Fault Diagnosis and Prognosis Techniques for Complex Engineering Systems

Front cover
Half title
Full title
Copyright
Contents
Contributors
Preface
Chapter 1 - Quality-related fault detection and diagnosis: a technical review and summary
1.1 Introduction
1.2 Basic methodology
1.3 Recent research
1.3.1 The KDD algorithm
1.3.2 The KLS-based approach
1.3.3 Reconstruction partial derivative contribution plot
1.3.4 Kernel sample equivalent replacement
1.4 Simulation
1.4.1 Introduction of the Tennessee-Eastman process
1.4.2 Fault detection results
1.4.3 Nonlinear fault detection using KSER
1.4.4 Fault diagnosis without smearing effect
Appendix A Description of the variables and faults
References
Chapter 2 - Canonical correlation analysis-based fault diagnosis method for dynamic processes
2.1 Introduction
2.2 Preliminaries
2.2.1 Basics of conventional CCA
2.2.2 Obtaining the positions and images in CCA
2.2.3 Details of the SVD-based technique
2.2.4 The CCA-based fault diagnosis method
2.2.5 Main steps of the CCA-based fault diagnosis method
2.3 CCA-based fault diagnosis method for dynamic processes
2.3.1 DCCA-based fault detection
2.3.2 The GRU-aided CCA fault detection method
2.4 Experimental results and analysis
2.4.1 The CSTR process
2.4.2 The TDCS process
2.5 Conclusion
Acknowledgments
References
Chapter 3 - H∞ fault estimation for linear discrete time-varying systems with random uncertainties
3.1 Introduction
3.2 Robust fault detection for LDTV systems with multiplicative noise
3.3 Robust fault detection for LDTV systems with measurement packet loss
3.4 Fixed-lag fault estimator design for LDTV systems under an unreliable communication link
3.5 Conclusion
Acknowledgments
References
Chapter 4 - Fault diagnosis and failure prognosis of electrical drives
4.1 Introduction
4.1.1 Operation under field orientation control
4.1.2 Operation under Direct Torque Control
4.2 What can fail and how
4.2.1 Electric power converters
4.2.2 Electrical machines
4.2.3 Capacitors
4.2.4 Batteries
4.3 Diagnosis methodology and tools
4.3.1 Signal selection
4.3.2 Signal features
4.3.3 Classification
4.4 Faults, their manifestation, and diagnosis
4.4.1 Winding faults in AC machines
4.4.2 Bearing faults
4.4.3 Insulation
4.4.4 Power electronics
4.4.5 Induction motor drives
4.4.6 PMAC drives
4.4.7 Switched reluctance machines
4.5 Failure prognosis, fault mitigation, and reliability
4.5.1 From diagnosis to prognosis
4.5.2 Prognosis tools
4.5.3 Applications and new developments
4.5.4 Decisions based on prognosis and mitigation
References
Chapter 5 - Intelligent fault diagnosis for dynamic systems via extended state observer and soft computing
5.1 Introduction
5.2 Extended state observer
5.2.1 ESO design
5.2.2 Estimation error convergence
5.3 Case study: three-tank dynamic system
5.4 Fault detection by means of ESO
5.4.1 Fault detection scheme
5.4.2 Fault detection without exact knowledge of the plant model
5.5 FAULT isolation and fault identification
5.5.1 Generation of reference values
5.5.2 Fault isolation by means of fuzzy inference and ESO
5.5.3 Fault identification via neural networks
5.6 Simultaneous faults of different types
5.6.1 Isolation of process faults
5.6.2 Isolation of sensor faults
5.6.3 Isolation of actuator faults
5.7 Isolation of simultaneous process faults and actuator faults
5.7.1 Characteristics of process faults and actuator faults
5.7.2 Utilizing an outflow sensor to isolate actuator faults
5.8 Conclusion and future work
References
Chapter 6 - Fault diagnosis and failure prognosis in hydraulic systems
6.1 Application status of sensor detection technology
6.1.1 Relevant standards of hydraulic machinery sensor detection technology
6.1.2 Instrumentation for the hydraulic turbine prototype
6.1.3 On-site detection for hydraulic turbines
6.2 Cavitation research
6.2.1 Establishment of cavitation theory
6.2.2 Numerical simulation of the cavitation mechanism
6.2.3 Cavitation model establishment and optimization
6.2.4 Engineering application of numerical simulation for cavitation flow in a hydraulic turbine
6.3 Intelligent evaluation and diagnosis technology
6.3.1 Current theoretical research hotspot
6.3.2 Commercial intelligent evaluation and fault diagnosis systems
6.3.3 Application status and deficiency of current diagnosis systems
6.4 Prognostics research
6.4.1 Prediction based on the classical linear time series model
6.4.2 Time Series Prediction Based on Intelligent Technology
6.4.3 Fuzzy time series prediction based on fuzzy set theory
6.4.4 Combination forecast
References
Chapter 7 - Fault detection and fault identification in marine current turbines
7.1 The HT-based detection method
7.1.1 Problem description
7.1.2 The HT-based detection method
7.1.3 Simulation results and analysis
7.2 The wavelet threshold denoising-based dectection method
7.2.1 Problem description
7.2.2 The wavelet threshold denoising-based detection method
7.2.3 Simulation results and analysis
7.2.4 Experimental results and analysis
7.3 The identification method of blade attachment based on the sparse autoencoder and softmax regression
7.3.1 Problem description
7.3.2 The recognition method based on the sparse autoencoder and softmax regression
7.3.3 Experimental results and analysis
7.4 The identification method of blade attachment based on depthwise separable CNN
7.4.1 Problem description
7.4.2 The recognition method based on depthwise separable CNN
7.4.3 Experimental analysis
7.5 Conclusion and future works
References
Chapter 8 - Quadrotor actuator fault diagnosis and accommodation based on nonlinear adaptive state observer
8.1 Introduction
8.2 Mathematical model of a quadrotor
8.2.1 The nonlinear quadrotor model
8.2.2 The actuator fault model
8.3 Naso-based FTC
8.3.1 The fault detection module
8.3.2 The fault diagnosis module
8.3.3 The fault accommodation module
8.4 Validation
8.4.1 Numerical simulation results
8.4.2 Flight test
8.5 Conclusion
References
Chapter 9 - Defect detection and classification in welding using deep learning and digital radiography
9.1 Introduction
9.1.1 Welding Process
9.1.2 Digital Radiography
9.2 Literature Review
9.3 Database Preparation
9.4 Experimental Study
9.4.1 Deep Learning Architecture
9.4.2 Training
9.4.3 Network HP Optimization
9.5 Experimental Implementation
9.6 Conclusion
References
Chapter 10 - Real-time fault diagnosis using deep fusion of features extracted by PeLSTM and CNN
10.1 Introduction
10.2 Basic theory
10.2.1 Convolutional neural network
10.2.2 Long short-term memory
10.3 Deep fusion of feature extracted by PeLSTM and CNN
10.3.1 2D screenshot image construction
10.3.2 The feature fusion algorithm based on CNN and PeLSTM
10.4 Experimental testing
10.4.1 Rolling bearing test and analysis
10.4.2 Gearbox test and analysis
10.5 Conclusion and future work
Acknowledgment
References
Index
Back cover