MHSACAE-CNN在噪声下的电机轴承故障诊断  

Fault diagnosis of motor bearing under high noise based on MHSACAE-CNN

作　　者：文斌 李知聪[1] 朱晗 曹仁轩 WEN Bin;LI Zhi-cong;ZHU Han;CAO Ren-xuan(College of Electricity Engineering and New Energy,China Three Gorges University,Yichang 443002,China;Hubei Provin-cial Collaborative Innovation Center for New Energy Microgrid,China Three Gorges University,Yichang 443002,China)

机构地区：[1]三峡大学电气与新能源学院,湖北宜昌443002 [2]三峡大学新能源微电网湖北省协同创新中心,湖北宜昌443002

出　　处：《振动工程学报》2023年第4期1169-1178,共10页Journal of Vibration Engineering

基　　金：国家自然科学基金资助项目(61876097)。

摘　　要：电机的运行情况复杂,实际运行工况下会有大量的噪声,导致其轴承故障诊断精度下降。为了改善这一问题,提出了一种基于多头自注意力机制的一维全卷积自编码网络(One-dimensional Fully Convolutional Autoencoding Network Basedon Multi-head Self-attention,MHSACAE)与卷积神经网络(Convolutional Neural Network,CNN)结合的轴承故障诊断方法。该方法先采用MHSACAE网络进行降噪,再通过CNN进行故障诊断。其中MHSACAE去噪网络采用无监督训练的方式,充分考虑了实际工况和序列数据内在联系,在实现对噪声的滤除效果的同时,最大限度地保留下了原始的故障信息,使得CNN可以实现在噪声情况下对电机轴承故障的高精度诊断。通过与其他轴承故障诊断方法在噪声情况下进行对比,证明提出的方法具有更好的效果。The operation of the motor is complicated and there will be a lot of noise under actual operating conditions.The noise causes low-accuracy of bearing fault diagnosis.In order to improve this problem,a bearing fault diagnosis method based on the multi-head self-attention mechanism of one-dimensional fully convolutional self-encoding network(MHSACAE)combined with convolutional neural network(CNN)is proposed.Firstly,we use the MHSACAE for noise reduction.And then we use CNN for fault diagnosis.Particularly,the MHSACAE adopts an unsupervised training method.The method fully considers actual working conditions and the inherent connection of the sequence data,while the ability to filter noise is achieved and the original fault infor⁃mation is retained to the greatest extent.So that CNN can realize high-precision diagnosis of motor bearing faults under noise condi⁃tions.Finally,the comparison with other bearing fault diagnosis methods under noisy conditions proves that the proposed method has better results.

关 键 词：故障诊断 轴承 自注意力 噪声 卷积神经网络 

分 类 号：TH165.3[机械工程—机械制造及自动化] TH133.33