Robust Fault Detection For Networked Systems Based On Event-triggered Mechanism

Networked control system is a closed-loop control system formed by sharing communication network between controlled object and other components.Compared with the traditional control systems,networked control system has the advantages of simple connection,high flexibility,easy expansion and convenient maintenance.However,the introduction of the communication network also inevitably leads to a series of new questions,like transmission delays,data packet dropout,quantization errors,bandwidth limit and so on,which will degrade system performance and even cause system instability.Fault detection is to judge whether the system fails,which is an important basis for the establishment of the early warning mechanism of system security operation.Due to the introduction of communication network,the conventional fault detection approach is difficult to be directly applied into networked control system.Therefore,for non-ideal network factors,it is of great theoretical significance and application value to study the fault detection method of networked systems.In view of the current researches for fault detection of networked systems,their focus is on cycle-triggered discrete-time systems.This paper will study the fault detection of continuous-time networked systems based on event-triggered mechanism.Compared with the conventional periodic communication scheme,the event trigger communication mechanism used in this paper can realize the on-demand sending of packets.This can greatly reduce network load and save network resources.The main research work of this thesis can be summarized as follows:An effective transmission strategy based on event-triggered mechanism is introduced to investigate the fault detection problem for networked invariant systems with limited communication bandwidth and sensor saturation.Firstly,Lyapunov stability theory is used to analyze the stability of filtering error system,and sufficient conditions for the existence of robust fault detection filters can be obtained by means of linear matrix inequality.Next,the convex optimization method is adopted to obtain the parameters of target fault detection filters.Then,based on the designed fault detection filter,a residual signal is generated to determine whether the system has failed.Finally,a numerical simulation example is provided to illustrate the effectiveness of the fault detection method.The fault detection problem of a class of networked Markov jump systems is studied for the abrupt change of system parameters caused by abrasion of system components or sudden change of environment in the process of industrial application.The change of system parameters is subject to the continuous-time Markov chain,and the optimal fault detection filter parameters are obtained by using stochastic Lyapunov stability theory and linear matrix inequality method in the filter design process.Then,the designed fault detection filter is taken as the residual generator to detect the system faults by residual signal.Finally,examples of numerical simulation and vertical take-off and landing helicopter system are used to verify the effectiveness and practicality of fault detection method,respectively.The fault detection problem for a class of networked LPV systems is investigated by establishing a linear parameter varying(LPV)model to describe the case for linearization of nonlinear systems along parameter trajectories.Due to the state space of LPV system is infinite dimension,the parametric linear matrix inequality is solved by using the base function and grid technology with LPV Toolbox in the design process of fault detection filter.Based on this,the infinite dimensional problem can be transformed into a finite dimensional problem and the sufficient conditions for the existence of the fault detection filter are derived.Then,the faults of the system are detected by threshold logic method.Finally,a simulation example is given to proof the validity of fault detection method through GE-90 turbofan engine system for Boeing 777.