Research On Fault Detection For Delta Operator Systems With Packet Dropouts And Limited Communication

With the development of the communication technology and the size of control systems becoming larger,the digital network is applied to transfer information between different nodes of the control systems and the network control systems(NCSs)emerge at this moment.Different from the traditional point-to–point control systems,NCSs possess many advantages such as reduction of system wring,simple maintenance and shared information.However,the network induced issues such as time-delay,packet dropouts and communication constraint,will degrade the performance and even make the NCSs unstable.Because the NCSs are widely used in people's life,the sudden appeared faults will inevitably affect the property of NCSs and even lead to serious consequences.It is particularly important to detect the faults and give an alarm as soon as possible.The digital network which is used as the transmission medium makes it essential to convert the continuous-systems into the discrete-model in the research.Most of the available literatures used the shift operator approach to discretize continuous-systems.However,the shift operator cannot make the discrete-systems approach to the corresponding continuous-systems at high sampling rate and even makes the crowding poles around the boundary of the stability circle.While the discretization method of delta operator can avoid the disadvantages and makes the discrete-model and the continuous-model into a formulation.This thesis focuses on the research of fault detection for delta operator systems with packet losses and limited communication.The innovation points and major contributions of the thesis are as follows:Firstly,the fault detection(FD)observer is designed for delta operator systems with packet dropouts and limited communication.The systems are modeled as Markov jump systems with packet dropouts obeying Markov chain.By using the stochastic stability theory and linear matrix inequalities(LMIs)method,the filter gain matrix can be achieved which can ensure that the systems satisfy the stochastic stability and H_? performance.Besides,the simulation proves the effectiveness of the designed method.Then,based on the former research,the thesis proposes the time-delay delta operator formulated model with data dropouts and limited communication existing in both the sensor to the controller link and the controller to the actuator link.Then,the FD filter is addressed to detect the faults.According to the Lyapunov functional technology in delta domain and the LMIs approach,sufficient conditions which can make the systems meet the asymptotic stability and the H_? performance are gained.The parameters are gotten by the optimization issue of LIMs toolbox.Lastly,the thesis studies the robust FD problem of the NCSs with time-delay,packet dropouts and limited communication.The considered networked control systems are modeled as parameter uncertain delta operator systems with stochastic packet dropout meeting Bernoulli distribution.By using the multi-Lyapunov functional and LMIs approach,sufficient conditions for the existence of the designed fault detection filter and its explicit expression are obtained.Furthermore,an improved evaluation function and threshold are given to detect the faults.The numerical simulation shows that the presented algorithm has the robustness to disturbances.