运用人工神经网络诊断电机轴承故障  

DIAGNOSING FAULTS OF ROLLING BEARING IN ELECTRICAL MACHINES BY ANN

作　　者：孙健[1] 邱阿瑞[1] 

机构地区：[1]清华大学电机系

出　　处：《电工电能新技术》1996年第4期1-6,共6页Advanced Technology of Electrical Engineering and Energy

基　　金：国家攀登计划项目

摘　　要：本文运用人工神经网络方法实现对滚动轴承故障的自动分类与诊断。文中给出了神经网络诊断系统的框图，电机滚动轴承振动信号经滤波和包络处理之后，采用ＲＮＮ网络对包络谱数据进行压缩处理，然后利用基于ＢＰ算法的前馈网络实现对轴承故障的自动分类和诊断。实验结果表明，这种诊断方法是行之有效的。The automatic classification and diagnosis of the rolling bearing faults for electrical machines are performed by ANN and the ANN diagnosing system block diagram is given in this paper. The vibration signal of the rolling bearing in motors is processed by the band-pass filter and envelope analysis. The data of the envelope spectrum are compressed by RNN , and then the automatic classification and diagnosis of the rolling bearing faults are realized by MLPs, which is based on BP algorithm. The results of the test show that this approcach of diagnosis is feasible and effective.

关 键 词：神经网络 故障诊断 电机 轴承 

分 类 号：TM307.1[电气工程—电机]