Research On Event-Triggered Fault Detection For Stochastic Systems In Network Environment

With the development of control system integration and complexity,the performance has also been a qualitative leap.However,the safety and reliability of the control system are also facing great challenges,which promotes the development of fault detection of control system.In distributed control systems,the controlled object and the system components distributed in different areas are linked by using network communication technology to build a network control system.The data transmission in the real network is often limited to a certain extent.Therefore,it may cause delay or packet dropout and other problems when data is transmitted,the difficulty of analysis of the control system is greatly increased.The event-triggered fault detection of stochastic systems with limited network communication is discussed in this paper,and takes network induced delay and packed dropout into account in fault detection system.Firstly,for linear stochastic systems with limited communication resources in the network environment,fault detection filter is established to detect system fault when packet dropout is considered.An event trigger with fixed threshold is designed at the output of the sensor to reduce the communication pressue.The Bernoulli distribution is adopted to simulate the possible packet dropout.A weighted fault model is devised to advance the speed and accuracy of fault detection.Based on the above consideration and analysis,the model of fault detection and stability condition can be established.According to Lyapunov stability theory and LMI,the criteria that meet the requirements of fault detection system stability and fault detection are derived.And the design scheme of filter parameters is given.By solving LMI,then calculate the relevant parameters of fault detection system.Secondly,for linear stochastic with limited communication resources in the network environment,the filter and controller are established simultaneously.A two terminal eventtriggered mechanism is established.The method of setting event-triggered devices at the output side and the control terminal respectively is used to further alleviate the problem of network bandwidth constraints.Considering the influence of the two terminal asynchronous event-triggered mechanism on the fault detection effect and control performance,the collaborative design of filter and controller is completed.Based on Lyapunov stability theory and LMI,the stability criteron of fault detection system is derived,and the fault detection index and control performance are determined.Then the relevant parameters of the system are calculated.Finally,aiming at fault detection for nonlinear continuous stochastic systems under the condition of limited network resource,packet dropout and time delay based on adaptive event trigger is proposed.Bernoulli distribution is used to simulate the possible packet dropout problem and the probability of the nonlinear term.In the communication channel,an adaptive event-triggered mechanism is designed to further alleviate the network communication pressure.The nonlinear term is constrained by using the quadratic constraint term.Based on the comprehensive consideration of the above problems,a fault detection model is constructed,and the stability and fault detection requirements are satisfied.Then the design scheme of filter parameters is given,and the relevant parameters of fault detection system are calculated by solving LMI.