Bookstore

This book presents a methodology for fault diagnosis - detection, isolation, and identification - for dynamic systems whose models are faced with parametric uncertainties. This class of problems is often encountered in practice, and has not been exclusively dealt with in the past. The book begins with a review of existing literature on model-based, in particular, observer-based fault diagnosis for linear and nonlinear dynamic systems. It then focuses on a novel observer-based fault diagnosis technique for linear time-invariant systems with parametric uncertainties. The underlying observer-design utilizes recent results in the area of robust control of dynamic systems with parametric uncertainties, and is grounded in Kharitonov's theorem of stability of Interval polynomials. This paradigm of fault diagnosis is then extended to nonlinear dynamic systems with parametric uncertainties. The remaining part of the book extends the proposed model-based fault diagnosis methodology to empirical state-space models derived from input-output data using subspace model identification.

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