Post on 19-Dec-2015
1
Wireless Structural Sensors using Reliable Communication Protocols for
Data Acquisition and Interrogation
Yang Wang, Prof. Kincho H. Law
Department of Civil and Environmental Engineering, Stanford University
Prof. Jerome P. Lynch
Department of Civil and Environmental Engineering, University of Michigan
IMAC XXIII, Orland, FL, February 2, 2005
2
AgendaAgenda
Research background Hardware design of the latest wireless
sensing unit prototype Software design of the latest wireless SHM
system Laboratory validation tests Field validation tests Future direction
3
AgendaAgenda
Research background Hardware design of the latest wireless
sensing unit prototype Software design of the latest wireless SHM
system Lab validation tests Field validation tests Future direction
4
Structural Health Monitoring (SHM)Structural Health Monitoring (SHM)
S. Chase (2001), National Bridge Inspection Program (NBIP): Nearly 60,000 bridges in U.S. evaluated as structurally deficient.
Over 580,000 highway bridges in U.S. mandated by Federal Highway Administration for biannual evaluations.
Human Visual Inspection: resource consuming; visible damages only
Advanced Sensing Technology for Autonomous SHM:
Rapid, accurate, low-cost
5
From Wire-based Sensing to Wireless SensingFrom Wire-based Sensing to Wireless Sensing
E. G. Straser, and A. S. Kiremidjian (1998), Installation of wired system can take about 75% of testing time for large structuresM. Celebi (2002):
Estimation for each sensor channel and data recording system: $2,000; Installation (cabling, labor, etc.) per wired channel: $2,000. (Total: $4000)
Traditional DAQ System: wire-based
Future Wireless DAQ System
This wireless SHM prototype system is Jointly developed by researchers in Stanford University and the University of Michigan
6
Challenges in Wireless SHM (1)Challenges in Wireless SHM (1)
Limited power consumption: The wireless units most likely run on batteries.
BETTER PERFORMANCE: Long-distance high-speed wireless acquisition; Extensive local data processing.
LOWER POWER: Wireless communication consumes lots of power; Likewise for extensive local data processing
7
Challenges in Wireless SHM (2)Challenges in Wireless SHM (2)
Synchronization for data collected from multiple sensing units
Communication range
Limited bandwidth of wireless communication: impedes high-speed real-time data collection from multiple sensors
Failures in wireless data transmission
Communication protocol for the network: real-time data collection; multiple sensing units; synchronization; robust data transmission
8
Dr. E. G. Straser, Prof. A. Kiremidjian (1998)
Dr. J. P. Lynch, Prof. K. H. Law et al. (2002)
Y. Wang, Prof. J. P. Lynch, Prof. K. H. Law (2005)
Wireless SHM Unit Prototypes from Stanford and UMichWireless SHM Unit Prototypes from Stanford and UMich
L. Mastroleon, Prof. A. Kiremidjian et al (2004)
9
Design ObjectiveDesign Objective
Relatively low power
Long communication range for civil structural applications
Robust communication protocols for reliable data acquisition
Near-real-time, non-stopping wireless data collection, from multiple sensors, at an acceptable sampling frequency
Near-synchronized data collection
High-precision analog-to-digital conversion from multiple heterogeneous analog sensors
Considerable local data processing capability
Point-to-multipoint, and peer-to-peer communication
Low cost
10
AgendaAgenda
Research background Hardware design of the latest wireless
sensing unit prototype Software design of the latest wireless SHM
system Lab validation tests Field validation tests Future direction
11
Parallel Port4-channel 16-bitAnalog-to-Digital
ConverterADS8341
4-channel 16-bitAnalog-to-Digital
ConverterADS8341
40 kbps 9XCiteWireless
Modem with 900MHz Tranceiver
8-bit Micro-controllerAtmega128
Sensing Interface Computational CoreWireless
Communication
0 - 5VAnalog
Sensors
SPIPort
128kB External SRAMCY62128B
UARTPort
Functional Diagram of the Latest PrototypeFunctional Diagram of the Latest Prototype
Major power consumption when components are active:
9XCite – 45mA average between transmitting and receiving.
Atmega128 – 15mA running at 8MHz
CY62128B – 15mA
12
Power consumption: 75 – 80mA when active; 0.1mA standby
Communication range: 90m indoor, 300m outdoor
Near-real-time, non-stopping wireless data collection, from multiple sensors: up to 24 sensors at 50Hz sampling frequency
Near-synchronized data from multiple wireless sensing units
16bit Analog-To-Digital conversion, 4 A2D channels, 100kHz Fs
Local data processing: 4096-point float-point number FFT
Point-to-multipoint, and peer-to-peer communication
Total hardware cost: $130.00 each for small quantity production
Hardware Performance SummaryHardware Performance Summary
13
Picture of the Prototype with Wireless ModemPicture of the Prototype with Wireless Modem
14
Atmega128 Micro-
controller
Sensor Connector
4 Channel, 16-bit A2D Converter
Address Latch for the External
Memory
128kB External Memory
Power Switch
Connector to Wireless Modem
Prototype Double-layer Circuitry BoardPrototype Double-layer Circuitry Board
15
Container Dimension
4.02” x 2.56” x 1.57”
(10.2cm x 6.5cm x 4.0cm)
Antenna Length: 5.79” (14.7cm)
Wireless Sensing Unit Prototype PackageWireless Sensing Unit Prototype Package
Power supply requirement: 5V
16
AgendaAgenda
Research background Hardware design of the latest wireless
sensing unit prototype Software design of the latest wireless SHM
system Lab validation tests Field validation tests Future direction
17
Wireless Sensing NetworkWireless Sensing Network
Server-side software: Computer server manages network of multiple wireless sensing units
Firmware: Atmega128 micro-controller organizes different hardware modules of the wireless sensing units
Simple star topology
18
Objectives of Communication Protocol DesignObjectives of Communication Protocol Design
Major Challenge: Address Unreliability of Wireless Communication
Sufficiently reliable: detect the wireless communication failures; Successfully recover the system whenever a communication failure happens
Central server: active – responsible for reliability
Sensing units: passive – save power
Synchronization among different wireless sensing units
Retry and acknowledgement protocol to ensure the fidelity of data transfer
19
Reliable Beacon Signal Synchronization ProtocolReliable Beacon Signal Synchronization Protocol
•Straser and Kiremidjian (1998)
BroadcastBeacon signal to
all WSUs Begin sensor datasampling and
storage
Verify with eachWSU if it
received theBeacon signal
A WSU didn't receivethe Beacon signal
Inform WSUs* oneby one to restart
* WSU: Wireless Sensing Unit** CS: Central Server
Y
Receivedbeacon signal
Wait andRespond to
Beaconverification
Restart andacknowledge with
CS**, wait forBeacon signal
Wait for datacollection
command from CS
Receive restartcommand
Y
Start data collection fromWSUs one by one, and
round by roundTransmit data to
CS
Received datacollection request and
data is ready
Central Server Wireless Sensing Unit
Finish one round ofdata transmission
Wireless communicationand its direction
Y
Y
Approximate beginningsynchronization error:
20 micro-Seconds.
20
System complexity incurred by the unreliability of the wireless communication -> state machine concepts
A set of states A set of transitions among these states. At any point in time, the state machine can only be in one of the possible
states. In response to different events, the machine transits between its discrete
states. Visualize the communication protocol
State Machine Concept for Software DesignState Machine Concept for Software Design
State 2State 1Action
Condition
21
Communication State Diagram for Wireless Sensing UnitCommunication State Diagram for Wireless Sensing Unit
State1Wait forBeacon
State0Bootup
State2Wait forBeaconVerifi.
State3Wait for
DataCollect.
Initialize memory
No condition
Send LostBeacon
Received BeaconVerification
Start data collection
Received Beacon
Send BeaconConfirmation
Received BeaconVerification
Send BeaconConfirmation
Received BeaconVerification
Send AckRestart
Received Restart
Send AckRestart
Received Restart
Send DataReady
Received Request for datacollection and data is ready
DataTransfer
Send AckRestart
Received Restart
Service
Event
22
Dual Stack Memory Allocation for Non-stopping DAQDual Stack Memory Allocation for Non-stopping DAQ
Instant 1
Storingsensor data
Storingsensor data
Memory Map for Sensor 1 Memory Map for Sensor 2
Currently transfering
Storingsensor data
Storingsensor data
Instant 2
Stack 1-1 Stack 2-1
Stack 1-2
Stack 1-2
Stack 1-1 Stack 2-2Stack 2-1
Stack 2-2
Currently transfering
•Straser and Kiremidjian (1998)
•Lynch and Law et al. (2002)
23
AgendaAgenda
Research background Hardware design of the latest wireless
sensing unit prototype Software design of the latest wireless SHM
system Lab validation tests Field validation tests Future direction
24
Wireless Sensing Unit
Crossbow CXL01LF1
Accelerometer
Wireless Modem Receiving Data from
Wireless Sensing Units
Lab Validation TestsLab Validation Tests
25
Sampling frequency: 200Hz. Sampling resolution: 16 bits
Real-time non-stopping data collection from four wireless sensing units
Crossbow CXL02LF1 accelerometer, RMS noise floor of 1mg
Bosch SMB110 accelerometer, RMS noise floor of 7mg
Some Test ParametersSome Test Parameters
26
Discrete Fourier Transform of the third-floor acceleration during free-vibration
Three theoretical natural frequencies: 2.08Hz, 5.71Hz, and 8.18Hz
First Lab Test: Free VibrationFirst Lab Test: Free Vibration
27
Figure 7 – Ground acceleration time history
Second Lab Test: Chirping-signal Ground MotionSecond Lab Test: Chirping-signal Ground Motion
28
Comparison of measured and theoretical absolute acceleration at third floor
Absolute Acceleration at Third FloorAbsolute Acceleration at Third Floor
Measured
Computed
29
1st floor 2nd floor 3rd floor
Measured (m/s2) 2.14 3.35 4.74
Theoretical (m/s2) 2.07 3.76 4.93
Relative Difference 3.32% 11.5% 3.93%
Comparison of measured and theoretical maximum absolute acceleration
Maximum Absolute Acceleration at Three FloorsMaximum Absolute Acceleration at Three Floors
30
Acceleration Data and FFT Results of the Shake-table Excited by 5.7Hz Sinusoidal Input
Sampling Frequency: 100Hz. Total time: 40.96 Seconds. Total Number of Pts: 4096
Third Lab Test: Sinusoidal Ground MotionThird Lab Test: Sinusoidal Ground Motion
0 5 10 15 20 25 30 35 40 45-0.2
-0.15
-0.1
-0.05
0
0.05
0.1
0.15Time History
Acc
eler
atio
n (g
)
time (Sec.)
0 1 2 3 4 5 6 7 8 9 100
100
200
300
400
500Absolute Values of FFT to Time History Computed Onboard By Unit No.1
Mag
.
Freq.(Hz)
31
Third-floor Acceleration Data
Sampling Frequency: 100Hz. Total time: 40.96 Seconds. Total Number of Pts: 4096
Third Test: Sinusoidal Ground MotionThird Test: Sinusoidal Ground Motion
0 5 10 15 20 25 30 35 40 45-0.5
0
0.5Time History
Acc
eler
atio
n (g
)
time (Sec.)
0 1 2 3 4 5 6 7 8 9 100
200
400
600
800
1000
1200
1400Absolute Values of FFT to Time History Computed Onboard By Unit No.3
Mag
.
Freq.(Hz)
32
AgendaAgenda
Research background Hardware design of the latest wireless
sensing unit prototype Software design of the latest wireless SHM
system Lab validation tests Field validation tests Future direction
33
Geumdang Bridge Test, KoreaGeumdang Bridge Test, Korea
12
14 12
11 10 9 8 7 6 5 4 3
13
NORTH
Pier Pier PierAbutment
46.0m18.7m 18.7m
34
Sensor PropertyPCB Piezoelectric
(Cable System)PCB MEMS Capacitive
(Wireless System)
Maximum Range 1 g 3 g
Sensitivity 10 V/g 0.7 V/g
Bandwidth 2000 Hz 80 Hz
RMS Resolution (Noise Floor) 50 g 0.5 mg
Wire-based System versus Wireless SystemWire-based System versus Wireless System
Sampling Frequency 200Hz 70Hz
35
Bridge Traffic ExcitationBridge Traffic Excitation
36
155 160 165 170 175 180 185 190-0.04
-0.02
0
0.02
0.04KAIST Data Aquisition System
Time (sec)
Acc
eler
atio
n (g
)Sensor Location #8
155 160 165 170 175 180 185 190-0.04
-0.02
0
0.02
0.04Wireless Data Aquisition System
Time (sec)
Acc
eler
atio
n (g
)
Sensor Location #8
Time History Comparison Between Two SystemsTime History Comparison Between Two Systems
Wireless System
Wire-based System
Difference in sensor and signal conditioning
37
0 2 4 6 8 10 12 14 160
1
2
3
4
FFT - KAIST DAQ
Frequency (Hz)
Mag
nitu
de
Sensor Location #8
0 2 4 6 8 10 12 14 160
0.2
0.4
0.6
0.8
FFT - Wireless DAQ
Frequency (Hz)
Mag
nitu
de
Sensor Location #8
Comparison of FFT to Time HistoryComparison of FFT to Time History
Wireless System
Wire-based System
38
Validation Results SummaryValidation Results Summary
Rapid installation, and low cost
Communication range: 50 – 60m in box girder
Networked real-time and non-stopping data collection: 14 wireless sensors at 70Hz sampling frequency
Data is near-synchronized
Robust communication protocols for reliable data acquisition
Local data processing capability
Analog-to-digital precision
Point-to-multipoint, and peer-to-peer communication
39
AgendaAgenda
Research background Hardware design of the latest wireless
sensing unit prototype Software design of the latest wireless SHM
system Lab validation tests Field validation tests Future direction
40
Future DirectionFuture Direction
To be improved for current prototype system:
Power design, anti-noise circuits design
Accommodation for digital sensors
Greater wireless communication range, higher data rate
Multiple receivers at different channels to improve data rate
Large-scale data collection from densely allocated sensors
More local data analysis and damage identification algorithm
Handling mal-functioning sensing units
Further field validation tests
41
AcknowledgementAcknowledgement
National Science Foundation CMS-9988909 and CMS-0421180.
The University of Michigan Rackham Grant and Fellowship Program
The Office of Technology Licensing Stanford Graduate Fellowship
Smart Infrastructure Technology Center at Korean Advanced Institute of Science and Technology (KAIST), Korea Highway Corporation
42
The EndThe End