UNCLASSIFIED FileName.pptx UNCLASSIFIED UNCLASSIFIED UNCLASSIFIED Presented to: Prognostic Working Group 15 October 2014 U.S. Army Aviation and Missile Research, Development, and Engineering Center Presented by: Jean P. Vreuls Lead Systems Engineer [email protected]// 256-990-6195 Diagnostic / Prognostic Laboratory U.S. Army Aviation and Missile Research, Development, and Engineering Center Structural Health Monitoring (SHM) It’s Eat Our Lunch!
46
Embed
UNCLASSIFIED FileName.pptx UNCLASSIFIED UNCLASSIFIED Presented to: Prognostic Working Group 15 October 2014 U.S. Army Aviation and Missile Research, Development,
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
UNCLASSIFIED FileName.pptx
UNCLASSIFIED
UNCLASSIFIED
UNCLASSIFIED
Presented to:Prognostic Working Group
15 October 2014
U.S. Army Aviation and Missile Research, Development, and
Structural Health Monitoring (SHM) promises to do the following:
1. Reduce Unnecessary Inspections –– By monitoring the structure maintainers can move away from usage based inspection and
only perform them when damage in suspected. This removes a potential source of damage since a disassembly can often result in damage the structure (dents, scratches that break the corrosion barrier, etc.)
2. Increased asset availability – with less scheduled maintenance an asset is available for duty
3. Reduced burden on the Warfighter– An automated inspection process frees up a serviceman for other more important tasks
4. Increases safety – automated inspections reduces the risk of missing faults
5. Reduces costs – An automated SHM enables the prediction of when a component will fail. Maintainers with
this knowledge can anticipate maintenance actions and reduces the amount of spares needed thus shortening the logistics chain. Another factor is the unscheduled maintenance is by far the most costly type in the Army. Just by reducing that will allow for a large savings.
Structural Health MonitoringStructural Health Monitoring
UNCLASSIFIED FileName.pptx
UNCLASSIFIED
1. Detection
– Is there a problem?
2. Localization
– Where is the problem?
3. Classification
– How bad is the problem?
4. Prognostication
– How long before I need a repair?
Levels of SHMLevels of SHM
UNCLASSIFIED FileName.pptx
UNCLASSIFIED
REPEATABLE DESIGN PROCESS
REPEATABLE DESIGN PROCESS
UNCLASSIFIED FileName.pptx
UNCLASSIFIED
Design FrameworkDesign Framework
MODELS &CONSTRAINTS
SENSORS OPTIMIZATION SIMULATE
SIGNAL PROCESSING
METHOD
ANALYZE
AMRDEC Design Optimizes, Physics-Based Models, and Sensors for Structural Health Monitoring…
UNCLASSIFIED FileName.pptx
UNCLASSIFIED
• Methodology Differentiator: Optimization
– High sensitivity to likely damage areas (Hotspots)
– Ability to detect damage globally
– Minimum number of sensors / Minimize cost
– Reliability
– Design robustness to modeling error and manufacturing variations
OptimizationOptimization
p
DDQD
p
QDDQ
p
D
p
CCQC
p
QCCQ
p
C
p
Q T
uTuT
u
T
xTxT
xy
Repeatable Design Process
UNCLASSIFIED FileName.pptx
UNCLASSIFIED
Use
Physics
Based
Models
Optimal Sensors & Actuator Design
• Objective functions
• Algorithm
Evaluate
Design
Trade-off’s
Simulate
DesignsImplement
AMRDEC SHM Design ProcessAMRDEC SHM Design Process
Repeatable Verified Design Process…
UNCLASSIFIED FileName.pptx
UNCLASSIFIED
DECTECTIONDECTECTION
UNCLASSIFIED FileName.pptx
UNCLASSIFIED
Diagnostics and Prognostics Lab Demonstrations
Diagnostics and Prognostics Lab Demonstrations
Rotor Wing Aircraft Roof Strap and Drag Beam
UNCLASSIFIED FileName.pptx
UNCLASSIFIED
Drag BeamNo ‘Hot Spots’Known
Drag BeamNo ‘Hot Spots’Known
Random Sensor
Random Sensor
Random Sensor
Random Sensor
Optimum Sensor
Optimum Sensor
Actuator
Yellow = ActuatorRed = Optimum SensorsPurple = Random Sensors
Yellow = ActuatorRed = Optimum SensorsPurple = Random Sensors
0.5-10 kHz Excitement
UNCLASSIFIED FileName.pptx
UNCLASSIFIED
Unanticipated DamageLoosening of One Bolt Roof Strap
Unanticipated DamageLoosening of One Bolt Roof Strap
55 in-lbs 55 in-lbs45 in-lbs Finger Tight
• Optimal• Heuristic
UNCLASSIFIED FileName.pptx
UNCLASSIFIED
Damaged Roof StrapDamaged Roof Strap
0.25”
UNCLASSIFIED FileName.pptx
UNCLASSIFIED
Expected Damage LocationDemonstrated on the Roof Strap
Expected Damage LocationDemonstrated on the Roof Strap
No Damage 0.05” Cut
0.10” Cut
0.15” Cut
0.25” Cut
• Optimal• Heuristic
UNCLASSIFIED FileName.pptx
UNCLASSIFIED
Wing FittingWing Fitting
• Three specimens were tested
• Specimens taken from wing sections of aircraft that had been in-service
• Specimens were ~ 2m x 0.5m
• Skin panel
• 3 stiffeners
• U-channel fitting
• Model developed (DOF=42762)
• Expected damage locations known
UNCLASSIFIED FileName.pptx
UNCLASSIFIED
Test SpecimensTest Specimens
(a) Front side (in airstream) (b) Back side (inside wing)
Photographs showing the front side (a) and the back side of each specimen (b)
UNCLASSIFIED FileName.pptx
UNCLASSIFIED
Wing FittingSensor Location
Wing FittingSensor Location
The 5 Sensor 1 Actuator Design was chosen
8.7
27.2
15.0
1.8
2.9 3.8
16.1
29.1
1.8
20.7
UNCLASSIFIED FileName.pptx
UNCLASSIFIED
Wing Fitting ResultsWing Fitting Results
8 Test Hrs Before Failure
Stringer
Str
inge
r
Visual Detection
Optimal Design Detection
UNCLASSIFIED FileName.pptx
UNCLASSIFIED
DetectionFighter Aircraft Clevis
DetectionFighter Aircraft Clevis
UNCLASSIFIED FileName.pptx
UNCLASSIFIED
Detection – Results Detection – Results
• Detected at 16 kcycles• 0.03” Crack• 99.999% Confidence
baseline
dam
age
UNCLASSIFIED FileName.pptx
UNCLASSIFIED
LOCALIZATIONLOCALIZATION
UNCLASSIFIED FileName.pptx
UNCLASSIFIED
• Project Purpose– Does not replace current NDE/I– Guide maintainers and inspectors smartly
to the area of damage to perform NDE/I
• Paradigm– Works by identifying areas NOT having damage
Advantages– Does not need training data– Does not need high quality models
Damage LocalizationDamage Localization
UNCLASSIFIED FileName.pptx
UNCLASSIFIED
Damage & LocalizationDamage & Localization
98.4% of the area eliminated
UNCLASSIFIED FileName.pptx
UNCLASSIFIED
LocalizationRotor Wing Aircraft 409 Beam
UNCLASSIFIED FileName.pptx
UNCLASSIFIED
CORROSION CORRELATIONCORROSION
CORRELATION
UNCLASSIFIED FileName.pptx
UNCLASSIFIED
• Two unprotected 3” x 5” steel coupons• Three sensors per coupon• One piezo-electric actuator per coupon• Salt Fog applied at elevated temperature• Pictures taken three times a day
Corrosion Correlation TestCorrosion Correlation Test
UNCLASSIFIED FileName.pptx
UNCLASSIFIED
Test Results - Two SensorsTest Results - Two Sensors