Research Fault Diagnosis And Fault-tolerant Control For Networked Control Systems

Networked Control Systems, because of its low cost, reliable and flexible, facilitate fault diagnosis and other advantages, which is widely used in aerospace, manufacturing, electrical, artificial intelligence, and many other systems. But owing to the existence of the problem including time delay and lost package, compared with traditional, Networked Control Systems are more complex, more likely to lead to control the performance of the system down. Once the system failure, it may cause a huge loss that it is difficult to estimate, so that Networked Control System has become an urgent problem to be solved.In this paper, the networked control system fault detection and fault-tolerant control are studied. The main researches are concluded as follows:1) For a class of uncertain time-delay of nonlinear network control system, when the existing in the system of nonlinear uncertainties and external disturbance, in failure situation, we design fuzzy sliding-mode observer, based on the traditional observer added a feed-forward compensation, making the error of the system can still is asymptotically stable when the system in failure cases. By using the method of integral variable defined, sensor fault will be transformed into actuator fault diagnosis, based on the. Lyapunov stability theory to prove the design of stability, and gets the sufficient condition of the stability, through the linear matrix inequality (LMI) complete the design of the observer.2) The research with time delays and parameter uncertainties, existence of networked control system disturbances of robust fault-tolerant control problem of T-S model based on the research of networked control systems with blur. By the Lyapunov function stability theory and linear matrix inequality (LMI) method, and deduced can keep the asymptotic stability and at the same time satisfy the H∞performance index to the sufficient conditions. With numerical examples show that the feasibility of the design method.3) Based on RBF Neural Network of uniform approximation of the nature of the arbitrary nonlinear continuous function, we structure T-S model based on neural network adaptive observer, to estimate the fault of amplitude. At the same time, we propose fault adjustment based on the Lyapunov stability theorem of the algorithm. From the simulation results we can find it is apparently effective that the methods in fault estimate and fault-tolerant control.Finally, the results of the thesis are summarized and the further research problems are pointed out.