Research On Fault-tolerant Methods For Networked Congtrol System

With control system scale growing and development of computer network technology, networked control system (NCS) has been gradually applied to control projects. In NCS, measure and control signals are sent by network. For a lot of advantages, as modular structure, low cost and so on, NCS has got much more attention and great development. However, problems, such as packet loss and network-induced delay, have not been solved. These problems will badly interfere NCS control performance and stability.In the dissertation, the concept and structure of NCS are introduced; the types and characteristics of computer networks are summarized; NCS main problems and some solutions are described; NCS exponential stability and patient packet loss rate are researched.In NCS, the actuator nodes and control nodes are to be driven by events and sensor nodes by time. Based on the dynamic system theory, a model is established for the determinate sample period dynamic system with success and failure events.NCS stability is analyzed by using the Lyapunov theory and schur complement lemma. The sufficient conditions for NCS stability is given based on the linear matrix inequality.Bernoulli distribution packet loss process is considered as Markov process. The packet loss events are taken into possible states in the Markov process.Based on the schur complement lemma and Markov chain mean-square stability criterion, the NCS stability is discussed, and the sufficient stability conditions is given for it.Simulations are done for the asynchronous dynamic NCS by TrueTime toolbox of the Matlab. The simulation results show that a system is still stable if the packet loss rate is in a certain range.