Bayesian Computational Sensor Networks
Bayesian Computational Sensor Networks: Small-scale Structural Health Monitoring Wenyi Wang, Anshul Joshi, Nishith Tirpankar, Philip Erickson, Michael Cline, Palani Thangaraj, and Tom Henderson DDDAS ICCS Workshop Reyjkavik, Iceland 1-3 June 2015 AFOSR-FA9550-12-1-0291
1
DDDAS Workshop June 2015
Colleagues
Eddie Grant Tom Henderson Dan Adams
Anshul Joshi
2
Wenyi Wang
Nishith Tirpankar
Philip Erickson
V. John Mathews
Michael Cline
Palani Thangaraj
DDDAS Workshop June 2015
2020 Vision for SHM 3
http://www.cs.wright.edu/~jslater/SDTCOutreachWebsite/pic13.pdf
DDDAS Workshop June 2015
Goal: Replace Schedule-based Maintenance with Condition-based Maintenance
Schedule-based • Structural design • Perform full-scale fatigue testing and identify problem areas • Construct inspection schedule
Condition-based • Monitor aircraft continuously for problems • Do maintenance based on the state of the structure as need arises Increases availability Reduces maintenance costs Increases reliability
4
DDDAS Workshop June 2015
AF Structural Maintenance Offices (Slide Courtesy Jason Rich) AF Life Cycle Management Center Directorate (EN-EZ) Mr. Jorge Gonzalez Wright-Patterson AFB Product Support Engineering Division (EZP) Ms. Debbie Naguy Wright-Patterson AFB
Sustainment Technology Transition Branch (EZPT) Frank Erdman Wright-Patterson AFB
AF Corrosion Prevention and Control Office Wes Barfield Robins AFB
AF Nondestructive Inspection Office Mike Paulk Tinker AFB
AF Advanced Composites Office Jason Rich Hill AFB
What ACO Does (Slide Courtesy Jason Rich) • Mission defined by AFI-20-114 •
Provide organic subject matter experts to the AF; ensure AF interests and MAJCOM priorities are represented/protected for future composite-intensive weapon systems.
• How •
Provide Engineering and Technical Assistance to Composites Users
•
Provide and Support Advanced Composites Training
•
Support Advanced Composites Health & Safety Issues
•
Assist Technology Transfer
•
Conduct joint service conferences to discuss composite repair/sustainment issues
•
Manage 3 Technical Orders • TO1-1-690, General Advanced Composite Repair Manual • TO 00-80C, Crash Damaged & Disabled Aircraft Recovery
• TO 1-1-694 Low Observables Maintenance
SHM Needs in New Composite Airplanes
• Assess the flightworthiness of airplanes each time before they take off using ultrasound (Lamb waves)
From: http://www.stanford.edu/~gorin/papers/IWSHM2007OptEst.pdf
7
DDDAS Workshop June 2015
DDDAS Approach for Lamb Wave SHM • The multimodal and dispersive characteristics of Lamb waves may change due to changes in environmental conditions and structural properties. • This may result in failure of models (used to locate and characterize damage) based on previous knowledge of the medium. • In the data driven approach, the models are developed based on the data collected from the experiments. • The inference about the location, type and extent of damage is drawn based on these models. 8
DDDAS Workshop June 2015
Mode Identification Based on Group Velocity Estimation
Excitation signal (X) and sensed signal(Y)
•
•
Band pass filter X and Y
Hilbert transform of the filtered signals to calculate envelope of the signals
Cross correlation between the envelopes to estimate time delay (∆𝑡)
Group ∆𝐷 velocity= ∆𝑡
∆𝐷 is the distance separation between actuator and sensor Center frequency is shifted by some step and the process is repeated to get a range of velocities.
Inference from the plot: First wave packet in the sensed signal is S0 mode. Third wave packet is A0 mode
Estimated group velocity for mode identification
9
DDDAS Workshop June 2015
Separation of Overlapped Modes When Dispersion Characteristics for Structure are Approximately Known • • • • •
The technique assumes the knowledge of dispersion curves of the material. A filter based on the Wiener-Hopf criterion is used. Input to the filter: Modes generated using analytical model. Desired signal : Result of the experiments at the receiver sensor. The estimated output: Combination of individual estimated modes.
Here S0(t) and A0(t) are the modes calculated using analytical model and X(t) is the experimental result.
10
DDDAS Workshop June 2015
Theoretical Vs Estimated Velocity
Final Signal reconstruction is in the stage of implementation. Once individual modes are identified and extracted, damage mapping can be done.
11
DDDAS Workshop June 2015
A Damage Mapping Example
SHM systems can perform as well or better than C-scan-based NDI systems at much lower cost, faster speed, and without expert human supervision
12
DDDAS Workshop June 2015
DDDAS Major Objectives 1. Bayesian Computational Sensor Networks • Detect & identify structural damage • Quantify physical phenomena and sensors • Characterize uncertainty in calculated quantities of interest
13
DDDAS Workshop June 2015
Major Objectives (cont’d) 2. Active feedback methodology using model-based sampling regimes • Embedded and active sensor placement • On-line sensor model validation • On-demand sensor complementarity
14
DDDAS Workshop June 2015
Major Objectives (cont’d) 3. Rigorous uncertainty models of: • System states • Model parameters • Sensor network parameters (e.g., location, noise) • Material damage assessments
15
DDDAS Workshop June 2015
Major Objectives (cont’d): 4. Exploit Grid or Cloud Resources: • • • •
16
Real-time Data Storage Real-time On-line Sensor Reading Model Simulation and Calibration Information Feedback
DDDAS Workshop June 2015
BCSN for SHM
17
DDDAS Workshop June 2015
Futuristic Idea: sLAMBats!
18
DDDAS Workshop June 2015
Goal: BCSN in the Cloud
19
DDDAS Workshop June 2015
Experiments on an Aluminum Plate
• Aluminum panel of thickness 1.6mm • Fixed sensors (Acellent) • Mobile sensors (VS900-RIC Sensor: 100-900 kHz, ceramic face, integrated preamplifier; 34 dB gain) • Excitation signal – 5 cycle, Hanning window
20
DDDAS Workshop June 2015
sLAMBot: First Cut
21
DDDAS Workshop June 2015
Damage Detection (cont’d)
22
DDDAS Workshop June 2015
Dynamic Data-Driven Approach for SHM
Data collection using mobile SLAMBOT on surface of the structure
23
Identification of modes from a mixture of modes and reflected wave packets
Separation of overlapped modes (in case they overlap)
Locating damage and its extent using the information extracted from individual modes
DDDAS Workshop June 2015
Damage Detection: Based on Ultrasound Physics Model
24
DDDAS Workshop June 2015
Test Configuration
25
DDDAS Workshop June 2015
Ultrasound Signals Received
26
DDDAS Workshop June 2015
Recovered Ellipses
27
DDDAS Workshop June 2015
SLAM Exploitation • Lamb Wave Data (ellipses) provide landmarks that are incorporated into SLAM algorithm • Robot Motion Model based on standard 2track model • Uncertainties characterized by Lamb wave properties and motion error 28
DDDAS Workshop June 2015
sLAMBot II Camera
29
Ultrasound Sensors DDDAS Workshop June 2015
EKF SLAM • Estimate 𝑝(𝑥𝑡 , 𝑚|𝑧 1:𝑡 , 𝑢 1:𝑡 ) where 𝑥𝑡 : robot’s pose at time t 𝑚: map of environment 𝑧 1:𝑡 : sensor measurements 𝑢 1:𝑡 : control values 30
DDDAS Workshop June 2015
SLAM
See paper for details! (or Thrun, Burgard, & Fox)
31
DDDAS Workshop June 2015
Vision & Ultrasound SLAM Y-axis h
Map damage Locations & Uncertainty
X-axis 32
DDDAS Workshop June 2015
Broader SHM Issues • Update the wave propagation model (i.e., parameter calibration of physical properties) • Perform optimal sensor placement (i.e., robot motion control) • Store image and ultrasound data off-line • Exploit other possible localization possibilities (e.g., RSS from wireless) • Use structure-based sensor/actuators 33
DDDAS Workshop June 2015
Robot Monitoring Agents: Contract Net 1. Broadcasts task with eligibility spec, abstract, bid spec, & expiration time Manager Agent
2. Bid on tasks and provide info on their capabilities Contractor Agents 3. Manager awards contracts
Timeline 34
4. Contractor agents complete tasks DDDAS Workshop June 2015
Data Routing Model for Cloud Computing • Data sharing model • Highly customizable • Low latency between sensing and computational resources • Optimized socket applications
• Dynamic routing 35
DDDAS Workshop June 2015
Data Routing Model for Cloud Computing
36
DDDAS Workshop June 2015
Data Routing Model for Cloud Computing • Sensor Nodes: put data on queues • RabbitMQ Message Broker: fault tolerant, persistent messaging • ContractNet Manager: allocates work • RabbitMQ/ContractNet Management: web interface message monitor • Processor Nodes: remote machine process 37
DDDAS Workshop June 2015
Validation Experiment: SLAM
38
DDDAS Workshop June 2015
Validation Experiment: SLAM Lamb Wave Boundary Detection
39
DDDAS Workshop June 2015
Conclusions & Future Work Conclusions: • Combines SLAM with Lamb Waves • Multiagent Contract Net Model with Cloud Computing • Validation Experiment 40
DDDAS Workshop June 2015
Conclusions & Future Work Future Work: Refine sLAMBot
• Further analysis of Lamb wave uncertainties • Extend to multiple robots (plus quadrotors) • Extend to composite materials 41
DDDAS Workshop June 2015
Conclusions & Future Work Future Work: Major Extensions
• Add a logical sentence layer and use argumentation techniques to analyze and combine inconsistent sources • Expand to large-scale infrastructure health monitoring 42
DDDAS Workshop June 2015
Bayesian Argumentation Framework (with X. Fan)
43
DDDAS Workshop June 2015
Other DDDAS-based SHM Research Directions
44
DDDAS Workshop June 2015
BRECCIA
45
DDDAS Workshop June 2015
Questions?
46
DDDAS Workshop June 2015
1
DDDAS Workshop June 2015
Colleagues
Eddie Grant Tom Henderson Dan Adams
Anshul Joshi
2
Wenyi Wang
Nishith Tirpankar
Philip Erickson
V. John Mathews
Michael Cline
Palani Thangaraj
DDDAS Workshop June 2015
2020 Vision for SHM 3
http://www.cs.wright.edu/~jslater/SDTCOutreachWebsite/pic13.pdf
DDDAS Workshop June 2015
Goal: Replace Schedule-based Maintenance with Condition-based Maintenance
Schedule-based • Structural design • Perform full-scale fatigue testing and identify problem areas • Construct inspection schedule
Condition-based • Monitor aircraft continuously for problems • Do maintenance based on the state of the structure as need arises Increases availability Reduces maintenance costs Increases reliability
4
DDDAS Workshop June 2015
AF Structural Maintenance Offices (Slide Courtesy Jason Rich) AF Life Cycle Management Center Directorate (EN-EZ) Mr. Jorge Gonzalez Wright-Patterson AFB Product Support Engineering Division (EZP) Ms. Debbie Naguy Wright-Patterson AFB
Sustainment Technology Transition Branch (EZPT) Frank Erdman Wright-Patterson AFB
AF Corrosion Prevention and Control Office Wes Barfield Robins AFB
AF Nondestructive Inspection Office Mike Paulk Tinker AFB
AF Advanced Composites Office Jason Rich Hill AFB
What ACO Does (Slide Courtesy Jason Rich) • Mission defined by AFI-20-114 •
Provide organic subject matter experts to the AF; ensure AF interests and MAJCOM priorities are represented/protected for future composite-intensive weapon systems.
• How •
Provide Engineering and Technical Assistance to Composites Users
•
Provide and Support Advanced Composites Training
•
Support Advanced Composites Health & Safety Issues
•
Assist Technology Transfer
•
Conduct joint service conferences to discuss composite repair/sustainment issues
•
Manage 3 Technical Orders • TO1-1-690, General Advanced Composite Repair Manual • TO 00-80C, Crash Damaged & Disabled Aircraft Recovery
• TO 1-1-694 Low Observables Maintenance
SHM Needs in New Composite Airplanes
• Assess the flightworthiness of airplanes each time before they take off using ultrasound (Lamb waves)
From: http://www.stanford.edu/~gorin/papers/IWSHM2007OptEst.pdf
7
DDDAS Workshop June 2015
DDDAS Approach for Lamb Wave SHM • The multimodal and dispersive characteristics of Lamb waves may change due to changes in environmental conditions and structural properties. • This may result in failure of models (used to locate and characterize damage) based on previous knowledge of the medium. • In the data driven approach, the models are developed based on the data collected from the experiments. • The inference about the location, type and extent of damage is drawn based on these models. 8
DDDAS Workshop June 2015
Mode Identification Based on Group Velocity Estimation
Excitation signal (X) and sensed signal(Y)
•
•
Band pass filter X and Y
Hilbert transform of the filtered signals to calculate envelope of the signals
Cross correlation between the envelopes to estimate time delay (∆𝑡)
Group ∆𝐷 velocity= ∆𝑡
∆𝐷 is the distance separation between actuator and sensor Center frequency is shifted by some step and the process is repeated to get a range of velocities.
Inference from the plot: First wave packet in the sensed signal is S0 mode. Third wave packet is A0 mode
Estimated group velocity for mode identification
9
DDDAS Workshop June 2015
Separation of Overlapped Modes When Dispersion Characteristics for Structure are Approximately Known • • • • •
The technique assumes the knowledge of dispersion curves of the material. A filter based on the Wiener-Hopf criterion is used. Input to the filter: Modes generated using analytical model. Desired signal : Result of the experiments at the receiver sensor. The estimated output: Combination of individual estimated modes.
Here S0(t) and A0(t) are the modes calculated using analytical model and X(t) is the experimental result.
10
DDDAS Workshop June 2015
Theoretical Vs Estimated Velocity
Final Signal reconstruction is in the stage of implementation. Once individual modes are identified and extracted, damage mapping can be done.
11
DDDAS Workshop June 2015
A Damage Mapping Example
SHM systems can perform as well or better than C-scan-based NDI systems at much lower cost, faster speed, and without expert human supervision
12
DDDAS Workshop June 2015
DDDAS Major Objectives 1. Bayesian Computational Sensor Networks • Detect & identify structural damage • Quantify physical phenomena and sensors • Characterize uncertainty in calculated quantities of interest
13
DDDAS Workshop June 2015
Major Objectives (cont’d) 2. Active feedback methodology using model-based sampling regimes • Embedded and active sensor placement • On-line sensor model validation • On-demand sensor complementarity
14
DDDAS Workshop June 2015
Major Objectives (cont’d) 3. Rigorous uncertainty models of: • System states • Model parameters • Sensor network parameters (e.g., location, noise) • Material damage assessments
15
DDDAS Workshop June 2015
Major Objectives (cont’d): 4. Exploit Grid or Cloud Resources: • • • •
16
Real-time Data Storage Real-time On-line Sensor Reading Model Simulation and Calibration Information Feedback
DDDAS Workshop June 2015
BCSN for SHM
17
DDDAS Workshop June 2015
Futuristic Idea: sLAMBats!
18
DDDAS Workshop June 2015
Goal: BCSN in the Cloud
19
DDDAS Workshop June 2015
Experiments on an Aluminum Plate
• Aluminum panel of thickness 1.6mm • Fixed sensors (Acellent) • Mobile sensors (VS900-RIC Sensor: 100-900 kHz, ceramic face, integrated preamplifier; 34 dB gain) • Excitation signal – 5 cycle, Hanning window
20
DDDAS Workshop June 2015
sLAMBot: First Cut
21
DDDAS Workshop June 2015
Damage Detection (cont’d)
22
DDDAS Workshop June 2015
Dynamic Data-Driven Approach for SHM
Data collection using mobile SLAMBOT on surface of the structure
23
Identification of modes from a mixture of modes and reflected wave packets
Separation of overlapped modes (in case they overlap)
Locating damage and its extent using the information extracted from individual modes
DDDAS Workshop June 2015
Damage Detection: Based on Ultrasound Physics Model
24
DDDAS Workshop June 2015
Test Configuration
25
DDDAS Workshop June 2015
Ultrasound Signals Received
26
DDDAS Workshop June 2015
Recovered Ellipses
27
DDDAS Workshop June 2015
SLAM Exploitation • Lamb Wave Data (ellipses) provide landmarks that are incorporated into SLAM algorithm • Robot Motion Model based on standard 2track model • Uncertainties characterized by Lamb wave properties and motion error 28
DDDAS Workshop June 2015
sLAMBot II Camera
29
Ultrasound Sensors DDDAS Workshop June 2015
EKF SLAM • Estimate 𝑝(𝑥𝑡 , 𝑚|𝑧 1:𝑡 , 𝑢 1:𝑡 ) where 𝑥𝑡 : robot’s pose at time t 𝑚: map of environment 𝑧 1:𝑡 : sensor measurements 𝑢 1:𝑡 : control values 30
DDDAS Workshop June 2015
SLAM
See paper for details! (or Thrun, Burgard, & Fox)
31
DDDAS Workshop June 2015
Vision & Ultrasound SLAM Y-axis h
Map damage Locations & Uncertainty
X-axis 32
DDDAS Workshop June 2015
Broader SHM Issues • Update the wave propagation model (i.e., parameter calibration of physical properties) • Perform optimal sensor placement (i.e., robot motion control) • Store image and ultrasound data off-line • Exploit other possible localization possibilities (e.g., RSS from wireless) • Use structure-based sensor/actuators 33
DDDAS Workshop June 2015
Robot Monitoring Agents: Contract Net 1. Broadcasts task with eligibility spec, abstract, bid spec, & expiration time Manager Agent
2. Bid on tasks and provide info on their capabilities Contractor Agents 3. Manager awards contracts
Timeline 34
4. Contractor agents complete tasks DDDAS Workshop June 2015
Data Routing Model for Cloud Computing • Data sharing model • Highly customizable • Low latency between sensing and computational resources • Optimized socket applications
• Dynamic routing 35
DDDAS Workshop June 2015
Data Routing Model for Cloud Computing
36
DDDAS Workshop June 2015
Data Routing Model for Cloud Computing • Sensor Nodes: put data on queues • RabbitMQ Message Broker: fault tolerant, persistent messaging • ContractNet Manager: allocates work • RabbitMQ/ContractNet Management: web interface message monitor • Processor Nodes: remote machine process 37
DDDAS Workshop June 2015
Validation Experiment: SLAM
38
DDDAS Workshop June 2015
Validation Experiment: SLAM Lamb Wave Boundary Detection
39
DDDAS Workshop June 2015
Conclusions & Future Work Conclusions: • Combines SLAM with Lamb Waves • Multiagent Contract Net Model with Cloud Computing • Validation Experiment 40
DDDAS Workshop June 2015
Conclusions & Future Work Future Work: Refine sLAMBot
• Further analysis of Lamb wave uncertainties • Extend to multiple robots (plus quadrotors) • Extend to composite materials 41
DDDAS Workshop June 2015
Conclusions & Future Work Future Work: Major Extensions
• Add a logical sentence layer and use argumentation techniques to analyze and combine inconsistent sources • Expand to large-scale infrastructure health monitoring 42
DDDAS Workshop June 2015
Bayesian Argumentation Framework (with X. Fan)
43
DDDAS Workshop June 2015
Other DDDAS-based SHM Research Directions
44
DDDAS Workshop June 2015
BRECCIA
45
DDDAS Workshop June 2015
Questions?
46
DDDAS Workshop June 2015
