Fault Diagnosis for Dynamic Systems withParametric Uncertainties: Robustness against Parametric Uncertainties

This book presents a methodology for fault diagnosis - detection, isolation, and identification - fordynamic systems whose models are faced with parametric uncertainties. This class of problems is oftenencountered in practice, and has not beenexclusively dealt with in the past. The book beginswith areview of existing literature on model-based, inparticular, observer-based faultdiagnosis for linear and nonlinear dynamic systems.It thenfocuses on a novel observer-based fault diagnosistechnique for linear time-invariant systems withparametric uncertainties. The underlyingobserver-design utilizes recentresults in the area of robust control of dynamicsystems with parametric uncertainties, andis grounded in Kharitonov's theorem of stability ofInterval polynomials. This paradigm of faultdiagnosis is then extended to nonlinear dynamicsystems with parametric uncertainties. The remainingpart of the book extends the proposed model-basedfault diagnosis methodology to empirical state-spacemodels derived from input-output data using subspacemodel identification.