基于Elman网络LM算法的异步电机故障诊断  

Asynchronous Motor Fault Diagnosis Based on LM Algorithm of Elman Neural Network

作　　者：闫俊荣[1] 闵勇[2] 郭西进[3] 

机构地区：[1]江苏师范大学电气工程及自动化学院,江苏徐州221116 [2]江苏师范大学机电工程学院,江苏徐州221116 [3]中国矿业大学信息与电气工程学院,江苏徐州221008

出　　处：《煤矿机械》2012年第8期253-255,共3页Coal Mine Machinery

基　　金：国家自然科学基金资助项目(60804022)

摘　　要：电机故障诊断问题在生产安全运行中非常重要,但难以建立准确数学模型,而神经网络能较好地解决故障诊断问题,Elman网络是一种动态递归神经网络,具有适应时变特性的能力,训练速度快、精度高,识别能力强。针对电机转子故障样本应用Elman网络并采用LM算法训练,将其训练效果与BP网络训练效果进行比较,显示了Elman网络和LM算法的优越性。It is an important problem of asynchronous motor fault diagnosis in safety production and operation,but this problem is described difficultly by an accurate mathematical model,and neural network can deal with it.Elman network is a dynamic recursion NN.It can adapt in time varying.Its training speed is fast,and precision is high,and identification is powerful.In this paper,an asynchronous motor fault sample was trained by Elman NN and LM algorithm,and training effect was compared with BP NN＇s,and it showed that Elman NN and LM algorithm were superior.

关 键 词：故障诊断 ELMAN网络 LM算法 

分 类 号：TM307[电气工程—电机]