Fault Detection,Estimation And Fault Tolerance Control Of Fuzzy Systems With Observable State

With the development of science and technology,the modern industrial sys-tems become much more complex.The requirement of control precision is getting higher and higher.At the same time,the system failures are inevitable because of the complexity of the modern equipment structure and high-power,high-load con-tinuous work.In many areas,failure of one component may result in a series of serious consequences.Fault detection,fault estimation and fault-tolerant control are effective methods to solve this problem.In addition,fuzzy system technique has provided an effective method to approximate a nonlinear system.In this method,a set of local models integrated with different degrees of validity are used to describe the nonlinear model.Thus,some technique of dealing with linear systems can be used to deal with problems in nonlinear systems.In recent decades,much attention has been paid to fuzzy systems.Under such a background,this dissertation consid-ered the problems of fault detection,estimation and fault-tolerant control for fuzzy systems.Firstly,based on the unknown input observer,address the problem of fault detection for switched fuzzy systems.Secondly,the dynamic unknown input observ-er is designed to estimate the sensor fault of switched fuzzy systems.Farther more,the problems of fault estimation and fault-tolerant control for the fuzzy system with actuator faults and sensor faults are considered.The main contributions of this dissertation are summarized as follows:(1)The problem of fault detection for switched fuzzy systems with unknown input is addressed by designing a unknown input fault detection observer.Based on the average dwell time and the piecewise Lyapunov function,the observer parameter calculation method is presented in terms of solutions to a set of linear matrix inequalities.The dimension of the designed observer can be flexibly selected within a certain range.The unknown input is decoupled from the residual.Both the weighted H? performance level and the weighted H_performance level are considered,which ensure the residual is robust to the disturbance and sensitive to the possible fault.(2)The problem of sensor fault estimation for switched fuzzy systems with un-known input is addressed.Firstly,the sensor faults are treated as a virtual state vector,then an augmented system can be obtained.Utilizing the defi-nition of left inverse matrix,the augmented system can be transform into a general system.A dynamic unknown input observer is designed to estimate the system states and sensor faults.The unknown input is decoupled from the estimation,error.Note that the traditional augmented descriptor observ-er cannot be used for the system with unknown input,since the unknown input decoupling condition cannot be satisfied.The observer design method proposed in this dissertation overcomes this problem.The proposed dynamic unknown input observer is more general than the traditional proportional ob-server,proportional integral observer and dynamic observer.These observers are only particular cases of our observer.(3)For the T-S fuzzy systems with actuator faults and sensor faults,the fuzzy-reduced-order robust fault estimation observer is designed.The observer can estimate the system states,actuator faults and sensor faults,simultaneously.The the influence of disturbances can be attenuated by H? performance.Com-pared with the existing methods,the proposed method can not only estimate system states and faults simultaneously,but also has a wider application range and a lower dimension.Using the information of estimation,an observer-based fault tolerant controller is designed.The controller can compensate the effect of the faults and guarantee the stability of closed-loop system.The observer and the controller are designed separately,which can reduce the computation complexity.(4)The problem of fault-tolerant anti-disturbance control for a class of fuzzy sys-tems with local nonlinear models is addressed.Base on the fuzzy fault esti-mation observer and disturbance observer,the systems states,actuator faults,sensor faults and modeling disturbances can be estimated,simultaneously.The H? performance is introduced to attenuate the effect of the unmodeled distur-bances.Furthermore,a fuzzy dynamic output feedback controller is designed,which can compensate the effect of the faults and disturbances.The regional pole constraint is considered,which can improve the transient performances of the error system and the closed-loop system.(5)For the nonlinear system with unknown nonlinear dynamic,the problem of fault-tolerant anti-disturbance control has been considered.Firstly,the gener-alized fuzzy hyperbolic model is used to approximate the unknown nonlinear function.Then a fuzzy adaptive observer and adaptive disturbance observer are designed simultaneously to estimate the system state,fault and distur-bance.In our method,the disturbance is assumed to be generated from an unknown exogenous system.Compared with the existing results,in which the disturbance is generated from a known exogenous system and assume all the states are measurable,the proposed adaptive disturbance observer is more ap-plicable in practice,since the exogenous system is independent of the controlled system and its parameter matrix is difficult to be measured.Furthermore,a nonlinear dynamic output feedback controller is designed,which contains a fault compensator and a disturbance compensator.For the case that system subject to mismatched input disturbance,the proposed disturbance local com-pensation technique can compensate the effect of a part of disturbance,and the H? performance is introduced to attenuate the effect the other part of disturbance.