故障诊断的多层神经网络研究  

Research on Fault-Diagnosing System Based on Multiple-Layer Artificial Neural Networks

作　　者：王志鹏[1] 孙惠国[1] 时斌 

机构地区：[1]郑州工业大学机械系

出　　处：《郑州工业大学学报》1998年第2期20-23,29,共5页Journal of Zhengzhou University of Technology

基　　金：河南省计委科技攻关项目

摘　　要：针对大型旋转机械多测点、多故障、同时性诊断问题，基于人工神经网络及贝叶斯分类规则，构造了一种由多个子网络组成的模块化、分层诊断网络，提高了子网络的分类能力，从而使整个网络有高精度、多故障同时性实时诊断的能力。Based on artificial neural networks and Bayesian decision rules, a hierarchical modular diagnosis network is proposed with respect to multiple-detectors multiple-faults simultaneous on-line diagnosis for the large-scale complicated rotary system. The subnetworks have good capability of classification, and the whole network is capable of multiple-faults simultaneous perfect diagnosing.

关 键 词：神经网络 故障诊断 多层神经网络 旋转机械 

分 类 号：TH17[机械工程—机械制造及自动化]