Research Of Observer-based Robust Fault Diagnosis And Fault Tolerant Control Technology

With the rapid development of system complexity and integration,fault diagnosis and fault tolerant control technologies have become attractive topics and have recently drawn considerable attention.Analytical redundancy-based fault diagnosis and fault tolerant control operate in the form of a software process model instead of the traditional hardware redundancy,which have the advantages of high efficiency and low cost.Fault diagnosis is divided into two parts,including fault detection and isolation(FDI)and fault estimation(FE).The former one is to decide whether faults have occurred in the system,and to determine when and where the faults occurred.Based on FDI,the latter one is to further estimate the amplitudes of the faults.Fault estimation plays a key role in the fault tolerant control system,which provides the real-time fault information for the additional control law.The complexity and high integration of the modern industrial systems lead to an increase in the possibility of system failures,which may change the system behavior and result in performance degradation or even instability,thus causing casualties.Therefore,the research on fault diagnosis and fault tolerant control is of great significance and challenge.The difficulties and deficiencies of the existing methods are the motivations of this study.This paper researches the observer-based robust fault diagnosis and fault tolerant control technology for uncertain linear and nonlinear systems respectively.The main works are as follows:Considering the linear time invariant(LTI)systems including modeling uncertainties and unknown inputs,a robust fault detection observer design is proposed.With the assumption of no modeling uncertainties and the definitions of the sensitivity to the fault and the robustness to the unknown inputs,a new iterative linear matrix inequality(LMI)algorithm is used to solve the design problem of the reference residual model as an H_?H_-optimization problem.Then combining with the reference residual model,the robust fault detection observer design can be described as a standard H_?model-matching problem.Simultaneously,the adaptive threshold generator is designed to decrease the rate of false alarming.Based on the above fault detection for uncertain LTI systems,the observer-based fault estimation problem is further studied.By constructing an augmented error system,the proposed algorithm not only relaxes the strict existence conditions of the conventional observer,but also separates the disturbance term including the model uncertainty.The improved LMI-based bounded real lemma is proposed to construct the augmented fault estimation observer(AFEO)to achieve the simultaneous estimation of system states and faults.After fault diagnosis,we extend the research to the adaptive observer-based fault estimation and accommodation.In order to improve robustness,rapidity and accuracy of traditional fault estimation algorithm,by utilizing a new fault estimator model,an improved adaptive fault estimation observer(IAFEO)based on the robust H_?theory and generalized inverse matrix theory is proposed to increase the performance.The introduction of the convergence rate helps to choose the uncertain parameter in the fault estimator model.Furthermore,an observer-based state feedback fault-tolerant control strategy is designed,which guarantees the stability and performance of the faulty system.Finally,the study on the observer-based fault estimation and fault tolerant control is extended to the cases of nonlinear systems and discrete-time systems.By introducing a nonsingular coordinate transformation,a new nonlinear reduced-order fault estimation observer(RFEO)for a class of uncertain nonlinear discrete-time systems is proposed with a wide application range in order to achieve an accurate estimation of both states and faults.Next,an improved algorithm is given to obtain the optimal estimation by using a novel iterative LMI technique.Furthermore,an RFEO-based output feedback fault-tolerant controller(ROFFTC),which is independent of the RFEO,can maintain the stability and performance of the faulty system.Note that,the fault estimation observer and the fault tolerant controller are independently designed,which is convenient when calculating parameters according to their respective properties.