摘要
阐述了机械设备故障诊断机理的发展，研究了一种新的状态识别方法——人工神经网络诊断法，基于并行运算的计算机新功能，是人工智能的一个重要方面，其核心思想是人脑结构的物理模拟。使用大量的专家经验对网络权值进行训练，以实现知识的存储；通过对ＢＰ算法的改进，用基因遗传工程思路指导神经网络的训练，能较好地模拟人类专家的思维过程。应用特例说明了本文所研究方法的实用性和可靠性。
In this paper we set forth the development of mechanism of mechanical equipment fauIt diagnosis,and a new condition distinguishing method Neural Net Method for fault diagnosis in mechanical equipment wasstudied.This method based on new function of computer-parallel operation, was an important part of AI(artificialintelligence).Its core thought was simulation of human brain in physics,The knowlege memory was realized bytraining the weighs of net through a huge amount of expert experience,On the basis of BP method we use GA(genetic-algorithm) method to train neural net,which can simulate better humen thought process. By practicalexemples,it may be seen that the method presented in the paper has been of practicality and reliability.
出处
《机械强度》
CAS
CSCD
北大核心
1995年第2期48-54,共7页
Journal of Mechanical Strength
关键词
故障诊断
识别
机械设备
神经网络
fault diagnosis,condition,distinguishing,equipment,machinery
