Process Monitoring and Fault Diagnosis Based on Multivariable Statistical Analysis (Engineering Applications of Computational Methods, 19)

Preface
Acknowledgments
Contents
Notations
	Abbreviations
	Mathematical Notations
List of Figures
List of Tables
1 Introduction
	1.1 An Overview of Process Monitoring and Fault Diagnosis
		1.1.1 Data-Driven Process Monitoring
		1.1.2 MSPC-Based Process Monitoring
		1.1.3 Related Monographs for MSPC-Based Monitoring
	1.2 Aim and Main Features of This Book
		1.2.1 Aim of This Book
		1.2.2 Main Features of This Book
	1.3 Organization of This Book
	References
2 An Overview of Conventional MSPC Methods
	2.1 Overview
	2.2 Multivariable Statistical Analysis Models
		2.2.1 Principal Component Analysis
		2.2.2 Independent Component Analysis
		2.2.3 Kernel Principal Component Analysis
		2.2.4 Total Projection to Latent Structure
	2.3 Process Monitoring and Diagnosis Methods
		2.3.1 PCA-Based Process Monitoring
		2.3.2 PLS-Based Process Monitoring
		2.3.3 Contribution Plot-Based Fault Identification
		2.3.4 Reconstruction Based Fault Diagnosis
	2.4 Summary
	References
3 System-Wide Process Monitoring and Fault Diagnosis
	3.1 Introduction
	3.2 Review of Related PCA-Based Models Years, Many PCA
		3.2.1 Dynamic PCA Model
		3.2.2 Dynamic-Inner PCA Model
		3.2.3 Recursive PCA Model
		3.2.4 Moving-Window PCA Model
	3.3 Time-Varying Process Monitoring Based on Adaptive Eigensubspace Extraction
		3.3.1 Dynamic Process Monitoring Technique
		3.3.2 Computer Simulations
		3.3.3 Conclusion
	3.4 GPCA-Based FS Decomposition and Its Fault Reconst. Application
		3.4.1 PCA-Based Fault Detection
		3.4.2 Subspace Extraction Approach of Responsible Fault Deviations
		3.4.3 Illustration of Tennessee Eastman Process
		3.4.4 Conclusion
	3.5 Summary
	References
4 Quality-Related Time-Varying Process Monitoring
	4.1 Introduction
	4.2 Review of the Related Work
		4.2.1 Modified PLS Model
		4.2.2 Recursive PLS Model
		4.2.3 Orthogonal Signal Correction Model
		4.2.4 Concurrent PLS Model
	4.3 Quality-Relevant Process Monitoring Based on OSC and RMPLS
		4.3.1 OSC-RMPLS Applied to Quality-Relevant Fault Monitoring
		4.3.2 Slow-Time-Varying Process Monitoring Technology
		4.3.3 Conclusion
	4.4 Recursive CPLS-Based Quality-Relevant and Process-Relevant Fault Monitoring
		4.4.1 RPLS and RCPLS Models
		4.4.2 TEP Simulation Application
		4.4.3 Conclusion
	4.5 Summary
	References
5 Quality-Related Dynamic Process Monitoring: Part I
	5.1 Introduction
	5.2 A Review of the Dynamic PLS Model
	5.3 Quality-Related/Process-Related Fault Monitoring with Online Monitoring Dynamic CPLS
		5.3.1 Online Monitoring Dynamic PLS Model
		5.3.2 Online Monitoring Dynamic Concurrent PLS Model
		5.3.3 Dynamic Process Monitoring Technology
	5.4 Simulations and Applications
		5.4.1 Simulations of Quality-Related and Process-Related Fault Monitoring
		5.4.2 Case Study on the TEP
		5.4.3 Conclusion
	5.5 Summary
	Appendix
	References
6 Quality-Related Dynamic Process Monitoring: Part II
	6.1 Introduction
	6.2 Preliminaries
	6.3 Orthogonal Multiblock Algorithm and Its Monitoring Strategy
		6.3.1 Modified DPLS Model
		6.3.2 Orthogonal Multiblock Dynamic PLS
		6.3.3 Quality-Related Process Monitoring and Its Strategy
	6.4 Simulations and Applications
		6.4.1 Numerical Simulations
		6.4.2 Simulation on TEP
		6.4.3 Conclusion
	6.5 Summary
	References
7 Quality-Related Complex Nonlinear Process Monitoring
	7.1 Introduction
	7.2 Review of the Main Quality-Related Complex Nonlinear Process Monitoring
		7.2.1 Kernel PLS Model
		7.2.2 Total KPLS Model
		7.2.3 Concurrent Kernel PLS Model
		7.2.4 New Modified KPLS Model
	7.3 General Quality-Related Nonlinear Process Monitoring Based on IO-KPLS
		7.3.1 Nonlinear Input–Output Modeling and Monitoring Design
		7.3.2 Simulations and Applications Studies
		7.3.3 Conclusion
	7.4 Summary
	References
8 Quality-Related Fault Subspace Extraction for Fault Diagnosis
	8.1 Introduction
	8.2 Review of Related Work
		8.2.1 PLS and IPLS Model
		8.2.2 Generalized PCA Model
	8.3 Novel FS Extraction for the Reconst.-Based Fault Diagnosis
		8.3.1 Proposed Fault Subspace Extraction Method
		8.3.2 Quality-Related Fault Diagnosis Strategy
		8.3.3 Simulation and Applications
		8.3.4 Conclusion
	8.4 KPI-Related Fault Subspace Extraction for the Reconst.-Based Fault Diagnosis
		8.4.1 IPLS Model for Monitoring
		8.4.2 Proposed Quality-Related Fault Diagnosis Approach
		8.4.3 Simulation and Application
		8.4.4 Conclusion
	8.5 Summary
	References
9 Non-Gaussian Process Monitoring and Fault Diagnosis
	9.1 Introduction
	9.2 Review of ICA-Based Fault Monitoring Models
		9.2.1 Dynamic ICA Model
		9.2.2 Kernel ICA Model
		9.2.3 Kernel Dynamic ICA Model
	9.3 Extraction of Reduced Fault Subspace Based on KDICA and Its Fault Diagnosis Application
		9.3.1 Fault Reconstruction Based on KDICA Model
		9.3.2 Extraction of Fault Subspace and Fault Diagnosis
		9.3.3 Case Study of the TE Benchmark Process
		9.3.4 Conclusion
	9.4 Fault Detection and Diagnosis Based on MKICR
		9.4.1 Establishment of the MKICR Model
		9.4.2 Quality-Related Fault Detection Based on MKICR
		9.4.3 Detectability Analysis of the Quality-Related Faults
		9.4.4 Fault Diagnosis Based on MKICR
		9.4.5 Simulations and Discussion
		9.4.6 Conclusions
	9.5 Summary
	Appendix
	References
10 Hybrid Gaussian/Non-Gaussian Quality-Related Nonlinear Process Monitoring
	10.1 Introduction
	10.2 Review of the Related Work
		10.2.1 Modified KPLS Model
		10.2.2 KICA Model
	10.3 Quality-Related Process Monitoring Based on a Bayesian Classifier
		10.3.1 Variable Separation
		10.3.2 Feature Extraction
		10.3.3 Constructing a Bayesian-Based Quality-Related Classifier
	10.4 Case Studies
		10.4.1 Numerical Simulation Experiment
		10.4.2 Application to Tennessee-Eastman Process
	10.5 Summary
	References
11 Conclusions and Future Work
	11.1 Conclusions
	11.2 Future Work