Research On Fault Tolerant Control Of Network Control System Based On ELM

In recent years,with the continuous development of control engineering,the consequences of engineering system faults have become increasingly serious.To strengthen the stability of engineering systems,scholars have gradually increased the research on fault diagnosis and fault-tolerant control.In the fault diagnosis,the system poles are used as fault diagnosis feature information,and the fault can be diagnosed by the method of pole configuration.Besides,the extreme learning machine(ELM)algorithm is widely used in the field of fault diagnosis because of its fast-computing speed and generalization performance.Reliable control systems can perform fault-tolerant control on the system engineering so that the system engineering can still meet certain performance standards after failure.However,there are still many areas that need to be improved.Therefore,research in this field has essential theoretical value and practical significance.Based on the methods of machine learning,fault diagnosis,and fault-tolerant control,this article made an in-depth study on the single-channel failure problem of the system.The fault diagnosis system and reliable controller designed in this paper can handle system faults.The main content of this article consists of the following three parts:1.Aiming at the failure of any single channel of the system,this paper proposes a fault diagnosis system that uses the poles of the closed-loop system as fault information to design a fault diagnosis system and realize fault-tolerant control.When any single channel of the actuator fails or the fault gain of different channels changes,the fault is judged by the change of the pole in a given area.The poles of the closed-loop system are used as the input of the ELM fault diagnosis model to realize fault-tolerant control of different channels of the system.Through the network control system(NCSs)simulation,verify the accuracy of the fault-tolerant control system and the effectiveness of the reliable controller.2.The optimization performance of the particle swarm optimization(PSO)is studied,and the parameters in the ELM are optimized to improve the pole learning rate of the ELM to achieve a better fault diagnosis effect.At the same time,a pole observer is designed to collect the poles of the system in time.It is verified by an example of the network control system that the accuracy of the PSO-ELM fault diagnosis system has been greatly improved,and the pole error observed by the pole observer is very small.3.Research is based on the original system model,adding external disturbances,and designing output feedback fault-tolerant controllers to realize system fault-tolerant control.Also,through the combination of the improved PSO algorithm and genetic algorithm(GA),the parameters of the ELM algorithm are further optimized to achieve precise fault-tolerant control.And in the simulation results,the reliability and accuracy of the method are shown.