无监督领域自适应轴承故障诊断方法研究  

RESEARCH ON UNSUPERVISED DOMAIN ADAPTIVE BEARING FAULT DIAGNOSIS METHOD

作　　者：吴晟凯 邵星[1] 王翠香[1] 皋军[1] WU ShengKai;SHAO Xing;WANG CuiXiang;GAO Jun(School of Information Engineering,Yancheng Institute of Technology,Yancheng 224051,China;School of Mechanical Engineering,Yancheng Institute of Technology,Yancheng 224051,China)

机构地区：[1]盐城工学院信息工程学院,盐城224051 [2]盐城工学院机械工程学院,盐城224051

出　　处：《机械强度》2024年第3期527-539,共13页Journal of Mechanical Strength

基　　金：国家自然科学基金项目(61502411,62076215);教育部新一代信息技术创新项目(2020ITA02057);江苏省研究生科研与实践创新计划项目(SJCX22_XZ035,SJCX22_XY061)资助。

摘　　要：针对基于深度学习的轴承故障诊断算法在不同工作条件和真实环境中故障样本缺乏标记的情况下诊断效果不佳的问题,提出了一种无监督的领域自适应轴承故障诊断方法,实现在无监督的情况下对不同工况的轴承进行故障诊断。首先,用快速傅里叶变换对数据进行预处理,并用卷积神经网络提取轴承故障的特征。然后,通过生成对抗网络中反转标签的方法使源域和目标域输出的特征分布趋同。最后,使用源域的分类器完成不同工况下的轴承故障诊断任务。为验证该方法有效性,在美国凯斯西储大学轴承数据集和德国帕德博恩大学轴承数据集上开展验证实验。验证结果表明,可使用无标签的目标域数据完成迁移任务,在两个数据集上表现出了较好的迁移效果,取得较高的诊断准确率。Aiming at the problem that the bearing fault diagnosis algorithm based on deep learning has poor diagnosis performance when the fault samples are lack of labels in different working conditions and real environmentsly,an unsupervised domain adaptive bearing fault diagnosis method was proposed to realize the unsupervised fault diagnosis of bearings under different working conditions.Firstly,the bearing fault sample data was preprocessed by fast Fourier transform and the features of bearing faults samples were extracted using convolutional neural network.Then,the feature distributions output of the source domain and the target domain were converged by the method of reversing labels in the generative adversarial network.Finally,the classifier of the source domain was exploited to complete the bearing fault diagnosis task under different working conditions.In order to verify the effectiveness of the proposed method,relevant comprehensive experiments were carried out on the bearing dataset of Case Western Reserve University of American and the bearing dataset of the University of Paderborn in Germany.The experimental results show that the proposed method can use the unlabeled target domain data to complete the transfer task,and it shows a good transfer performance on the two datasets and achieves a high diagnostic accuracy.

关 键 词：领域自适应 迁移学习 无监督学习 故障检测 旋转机械 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] TH165.3[自动化与计算机技术—计算机科学与技术] TH133.33[机械工程—机械制造及自动化]