Research On Fault Diagnosis And Fault Tolerant Control Based On Fuzzy Observer

Fuzzy observer based fault diagnosis and fault accommodation problems for a kind of nonlinear systems are studied in this dissertation. The proposed methods are simulated by Matlab software.At First, the history and the development of fault detection and isolation (FDI) and fault tolerant control (FTC) are surveyed. Also, fuzzy control theory, the adaptive control technique and neural network are introduced.The nonlinear system is modeled by T-S fuzzy model. A fuzzy detective observer and a fuzzy adaptive diagnostic observer are designed respectively. The former is used to detect the fault and based on the latter, a fault estimate scheme is proposed. Then, the fault estimation is used to reconfigure the control law to recover the system performance. The efficiency of the algorithm is demonstrated by an invert pendulum example.Furthermore, a fault detection method of T-S model with uncertainties is presented. A robust observer is designed for a class of uncertain nonlinear systems represented by T-S model, which is used to detect the system fault. Stability conditions of such observer are expressed in terms of linear matrix inequalities (LMIs). The simulation result validates the proposed observer has good ability against uncertainty and can detect the fault in time.Finally, a fault accommodation approach based on neural network fuzzy observer for nonlinear systems is presented. The neural network is constructed to approximate the fault on-line. The fuzzy observer can estimate the shape of the fault, which can be used to reconfigure the controller. The Lyapunov theorem is used to prove the stability of the faulty system under the fault accommodation (FA) scheme. An F-16 example illustrates the applicability of the given method.