基于WP-LSTM的偏心转子马达故障诊断方法  

Fault Diagnosis Method for Eccentric Rotor Motor Based on WP-LSTM

作　　者：冯战 王杰[1] 黄思思 方夏 FENG Zhan;WANG Jie;HUANG Si-si;FANG Xia(School of Manufacturing Science and Engineering,Sichuan University,Chengdu 610065,China)

机构地区：[1]四川大学机械工程学院,成都610065

出　　处：《组合机床与自动化加工技术》2020年第10期98-102,105,共6页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：四川省重点研发项目(2019YFG0356)。

摘　　要：为解决智能设备中的偏心转子马达故障检测准确性与效率低下等问题,提出一种基于小波包与长短时记忆网络(Wavelet Packet-Long Short Term Memory,WP-LSTM)的故障诊断方法。首先,将偏心转子马达的电压信号进行小波包分解,对高频信号进行重构。其次,将重构信号作为特征向量输入到3层LSTM网络中,依靠LSTM网络的记忆特性充分学习非稳态信号中具有时序性的故障特征信息,再利用模型诊断出马达断线、卷线、电刷不良和接触不良故障。最后,通过实验验证了所提方法的可行性,且准确率高达98.91%。与现有的马达故障诊断方法相比,基于WP-LSTM的诊断方法具有更好的诊断效果,对提高故障诊断的准确率有一定的作用。In order to solve the problem of fault detection accuracy and low efficiency of eccentric rotor motor in smart equipment,a fault diagnosis method based on Wavelet Packet-Long Short Term Memory(WP-LSTM)is proposed.Firstly,the voltage signal of the eccentric rotor motor is decomposed by wavelet packet,and the high frequency signal is reconstructed.Secondly,the reconstructed signal is input as a feature vector into the 3-layer LSTM network.The time-series fault feature information in the unsteady signal is learned by the memory characteristics of the LSTM network,and then the model is used to diagnose the motor disconnection and winding,poor brush and poor contact failure.Finally,the feasibility of the proposed method is verified by experiments,and the accuracy rate is as high as 98.91%.Compared with the existing motor fault diagnosis methods,the WP-LSTM-based diagnostic method has a better diagnostic effect and has a certain effect on improving the accuracy of fault diagnosis.

关 键 词：LSTM网络 小波包 故障诊断 神经网络 偏心转子马达 

分 类 号：TH162[机械工程—机械制造及自动化] TG506[金属学及工艺—金属切削加工及机床]