A FAULT DETECTION DIAGNOSIS PREDICT OBSERVER BASED ON RESOURCE ALLOCATION NETWORK

Qizhi Wang and Xiaoxia Wang

Keywords

Fault diagnosis, Lyapunov stability, RAN neural network, network control System, fault observer, LMI

Abstract

A fault observer based on RAN (resource allocation network) neural network prediction is proposed for fault detection in a control system with network delay. Firstly, the robust controller is designed to ensure the stability of the system under uncertainty parameters and external disturbances. Then, an observer-based robust control method is used to construct a fault observer. Thus the residual of NCS (network control system) is generated, and the controller and observer gains are obtained by Lyapunov stability criterion and robust toolbox LMI (linear matrix inequality) method. Finally, the effectiveness of the algorithm is verified by simulation. By using RAN prediction function, the sampling values of NCS sensors are predicted and a fault detection system based on prediction is designed. The stability of the observer error system is also proved to be effective by Lyapunov theorem.

Important Links:



Go Back