n X(i): a random variable F(k 2) uniformly distributed in [0, T(i)] F(k 3) n C: context switch time n F(i): frequency of occurrence of atomic C C+T(k 1) C+T(k 2). . . Jitter section i Assume that the probability of timer event occurring at any point in atomic section i is same, then jitter will follow C+X(i). Since jitter distribution of every atomic section begins from C, the frequency is highest near C and decreases as moving farther. And frequency drop at C+T(i) by F(i), since atomic section i will not have any distribution beyond C+T(i). Actually there is a peak at C, because when program is in preemptible section, it will immediately service timer event after context switch time C. Sample n n

Calculation of Transfer Timer n Let us assume each node store 4 Byte of data and 4 Byte of time stamp at 100 Hz. And assume there are 100 nodes, radio throughput is 1. 2 KB/s, and data is collected to one base station. If acceleration data worthy 5 minutes is collected, each node will transfer 240, 000 Bytes. 100 nodes will transfer 24, 000 Bytes. Since the end link to base station is a bottleneck, it will take more than 5 hours. We can see bandwidth is narrow compared to aggressive data sampling. Even if we alleviate this problem using multi-channel or multi-tier network, still we will be in short of bandwidth.

LRX component Transfers one data cluster, which is composed of several blocks. n One block fits into one packet, so the number of blocks is equal to window size. n Each data cluster has a data description. After looking at data description, receiver may deny data (receiver already has that data, or that data is not useful anymore). n

Sender

Receiver

Sender Receiver Open Ack for Open Data Block 1 Data Block 2 Data Block 3 Data Block 4 Ack for DONE Data DONE

Sender Receiver Open Ack for Open Data Block 1 Data Block 2

Sender Receiver Open Ack for Open Data Block 1 Data Block 2

Sender Data Block 1 Receiver Data Block 2 Data Block 3 Data Block 4 Ack for Data Block 2 Ack for DONE Data DONE

Sender Data Block 1 Receiver Data Block 2 Data Block 3 Data Block 4 Ack for Data DONE

Sender Data Block 1 Receiver Data Block 2 Data Block 3 Data Block 4 Ack for DONE Data DONE

Sender Data Block 1 Receiver Data Block 2 Data Block 3 Data Block 4 Ack for Data Block 2 Data Block 3 Ack for Data

Why Sender times out n There are two reasons why only sender times out and stimulate receiver for Ack. The first reason is shown in Figure 16. If sender doesn’t time out, for a receiver to make sure Ack is delivered to sender, receiver should get acknowledgement from sender for Ack itself. This is not good. So it is clear that sender should timeout. Given that sender times out, timeout of receiver makes no difference except that channel is wasted by unnecessary Ack from receiver. So timeout in only sender side is desirable. As a second reason, if receiver times out, in case like Figure 18 (if first Data after Ack is lost), second Data always collide with resent Ack of receiver. This is not a good phenomenon. Therefore, after sending last packet in each round, if acknowledgement does not come, sender sends the last packet in that round again to stimulate acknowledgement. However, this does not mean receiver has no timeout. Receiver waits sufficient amount of time, and if nothing happens, it regards the situation as a failure.

Imperfect Low-pass Filter Amplitude Filtering threshold Frequency

Time Synchronization n Temporal jitter is handled by high frequency sampling component. Spatial jitter should be solved by time synchronization. ITP [8] is a time synchronization protocol widely used in Internet. In wireless sensor network, there were several studies. In RBS [9], synchronization is done among receivers, eliminating sender’s jitter in media access. TPSN [10] put time stamp after obtaining channel. This gives even better synchronization accuracy than RBS (10μs compared to 20μs). Still there is a source of jitter at receiver side. As we saw in jitter for sampling, handling interrupt by radio can be delayed by atomic section of other activity. As suggested in [10], putting time stamp at MAC layer in receiver side will eliminate this jitter.

Table of Contents Overview n Data Acquisition n Accelerometer Board n High Frequency Sampling & Jitter n n Data Collection n Large-scale Reliable Data Transfer Signal processing & System Identification n Conclusion n Challenges & Future Work n

Acknowledgement n This work is supported, in part, by the National Science Foundation under Grant No. EIA 0122599.