Fault estimation filter design for discrete-time descriptor systems

This study deals with actuator fault estimation for discrete-time linear descriptor systems. The main contribution lies in the synthesis of a novel filter to estimate and isolate actuator faults for discrete linear descriptor systems. In this study, a restricted system equivalent (RSE) model is firstly obtained for the considered descriptor system, and then a fault estimation filter is designed based on the RSE representation. The gain matrix of the proposed filter has a special structure such that the residual vector will tend to the actuator fault with certain time delays which equal to the fault detectability indices. Moreover, this study proposes a robust filter design method to deal with the effect of the unknown disturbance. The proposed filter is able to detect and estimate the faults despite in the presence of disturbance. The design conditions for the proposed filter are formulated as a linear matrix inequality feasibility problem, which can be solved efficiently. Simulation results are given to illustrate the effectiveness of the proposed method.