Study On Robust Diagnosis And Fault-tolerant Control Technology For Dynamical Systems

The increasing demand for reliability, safety and efficiency has led to unprecedented challenges for modern complex engineering systems'operation. These conditions significantly increase the possibility of the faults in components or subsystems. Potential faults may cause the whole system fails and even serious losses. Due to the uncertainties of complex plant models and the external disturbance, it is difficult to satisfy the performances above. Therefore, the robust fault diagnosis.and fault-tolerant control (FTC) for dynamic systems has obtained greater attention and it is important to ensure the complex systems'reliability and safety.In this thesis, the problems of robust fault diagnosis and FTC for several types of dynamic systems are studied by using observer approach and adaptive technique. The main research results are given as follows:The problem of fault estimation algorithm and the sufficient condition of diagnosis system's stability for nonlinear dynamic systems is considered by using fault detection observer and adaptive fault diagnosis observer. Then, a flexible joint robotic example is given to illustrate the efficiency of the estimate approach.Robust guaranteed cost fault tolerant controller design method is proposed for fuzzy singular system with norm-bounded parameter uncertainties and time-varying delay. Using continuous fault model, global fuzzy state feedback controller is designed for the part failures of actuator and the upper bounded of cost function is obtained. The numerical example is provided to illustrate the design method.The problem of fault estimation and fault-tolerant control is researched for a class of T-S fuzzy time-delay systems with sensor faults. Using fuzzy descriptor system method, a fuzzy observer is constructed for the augmented fuzzy descriptor systems, which can estimate the state vector and sensor faults simultaneously. Based on the estimated sensor faults, a fault tolerant control scheme is proposed. Truck trailer systems illustrate the effectiveness of the given fault compensation method.For vehicle active suspension systems in finite frequency domain, fault estimation and fault-tolerant control problem is discussed. By designing an observer-based fault estimator, a H∞controller in generalized internal model control architecture is obtained. Using Gnenralized-Kalman-Yakubovich-Popov (GKYP) Lemma, sufficient conditions ensuring simultaneous finite frequency estimation and control performance are derived in terms of the LMIs and fault compensation is also proposed. Simulation results show the fesibility and efficiency of the given conclusions.