×

Robust model-based fault diagnosis for dynamic systems. (English) Zbl 0920.93001

The Kluwer International Series on Asian Studies in Computer and Information Science (ASIS). 3. Boston: Kluwer Academic Publishers. xviii, 356 p. Dfl 330.00; $ 160.00; £104.00 (1999).
The main theme of this book is to develop robust residual generation strategies for model-based fault diagnosis of dynamic uncertain systems. In analytical redundancy schemes, the resulting difference generated from the consistency checking of different variables is called a residual signal. The residual should be zero-valued when the system is normal, and should diverge from zero when a fault occurs in the system.
To detect and isolate faults in a dynamic system, based on the use of an analytical model, a declarative or residual signal must be used, which is derived from a combination of real measurements and estimates (generated by the model). The robustness problem can be solved by defining the independent sensitivities of the residual to uncertainties and faults. A robust Fault Detection and Isolation (FDI) scheme is characterized by a residual which is insensitive to uncertainties, but sensitive to faults. The aim of robust design of the FDI scheme is to reduce the effects of uncertainties on the residuals, and (or) enhance the effects of faults acting on the residuals. The success of fault diagnosis depends on the quality of the residuals. A preliminary requirement on residuals for successful diagnosis is the robustness with respect to modeling uncertainty.
After an introductory Chapter 1, Chapter 2 reviews the state of the art of model-based fault diagnosis techniques. The fault diagnosis problem is formalized in a uniform framework by presenting the mathematical descriptions and definitions. The fundamental issue of model-based methods is the generation of residual signals using the mathematical model of the monitored system. Chapter 3 introduces the approach to robust residual generation with the aid of the unknown input observer (UIO). The principle of the UIO is to make the state estimation error decoupled from the disturbance. Since the residual is defined as the weighted output estimation error, the residual is also decoupled from disturbances. Chapter 4 focusses on the disturbance decoupled residual generator design eigenstructure assignment. Chapter 5 demonstrates how to apply the disturbance decoupled method to a system with modeling uncertainty. Chapter 6 introduces an approach to the design of optimal residuals for detecting incipient faults, based on multi-objective optimization and genetic algorithms. Chapter 7 discusses the robust residual generation using optimally robust parity relations. Chapter 8 introduces frequency domain design and \(H_\infty\) optimization methods for robust fault diagnosis. Finally, the last Chapter 9 studies the fault detection and isolation problems for nonlinear dynamic systems.

MSC:

93-02 Research exposition (monographs, survey articles) pertaining to systems and control theory
93B07 Observability
90B25 Reliability, availability, maintenance, inspection in operations research
93B35 Sensitivity (robustness)

Software:

LMI toolbox
PDFBibTeX XMLCite