Fault Detection And Estimation Of Markovian Jump Systems With General Conditions Of Mode Signals

With the innovation and development of production technology,the control systems become increasing complexity,which in turn require higher efficient and reliable control performance.Thus more experts and scholars have paid attention to the systems,which are vulnerable to environmental changes.In the past few decades,various research techniques have emerged for different problems in such systems.The stochastic Markovian jump systems have been proved to be very suitable to model systems whose structures are susceptible to complex factors.As it is a hybrid dynamic system with special form,its running state takes continuous values,while the other state(i.e.mode)takes discrete values in a finite set.The switching among operation modes is subordinated to the Markovian switched rule.Meanwhile,due to the demand of safety and efficiency,it is important for systems to detect and estimate faults timely and accurately in modern industrial production.Based on Markovian jump systems theory and linear matrix inequality(LMI)method,this paper studies the design of estimator and filter,the following research results are obtained:(1)For a class of continuous-time Markovian jump systems,the problem of fault detection under general conditions of mode signal is studied.Different from the traditional methods,a partially mode-available and unmatched filter is presented and used as a residual generator.The fault detection and isolation(FDI)problem is converted into an H_?filtering problem by minimizing the error.Sufficient conditions for the existence of an FDI filter are derived in terms of LMIs.Numerical examples are exploited to demonstrate the effectiveness and superiority of the proposed methods.(2)This chapter focus on the fault estimation problem for a class of nonlinear systems with Lipschitz nonlinearities.A new fault estimation approach is proposed,and a random intermediate variable is constructed to make the model representation more general.The established intermediate estimator could estimate the state and fault simultaneously.The proposed algorithm could bear stochastic failure,which also has less conservatism.Finally,a simulation example is given to verify the practicality and superiority of the proposed method.(3)This chapter investigates the fault estimation problem for Markovian jump systems in the case of mode-dependent and mode-independent respectively.Firstly,a random intermediate variable depending on operation modes is introduced.Meanwhile,another is constructed by a series of mode-dependent parameters and its probability distribution.Different from the former,it is independent of system mode and does not need available online.Compared with existing fault estimation methods,the proposed algorithms could bear stochastic failure and estimate the system state and fault simultaneously.Two practical examples are exploited to demonstrate the effectiveness and superiority of the proposed methods.