Research On Observer-based Fault Detection And Fault Tolerant Control For Complex Dynamic Systems

In the practical engineering application,for some communication systems,circuit systems and large industrial complex dynamic systems,the faults will seriously affect the safety of the systems.In order to enhance the stability of the dynamic systems,it is essential to detect faults in time and take the necessary measures to overcome its effects.However,when the faults occur,the controllers designed by traditional methods can not guarantee the systems to achieve the desired excellent performance and even affect the stability of the systems.The methods of fault detection and fault-tolerant control can solve these problems effectively.To improve the fault tolerance of the dynamic systems,based on the Lyapunov stability criterion and using MATLAB tools,this thesis further studies the fault detection of the observer-based continuous/discrete-time conic-type nonlinear systems and the fault tolerant control of the Markov jump systems,the effectiveness of the theoretical results is confirmed by relevant examples.The main contributions of this thesis are summarized as follows:Chapter 2 analyses the problem of fault detection for a class of continuous-time conic-type nonlinear systems within the finite frequency domain.Employing the Takagi-Sugeno(T-S)fuzzy method,the conic-type dynamic error model is established and the appropriate residual signal is selected.To verify that the chosen residual signal is robust to external disturbances and sensitive to faults,we propose some sufficient conditions to achieve H?fault sensitivity performance and H_?disturbance attenuation performance.Then,the fault detection observer design issue is transformed into an optimization algorithm by utilizing linear matrix inequalities(LMIs)techniques.Subsequently,a simulation example concerned with a tunnel diode-circuit is provided to guarantee the effectiveness of the devised approach.Chapter 3 focuses on the issue of fuzzy fault detection filter for a class of discrete-time conic-type nonlinear systems.The Lyapunov function approach is applied to obtain the sufficient condition for the asymptotic stability of the conic dynamic error system,and the appropriate filter parameters are designed.Later,by using LMIs techniques,the fuzzy fault detection filter design problem is transformed into an optimization algorithm.Furthermore,the simulation results illustrate that the designed fuzzy fault detection filter is efficient for detecting faults.Chapter 4 studies the fault-tolerant control scheme for discrete-time Markov jump systems with sensor faults and actuator faults,respectively.The intention is to devise observers to estimate faults and states of the discrete-time Markov jump systems.To reduce the impact of the fault on the systems,the extension functional observers are first introduced.Considering that extension functional observers include unmeasurable information,the actual observers are established.Resorting to the estimated information from the actual observers,the fault-tolerant controller is devised to guarantee that the discrete-time Markov jump systems is stochastically stable.A simulation example via a DC motor device is employed to prove the correctness of the studied approach.Finally,the theoretical results of this thesis are summarized,and the next research direction is discussed.