Fault-Diagnosis Based on Neural Network in Fault-tolerant Control System
Fault-diagnosis plays an important role in the modem control system. A number of researchers have proposed fault diagnosis architecture for control system using the model-based analytical redundancy approach. However, in the real system the sufficiently accurate mathematical model is not generally available. This paper describes a fault-diagnosis system method using artificial neural network. The neural network replaces the analytical model. So it can be used in the complex system where analytical model are not or are hardly available. In this paper, a general architecture of fault diagnosis using neural network is presented. The neural-network approach for residual generation and residual evaluation is introduced as well.
Fault-diagnosis Neural network Residual generation Residual evaluation
Ma Linli Sun Yao
Automation college, Harbin Engineering University
国际会议
北京
英文
377-380
2007-11-04（万方平台首次上网日期，不代表论文的发表时间）