摘要
计算机分散控制系统（ＤｉｓｔｒｉｂｕｔｅｄＣｏｎｔｒｏｌＳｙｓｔｅｍ，简称ＤＣＳ）是在计算机技术、自动控制技术、通信技术和ＣＲＴ图形显示技术相结合的基础上开发出来的新型控制系统，是今后计算机控制系统发展的一大趋势．本文以计算机分散控制系统为背景，阐述了多级计算机测控系统的组成、原理及特点．在多级计算机测控系统的可靠性方面，着重阐述了利用人工神经网络的原理对实际测控单元进行故障诊断、报警及故障处理．本文就人工神经网络对给定知识的表达、联想记忆能力及网络的改进进行了研究，利用反向误差传播网络（Ｂａｃｋ－ＰｒｏｐａｇａｔｉｏｎＮｅｔｗｏｒｋ）对测控系统中四种常见故障，进行了实例分析和仿真，直至应用到实际系统中，取得了满意的仿真效果及实际故障诊断结果．
he Distributed Control System (DCS) is a new control system that was developed as a combination of variours techniques such as computer, control, communication and CRT (CathodeRay Tube) display. This is a trend for the Computer Control System. In this paper, the principle and charateristic of a multicomputer testing and control system is explained based on the distributed control system. In the respect of reliability of this system, the fault diagnosis, alarm and handling in the testing and control unit using the artificial neural network is emphatically presented. In this paper, knowledge representation, associative memory and network structure are studied. Utilizing the error back propagation network model, four types of typical faults in the testing and control system are analyzed and diagnosed, and the simulation results and the actual diagnostic results are quite satisfactory.
出处
《上海大学学报（自然科学版）》
CAS
CSCD
1997年第4期425-434,共10页
Journal of Shanghai University:Natural Science Edition
