Fault Diagnosis And Fault Tolerant Control For Non-Gaussian Linear Singular Time-delayed Stochastic Distribution Systems

In recent years, with the continuous improvement of automation, industrial engineering systems are becoming more and more complex, and regular maintenance measures can hardly meet the urgent need for safety and reliability in industrial production. How to diagnose fault in time and make the post-fault system still work well has raised people’s concerns. Thus, the research on fault diagnosis(FD) and fault tolerance control(FTC) has drawn engineers and scientific researchers’ attention.Stochastic distribution control(SDC) system is such a kind of stochastic systems, where the system output is the probability density function(PDF) of the output variable. The SDC system widely exists in practise industrial process such as papermaking process, grinding process and flaming process. The most widely research on stochastic distribution system is the output PDF tracking control, which means making the output PDF track the given PDF by designing a suitable controller.Because of the existence of belt material transporting, systems modelling, data processing and transmission, time-delay widely exists in various industrial control systems. The existance of time delay will have great impact on the system performance. Besides, time delay will extremely affect the effectiveness of FD and FTC of dynamic systems. Currently, few literatures reported the FD and FTC of the non-Gaussian singular time-delayed SDC systems.Against the background of grinding process, in this thesis the FD and FTC for non-Gaussian linear singular time-delayed SDC systems will be deeply studied.The specific contents are as follows:(1) For the non-Gaussian linear singular time-delayed SDC systems where the output PDF is approximated by the linear B-spline basis function. A fault diagnosis approach based on the adaptive observer is established to diagnose the size of fault. The parameter vector of fault diagnosis adaptive tuning rule and the observer gain can be obtained by solving linear matrix inequality(LMI). Using the estimated fault information and other measurable information, the fault tolerant controller based on proportional integral(PI) tracking control is designed so as to make the post-fault PDF still track the given PDF. The computer simulation is provided to demonstrate the effectiveness of the proposed approach.(2)For the non-Gaussian linear singular time-delay stochastic distribution system where the square root B-spline model is used to approximate the output PDF, and the PDF approximation error is taken into consideration. A fault diagnosis approach based on the adaptive observer is constructed to diagnose the size of fault. The fault tolerant tracking controller is constructed to make the distribution tracking error at each time instant satisfy a certain upper bound beyond a limited time. The Lyapunov stability theory is used to analyse the stability of the observation error system and the tracking error system. The gain matrices in fault estimation and fault tolerant tracking control can be obtained by solving the corresponding LMI. Finally, the Matlab simulation results have confirmed the FD and FTC algorithms.(3)For the non-Gaussian linear singular time-delayed SDC system with unknown external disturbance where the output PDF is approximated by the rational square-root B-spline basis function, a robust fault diagnosis algorithm based on the adaptive observer is presented. A full-order observer is constructed to estimate the exogenous disturbance. An adaptive observer is used to estimate the fault information. A fault tolerant tracking controller is designed to let the distribution tracking error at each time instant satisfy a certain upper bound. Finally, a simulation example is included to illustrate the effectiveness of the proposed algorithms and encouraging results have been obtained.