Observer-based Nonlinear System Fault Diagnosis

Stimulated by the growing demand for improving the reliability in modern industrial systems, fault diagnosis technique for dynamic systems has gotten great stride forward during the last decades. But most of fault diagnosis methods are based on linear system thoery, these methods can only be used in linear systems. Since practical engineering systems are mostly nonlinear in nature, and nonlinear feature is more obvious when system has faults, so it is essential to research fault diagnosis method for nonlinear dynamic systems.The speed of fault isolation and identification plays a decisive role in prevent fault from accident, but this issue has been paid few attention in the domain of fault diagnosis.Parameter interval based method is a new fault diagnosis method which can isolate fault fastly and can give the size of fault. Its thoery foundation is adaptive observer based fault diagnosis approach. So this paper did the research mainly on both adaptive observer based method and parameter interval based method. Main contribution of this paper:(1) Has studied fault isolation problem for process dynamic fault of parameter interval based method and extended the method to sensor fault and actuator fault problems.(2) Has built the simulation systems for parameter interval based fault isolation method and adaptive observer based method with robot system and other models. Using these models, we has shown the validity of parameter interval based fault diagnosis method on different models, and provided more simulation experiment circumstances for further research of this new method.