Robust Fault Detection For Large Scale Systems Under Network Environment

With the rapid development of computer and networked communications technology, and continuing to improve the control system requirements, traditional control system had been developing to the open distributed network control system (Networked Control Systems, NCS).In practice, many process control system is composed of a series of interrelated network interconnected systems composed of several subsystems affect each other, the data transmission network, sensors temporary failure and other reasons may cause the measurement and control signal data lost. The introduction of network and the complexity of the large-scale systems would greatly reduce the performance of the overall system robustness and the traditional fault detection algorithm, it is necessary to reassess and establish NCS large-scale systems fault detection theory.Now the research about network interconnected systems fault detection problem is still rare. Therefore, this problem about robust fault detection of a class of large-scale network systems were studied. The main work is as follows:(1) Research robust fault detection problem of large-scale system that does not contain data loss. For a class of large-scale systems consisting of N linear discrete subsystems with external interference, design fault detection filter. Based on Lyapunov stability theory, making the residual error system mean square exponentially stable and satisfy specified H∞performance index. Get The necessary and sufficient conditions for the existence of fault detection filter by linear matrix inequalities, and get the filter parameters matrix by solving linear matrix inequalities. Finally, apply the LMI Toolbox of Matlab to solve and simulink the example, the Matlab simulation results show that the algorithm is effective and feasible.(2) Research robust fault detection problem of large-scale system that contain measurement data loss under network environment. For a class of large-scale networked systems consisting of N linear discrete subsystems with measurement data loss, and measurement data loss assumptions to meet the known probability Bernoulli release, design fault detection filter. Based on Lyapunov stability theory, making the residual error system mean square exponentially stable and satisfy specified H∞performance index. Get the necessary and sufficient conditions for the existence of fault detection filter by linear matrix inequalities, and get the filter parameters matrix by solving linear matrix inequalities. Finally, apply the LMI Toolbox of Matlab to solve and simulink the example, the Matlab simulation results show that the algorithm is effective and feasible.