LIVE DAMAGE DETECTION Team Members: Eric Snyder , Hewei Ma, Nicole De Giulio Technical Advisor: Dr. Joel Harley (University of Utah) Faculty Advisor: Dr. Angela Rasmussen (University of Utah) Industry Advisor: Chris Deemer (Orbital ATK) Special Thanks: Jon Davies (University of Utah)
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
LIVE DAMAGE DETECTION
Team Members: Eric Snyder, Hewei Ma, Nicole De GiulioTechnical Advisor: Dr. Joel Harley (University of Utah)Faculty Advisor: Dr. Angela Rasmussen (University of Utah)Industry Advisor: Chris Deemer (Orbital ATK)Special Thanks: Jon Davies (University of Utah)
CALIFORNIA METHANE LEAK
1.6 million lbs leaked per day≈ 4.5 million cars¼ of California’s total emissions
(Sources: Time Magazine and CNN)
CALIFORNIA METHANE LEAK
PIPELINE INCIDENTS
Source: http://www.phmsa.dot.gov/ (Pipeline and Hazardous Materials Safety Administration)
FATALITIES
WHO BENEFITS?
Visual Inspection C-Scan Microwave
METHODS FOR DETECTION
Built a prototype to detect and monitor for damage over long periods of time.
GOAL OF PROJECT
Ultrasonic Guided Waves(Lamb Waves)
OUR METHODThin Aluminum Plate
ActiveEric Hewei
PassiveNicole
Implementation
ACTIVE SYSTEMGenerator
Sensor
Damage
COLLECT BASELINE
No damage present
RECEIVED DATA
Received with damage
Use Correlation Coefficient r:
𝑟𝑟 = �𝑛𝑛
𝑧𝑧𝑥𝑥 𝑛𝑛 𝑧𝑧𝑦𝑦 𝑛𝑛
∑ 𝑧𝑧𝑥𝑥2 𝑛𝑛 ∑ 𝑧𝑧𝑦𝑦2 𝑛𝑛
CORRELATION
LOW CORRELATION
HIGH CORRELATION
DAMAGE DETECTED
If r is below threshold, damage assumed
THRESHOLD
Direct Signal
Received Signal = Reflected Signal + Direct Signal + NoiseBaseline Signal = Direct Signal + Noise
BASELINE SUBTRACTION
Direct Signal
Received Signal - Reflected SignalBaseline Signal =
ISOLATE REFLECTED SIGNAL
SUBTRACTED SIGNAL
Damage Locating Technique
DELAY AND SUM
SIMPLE EXAMPLE
1. Delay Generated Signal2. Sum with each Reflection Signal3. Sum all points in resulting signal4. Repeat
DELAY AND SUM STEPS
d1
d1_a
d1_b
d1 = d1_a + d1_b
SIMULATED RESULTS
10 -3
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Ampl
itude
-2
0
2
Baseline
10 -3
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Ampl
itude
-2
0
2
Received
Time (s) 10 -3
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Ampl
itude
-2
0
2
Subtracted
Time (s) 10 -3
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Ampl
itude
-0.5
0
0.5
Subtracted10 -3
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Ampl
itude
-5
0
5
Generated
x direction (meters)0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
y di
rect
ion
(met
ers)
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
Plot w/ Damage
Plot w/ Damage
Generator Location
Estimated Impact Location
Actual Impact Location
Sensors
Detects damage in simulations using correlation
Locates damage in simulations using Delay and Sum
CONCLUSION
“An ounce of prevention is worth a pound of cure.”
Benjamin Franklin
[1] T. Clarke and P. Cawley, Enhancing the Defect Localization Capability of a Guided Wave SHM System Applied to a Complex Structure, Structural Health Monitoring: Sage, 2010.[3] C. Liu, J.B. Harley, M. Berges, D.W. Greve, W.R. Junker, I.J. Oppenheim, “A robust baseline removal method for guided wave damage localization,” SPIE, vol. 9061, 2014.[4] E. B. Flynn, et al., “Maximum-likelihood Estimation of Damage Location in Guided-Wave Structural Health Monitoring”, Proceedings of the Royal Society, 2011.[5] J. E. Michaels, “Detection, Localization and Characterization of damage in Plates with an In Situ Array of Spatially Distributed Ultrasonic Sensors,” ECE, Georgia Institute of Technology, Atlanta, GA, 2008.[6] J. E. Michaels, A. J Croxford, Paul D Wilcox, “Imaging Algorithms for Locating Damage via in situ Ultrasonic Sensors,” ECE, Georgia Institute of Technology, Atlanta, GA, 2008.[7] J. E. Michaels, J. S. Hall, T. E. Michaels, “Adaptive Imaging of Damage from Changes in Guided Wave Signals Recorded from Spatially Distributed Arrays,” EVE, Georgia Institute of Technology, Atlanta, GA, 2009.
REFERENCES
LIVE DAMAGE DETECTION
Team Members: Hewei Ma, Nicole De Giulio, Eric SnyderTechnical Advisor: Dr. Joel Harley (University of Utah)Faculty Advisor: Dr. Angela Rasmussen (University of Utah)Industry Advisor: Chris Deemer (Orbital ATK)Special Thanks: Jon Davies (University of Utah)
ActiveEric Hewei
PassiveNicole
Implementation
PASSIVE SYSTEM• Damage Detection
• Damage Localization
[9] Traditional Damage Detection Method
PASSIVE EXPLANATION
Time (s) 10 -3
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Ampli
tude
10 -3
-2
-1
0
1
2
3
4
5
Simulated Lamb Wave w/ Noise
Tippmann’s Algorithm
Threshold Voltage
TRIGGER
NARROWBAND FILTER
10 -4
0 0.5 1 1.5 2 2.5 3
Ampl
itude
0
0.05
0.1
Before Filtering
Time (s) 10 -4
0 0.5 1 1.5 2 2.5 3
Ampl
itude
0
0.05
0.1
After Filtering
ENVELOPE METHOD
ENVELOPE METHOD
MULTILATERATION
∆12 = 𝑥𝑥𝑥 − 𝑥𝑥𝑥𝑥 2 + 𝑦𝑦𝑥 − 𝑦𝑦𝑥𝑥 2 − 𝑥𝑥𝑥 − 𝑥𝑥𝑥𝑥 2 + 𝑦𝑦𝑥 − 𝑦𝑦𝑥𝑥 2
(x1,y1)
(x2,y2)
(xp,yp)
EQUATIONS
CURVE INTERSECTION
x direction (meters)0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
y di
rect
ion
(met
ers)
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
Plot w/ Damage
Plot w/ Damage
Estimated Impact Location
Actual Impact Location
Sensors
Actual Location: (0.3, 0.3)
• Damages Detected
• Multilateration Algorithm
• 96% Accuracy in MATLAB
• >90% Accuracy in LabVIEW
CONCLUSION
“With a working passive damage monitoring
system, technicians don’t need to climb 300
feet high wind turbines to find a small impact.
They could do it remotely and easily.” By Hewei Ma
[1] J. D. Tippmann, X. Zhu and F. Lanza di Scalea, ‘‘Application of damage detection methods usingpassive reconstruction of impulse response functions’’, University of California, San Diego CA 92093, USA[2] Marchi L D, Marzani A, Speciale N and Viola E 2011 A passive monitoring technique based on dispersion compensation to locate impacts in plate-like structures Smart Mater. Struct. 20 035021[3] Ciampa F and Meo M 2010 Acoustic emission source localization and velocity determination of the fundamental mode A0 using wavelet analysis and a Newton-based optimization technique Smart Mater. Struct. 19 045027[4] Lee, H. Accuracy limitations of hyperbolic multilateration systems. IEEE Transactions on Aerospace and Electronic Systems, AES-11, 1 (Jan. 1975), 16–29.
REFERENCES
[5] B. Gueye et al., “Constraint-based Geolocation of Internet Hosts,” IEEE/ACM Trans. Net., vol. 14, no. 6, Dec. 2006, pp. 1219–32.[6] B. Xu, L Yu, and V Giurgiutiu, “Advanced Methods for Time-Of-Flight Estimation with Application to Lamb Wave Structural Health Monitoring,” Stanford University, Palo Alto, CA. Rep. 2009.[7]S. M. Ziola, M. R. Gorman, “Source Location in Thin Plates Using Cross correlation”, Naval Postgraduate School, Monterey, CA, Dec, 1991[8] T Clarke and P Cawley, Enhancing the Defect Localization Capability of a Guided Wave SHM System Applied to a Complex Structure, Structural Health Monitoring: Sage, 2010.[9] LM Wind Power, http://www.sunwindenergy.com/wind-energy/high-tech-remedies, online resources, 2016.
[10] V. Kratochvil, “Public Domain Pictures”, http://technology.nasa.gov/t2media/tops/img/LAR-TOPS-128/TOP1_front.jpg, Online resources, 2016
REFERENCES
LIVE DAMAGE DETECTION
Team Members: Nicole De Giulio, Hewei Ma, Eric SnyderTechnical Advisor: Dr. Joel Harley (University of Utah)Faculty Advisor: Dr. Angela Rasmussen (University of Utah)Industry Advisor: Chris Deemer (Orbital ATK)Special Thanks: Jon Davies (University of Utah)
ActiveEric Hewei
PassiveNicole
Implementation
• Simulation proved effective• Active ~90% accuracy• Passive ~90% accuracy
• Next stage: experiments
No knowledge prior to this project
LABVIEW
DAQmx
DATA ACQUISITION
Passive SystemTriggering
Active SystemGenerating Signal
DATA ACQUISITION
Receiving Generating
CONSTRUCT PLATE
ACTIVE SYSTEM
Plot
LocateActive
Delay&Sum
Window
Filter
Received
GaussianWave
DataSetup
ACTIVE SYSTEM
GAUSSIAN WAVE
Time (s) 10 -4
4 4.2 4.4 4.6 4.8 5 5.2 5.4 5.6 5.8 6
Ampl
itude
(V)
-1
-0.5
0
0.5
1
Generated Signal
RECEIVED SIGNAL
FILTER
DAMAGE DETECTION
Filtered Received
Filtered Baseline
Correlation Coefficient
𝑟𝑟 = �𝑛𝑛
𝑧𝑧𝑥𝑥 𝑛𝑛 𝑧𝑧𝑦𝑦 𝑛𝑛
∑ 𝑧𝑧𝑥𝑥2 𝑛𝑛 ∑ 𝑧𝑧𝑦𝑦2 𝑛𝑛
WINDOW
Time (s)0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008 0.009 0.01
Ampl
itude
(V)
10 -3
-4
-3
-2
-1
0
1
2
3
Windowed Signal
DELAY AND SUM
ACTIVE RESULTS
PASSIVE SYSTEM
Trigger
Window
Envelope & Filter
Time of Arrival
Multilateration
Plotting
PASSIVE SYSTEM
TRIGGER
Time (s)0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008 0.009 0.01
Ampl
itude
(V)
-0.3
-0.2
-0.1
0
0.1
0.2
0.3
Received Signal
Trigger Threshold
Time (s)0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008 0.009 0.01
Ampl
itude
(V)
-0.02
-0.01
0
0.01
0.02
Windowed Signal
WINDOW
ENVELOPE AND FILTER
Time (s)0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008 0.009 0.01
Ampl
itude
(V)
10 -3
-1
-0.5
0
0.5
1
Filtered and Enveloped Signal
TIME OF ARRIVAL
MULTILATERATION
∆12 = 𝑥𝑥𝑥 − 𝑥𝑥𝑥𝑥 2 + 𝑦𝑦𝑥 − 𝑦𝑦𝑥𝑥 2 − 𝑥𝑥𝑥 − 𝑥𝑥𝑥𝑥 2 + 𝑦𝑦𝑥 − 𝑦𝑦𝑥𝑥 2
Possible grid pointsX
Y
Signal Estimated Point
PASSIVE RESULTS
LIVE DEMO
• Runs live• Detects damage on the plate• Active and Passive have ~100% detection accuracy
• Passive has ~89% locating accuracy
CONCLUSION
ActiveEric Hewei
PassiveNicole
Implementation
• Why?• Who?
CONCLUSION
• “Turn your wounds into wisdom.”-Oprah Winfrey
• [1] M. Niethammer, L.J. Jacobs, “Time-frequency representation of Lamb waves,” School of Civil & Environmental Engineering, GIT, Atlanta, GA.
• [2] P.D. Wilcox et al., “Mode and transducer selection for long range lamb wave inspection,” Journal of Intelligent Material Systems and Structures, vol. 12, 2001.
• [3] P. Cawley, “Practical long range guided wave inspection-managing complexity,” Department of Mechanical Engineering, Imperial College, London.
• [4] P. Wilcox et al., “The effect of dispersion on long-range inspection using ultrasonic guided waves,” NDET&E International, vol. 34, 1-9, 2001.
• [5] P. Cawley, D. Alleyne, “The use of Lamb waves for the long range inspection of large structures,” Ultrasonics international, vol. 34, pages 287-290, 1996
REFERENCES
Related Documents
