Model-Based Fault Diagnosis in Dynamic Systems Using Identification Techniques

by Simani, Silvio; Fantuzzi, Cesare; Patton, Ronald J.

• ISBN13: 9781852336851
• ISBN10: 1852336854
• Format: Hardcover
• Copyright: 2003-01-01
• Publisher: Springer Verlag

Summary

Safety in industrial process and production plants is a concern of rising importance, especially if people would be endangered by a catastrophic system failure. On the other hand, because the control devices which are now exploited to improve the overall performance of industrial processes include both sophisticated digital system design techniques and complex hardware (input-output sensors, actuators, components and processing units), there is an increased probability of failure. As a direct consequence of this, control systems must include automatic supervision of closed-loop operation to detect and isolate malfunctions as early as possible. One of the most promising methods for solving this problem is the "analytical redundancy" approach, in which residual signals are obtained. The basic idea consists of using an accurate model of the system to mimic the real process behaviour. If a fault occurs, the residual signal, i.e., the difference between real system and model behaviours, can be used to diagnose and isolate the malfunction. This book focuses on model identification oriented to the analytical approach of fault diagnosis and identification. The problem is treated in all its aspects covering:· choice of model structure;· parameter identification;· residual generation;· fault diagnosis and isolation. Sample case studies are used to demonstrate the application of these techniques. Model-based Fault Diagnosis in Dynamic Systems Using Identification Techniques will be of interest to researchers in control and fault identification. Industrial control engineers interested in applying the latest methods in fault diagnosis will benefit from the practical examples and case studies.

Table of Contents

Symbols and Abbreviations	xv
Introduction	1	(18)
Nomenclature	3	(2)
Fault Detection and Identification Methods based on Analytical Redundancy	5	(2)
Model-based Fault Detection Methods	7	(1)
Model Uncertainty and Fault Detection	8	(1)
The Robustness Problem in Fault Detection	9	(2)
System Identification for Robust FDI	11	(1)
Fault Identification Methods	12	(1)
Report on FDI Applications	13	(3)
Outline of the Book	16	(2)
Summary	18	(1)
Model-based Fault Diagnosis Techniques	19	(42)
Introduction	19	(1)
Model-based FDI Techniques	20	(1)
Modelling of Faulty Systems	21	(7)
Residual Generator General Structure	28	(3)
Residual Generation Techniques	31	(13)
Residual Generation via Parameter Estimation	32	(3)
Observer-based Approaches	35	(5)
Fault Detection with Parity Equations	40	(4)
Change Detection and Symptom Evaluation	44	(1)
The Residual Generation Problem	45	(6)
Fault Diagnosis Technique Integration	51	(8)
Fuzzy Logic for Residual Generation	51	(2)
Neural Networks in Fault Diagnosis	53	(1)
Neuro-fuzzy Approaches to FDI	54	(2)
Structure Identification of NF Models	56	(1)
NF Residual Generation Scheme for FDI	57	(2)
Summary	59	(2)
System Identification for Fault Diagnosis	61	(54)
Introduction	61	(1)
Models for Linear Systems	62	(2)
Parameter Estimation Methods	64	(11)
System Identification in Noiseless Environment	65	(3)
System Identification in Noisy Environment	68	(5)
The Frisch Scheme in the MIMO Case	73	(2)
Models for Non-linear Dynamic Systems	75	(14)
Piecewise Affine Model	75	(4)
Model Continuity and Domain Partitioning	79	(3)
Local Affine Model Identification	82	(3)
Multiple-Model Identification	85	(4)
Fuzzy Modelling and Identification	89	(23)
Fuzzy Multiple Inference Identification	90	(2)
Takagi-Sugeno Multiple-Model Paradigm	92	(3)
Fuzzy Clustering for Fuzzy Identification	95	(5)
Product Space Clustering and Fuzzy Model Identification	100	(3)
Non-linear Regression Problem and Black-Box Models	103	(4)
Fuzzy Model Identification From Clusters	107	(5)
Conclusion	112	(3)
Residual Generation, Fault Diagnosis and Identification	115	(42)
Introduction	115	(1)
Output Observers for Robust Residual Generation	116	(3)
Unknown Input Observer	119	(3)
UIO Mathematical Description	120	(2)
UIO Design Procedure	122	(1)
FDI Schemes Based on UIO and Output Observers	122	(5)
Sliding Mode Observers for FDI	127	(3)
Sliding Mode Observers	128	(2)
Kalman Filtering and FDI from Noisy Measurements	130	(1)
Residual Robustness to Disturbances	131	(10)
Disturbance Distribution Matrix Estimation	132	(1)
Additive Non-linear Disturbance and Noise	133	(1)
Model Complexity Reduction	133	(1)
Parameter Uncertainty	134	(1)
Distribution Matrix Low Rank Approximation	135	(1)
Model Estimation with Bounded Uncertainty	135	(1)
Disturbance Vector and Disturbance Matrix Estimation	136	(3)
Distribution Matrix Optimisation	139	(1)
Disturbance Distribution Matrix Identification	139	(2)
Residual Generation via Parameter Estimation	141	(1)
Residual Generation via Fuzzy Models	142	(1)
FDI Using Neural Networks	143	(4)
Neural Network Basics	145	(2)
Fault Diagnosis of an Industrial Plant at Different Operating Points Using Neural Networks	147	(3)
Operating Point Detection and Fault Diagnosis	147	(2)
FDI Method Development	149	(1)
Neuro-fuzzy in FDI	150	(6)
Methods of Neuro-fuzzy Integration	151	(1)
Neuro-fuzzy Networks	152	(2)
Residual Generation Using Neuro-fuzzy Models	154	(1)
Neuro-fuzzy-based Residual Evaluation	155	(1)
Summary	156	(1)
Fault Diagnosis Application Studies	157	(94)
Introduction	157	(1)
Physical Background and Modelling Aspects of an Industrial Gas Turbine	158	(10)
Gas Turbine Model Description	158	(10)
Identification and FDI of a Single Shaft Industrial Gas Turbine	168	(31)
System Identification	169	(7)
FDI Using Dynamic Observers	176	(7)
FDI Using Kalman Filters	183	(6)
Fuzzy System Identification and FDI	189	(2)
Sensor Fault Identification Using Neural Networks	191	(5)
Multiple Working Conditions FDI Using Neural Networks	196	(1)
FDI Method Development	196	(1)
Multiple Operating Point Simulation Results	197	(2)
Identification and FDI of Double Shaft Industrial Gas Turbine	199	(15)
Process Description	199	(2)
System Identification	201	(2)
FDI Using Unknown Input Observers	203	(5)
FDI Using Kalman Filters	208	(1)
Disturbance Decoupled Observers for Sensor FDI	209	(1)
Fuzzy Models for Fault Diagnosis	210	(4)
Modelling and FDI of a Turbine Prototype	214	(6)
System Modelling and Identification	215	(5)
Turbine FDI Using Output Observers	220	(22)
Case 1: Compressor Failure (Component Fault)	221	(2)
Case 2: Fault Diagnosis of the Output Sensor	223	(4)
Case 3: Turbine Damage (Turbine Component Fault)	227	(1)
Case 4: Actuator Fault (Controller Malfunctioning)	228	(5)
FDI in Noisy Environment Using Kalman Filters	233	(2)
Fault Isolation	235	(4)
Minimal Detectable Faults	239	(3)
FDI with Eigenstructure Assignment	242	(5)
Robust Fault Diagnosis of the Industrial Process	243	(4)
Robust Residual Generation Problem	247	(2)
Summary	249	(2)
Concluding Remarks	251	(10)
Suggestions for Future Work	253	(8)
Frequency Domain Residual Generation	253	(2)
Adaptive Residual Generators	255	(1)
Integration of Identification, FDI and Control	256	(1)
Fault Identification	256	(2)
Fault Diagnosis of Non--Linear Dynamic Systems	258	(3)
References	261	(18)
Index	279