Fault Detection Of Network Control System With Time Delay Based On The GA-elman Neural Network Prediction

Network control system(NCS) is a closed-loop control system that transmission information through a data communication network. With the widely application of the NCS, its fault detection can improve the security of the system. In this paper, considering a class of NCS with time delay that it maybe longer than one period, and assuming that the system has output delay between sensor and controller,then fault detection has carried out. The main work content is as follows:First, study the working mechanism of control system, and analysis the influence of time delay of the NCS. Then, set up a long time delay NCS fault detection system based on neural network prediction. Neural network is a new information system to simulate the human brain biological nervous system. It has good ability of self-learning, self-organization and very strong nonlinear mapping. it can be used in the NCS to predict sampling values of the control system to eliminate the time delay. But the BP neural network and Elman neural network has its own disadvantages, it is easy to fall into local minimal and the global search ability is poor.Therefore, in this paper, on the basis of in-depth study the genetic algorithm, GA- Elman neural network has been put forward, it combines the local search ability of neural network and the global search ability of genetic algorithm. It use genetic algorithm to optimize the neural network weights threshold to avoid neural network into a local minimal and speed up the convergence speed of the neural network.Compared with the other two kinds of neural network, GA- Elman neural network has good prediction precision to sampling values and less convergence steps.Between the sensor and controller add a GA- Elman neural network forecast module, according to the predicted value design a observer, and use the Lyapunov stability theory to prove the uniform stability of the closed-loop system. The simulation results shows that the observation system based on neural network prediction is a good tracking system, when fault occurs,it can timely detect fault. GA- Elman can effectively reduce the influence of time delay of the system.