火电厂生产过程故障诊断神经网络专家系统  

Fault Diagnosis Based on Neural Network and Expert System for a Thermal Power Plant

作　　者：姚志红[1] 陈铁军 冯冬青 

机构地区：[1]上海交通大学信息与控制工程系,200030 [2]郑州工学院

出　　处：《控制与决策》1997年第3期252-255,273,共5页Control and Decision

基　　金：河南省自然科学基金资助项目

摘　　要：将人工神经网络和专家系统结合应用于火电厂生产过程的故障诊断,提出并设计了热工过程故障诊断系统,同时对常规BP算法做了改进,神经网络则采用分层化与并联运用相结合的结构,从而使系统诊断速度和准确率得到提高。仿真结果表明了该系统设计的合理性和可行性。The paper is about the study of controlling the very complex working process in thermal power plant, the study and design of the fault diagnosis system adopts a method of combining the neural network with expert system. Based on dividing groups, the general structure of neural network is presented, in which the method of parallel connections are simplified by reducing unnecessary weight distributions. The pace auto - adjusted algorithm is introduced into BP network. In addition, outer -combine is replaced by inter -com-bine of the neural network and expert system. So the fault diagnosis system has a high calculation rate and fault diagnosis precision. The simulation results show the design of the fault diagnosis system is reasonable and feasible.

关 键 词：故障诊断 神经网络 专家系统 生产过程 火电厂 

分 类 号：TM621.6[电气工程—电力系统及自动化]