View
4
Download
0
Category
Preview:
Citation preview
Structural Health Monitoring
Ph.D. Student: Zhongqing SUSupervisor: Dr. Lin Ye
Laboratory of Smart Materials & Structures (LSMS)School of Aerospace, Mechanical & Mechatronic Engineering
University of Sydney, Australia
IntroductionAn Active Online Structural Health Monitoring System Is Developed for the In-Service Composites Using Smart Materials (Piezoelectrics), to Syncretise the Designated Functions With Self-diagnosis and Self-rehabilitation Capabilities, Based on Which the Occurrence and Severity of the Damage Can Be Quantitatively Real-Time monitored.
Digital Damage Fingerprints (DDF)
Damage Parameters Database (DPD)
Artificial Intelligence Algorithm
Advanced Signal/Image Processing Algorithm
Active Sensor/Actuator Network
Active Real-Time Diagnosis System
Highlights
Application FieldsAerospaceAeronauticalMaritimeMechanicalCivil
Delamination
PZT Actuator PZT Actuator
PZT Sensor PZT Sensor
Multi-Lamb
Modes
Delamination
Basic Principle
(Lamb Wave Propagation-Based Identification)
Start
Configuration Design for Built-in Diagnosis Network System
ExperimentalMeasure
Signal Processing via Wavelet Technique
SpectrographicAnalysis
DamageDiagnosis End
Proposed Damage Identification Scheme
Damage ParametersInformation Database
and LibraryDirect Diagnosis
Σ11b
1T11n
Σ12b
1T12n
Σ1jb
1T1jn
.
.
.
Σ21b
2T21n
Σ22b
2T22n
Σ2kb
2T2kn
.
.
....
mim
3im
2im
1im
.
.
Σ31b
3T31n
Σ32b
3T32n
Σ3nb
3T3nn
.
.
....
nov
2ov
1ov
InputLayer
First NeuralProcessing Layer
Output Layer
111−w
1jmw −
211−w
2kjw−
311−w
3nkw−
Characteristicsof Signals
( )ζ
Second NeuralProcessing Layer
Diagnostic Results
ξ( )
( )φ
Artificial Neural Network Technique
– A Simulation of Human Brain
104 105 106 107
0.0
0.2
0.4
0.6
0.8
1.0
Nor
mal
ised
Am
plitu
de
Frequency ( Hz )
0 200 400 600 800 1000
-1.0
-0.5
0.0
0.5
1.0
Nor
mal
ised
Am
plitu
de
Time [ µs ]
5 mm-Depth
FFT
Wavelet Transform
Signal Identification
(Wavelet Algorithm)
Specimen
(Composite Laminate)
Main Control System
Software interface
CCU(HP Vectra Workstation)
Agilent VXI
Agilent ConditionerHP3242
Composite Laminate
P1
Delamination
Piezo System Amplifier EPA-104
SwitchSwitchP1 P3
P2
Fixed Support
Active Online Damage Diagnosis System
(VXI Platform)
Actual Example
Bonded
Piezoelectrics
CCU(HP Vectra Workstation)
Agilent VXI
Agilent ConditionerHP3242
Composite Laminate
P1
Delamination
Piezo System Amplifier EPA-104
SwitchSwitchP1 P3
P2
Fixed Support
Active Diagnosis System
In-Service Aircraft
Identification
Artificial Neural Network
Damage parameters database
Self-Rehabilitation
Failure Diagnosis
Lower Delaminated Surface
Neighboring Area
Upper Delaminated Surface
Case Study (1)
(Delamination Detection in Composites)
Actual Damage
Diagnosis B
Diagnosis C
Diagnosis DDiagnosis E
Case Study (2)
(Hole Detection in Composites)
Diagnosis A/B/C/D/E: Identification using 20/30/40/60/80 damage patterns for sub-database training
Damage Patterns for Network Training
Diagnosis A
Recommended