Fault Diagnosis of Sampled Data Systems

This thesis is concerned with fault detection of sampled data systems with special emphasis on the time varying characteristics of the system. A discretization method based on the norm invariant transformation is proposed for time invariant systems with aperiodic sampler and time varying systems with periodic sampler. The outcome of this transformation for both cases is a discrete time varying system.;The problem of fault detection for discrete time varying systems is studied and the proposed approach along with the norm invariant transformation is used for fault detection of sampled data system with time varying characteristics. Also, because of the time varying nature of Kalman filter, its application as an observer for fault detection of discrete time varying systems is investigated.;Finally, the fault detection method with norm invariant transformation is used for fault diagnosis of network control system with network induced delay.