基于一维RepVGG协同领域自适应的电机滚动轴承故障诊断  

Motor Rolling Bearings Fault Diagnosis Based on One-dimensional RepVGG and Domain Adaptation

作　　者：周涛 罗响 朱莉[1] ZHOU Tao;LUO Xiang;ZHU Li(School of Electric Information and Electrical Engineering,Shanghai Jiaotong University,Shanghai 200240,China)

机构地区：[1]上海交通大学电子信息与电气工程学院,上海200240

出　　处：《微特电机》2023年第4期1-7,共7页Small & Special Electrical Machines

摘　　要：在使用传统机器学习类方法对电机滚动轴承故障进行诊断时,电机运行工况的变化以及采集信号时的噪声干扰,会出现源域训练集和目标域测试集分布不一致的问题。提出了基于一维RepVGG协同领域自适应的故障诊断方法。RepVGG具有精度高和速度快的特点,使用一维RepVGG实现对电机滚动轴承信号的特征提取;基于提取的特征,在网络顶层结构中使用集成优化目标函数来实现域自适应,并完成轴承故障诊断。基于凯斯西储大学轴承数据集,对该方法进行了实验验证。实验结果表明,在电机变工况运行时,改进方法为诊断性能优于现有其他诊断方法。When traditional machine learning method is used to diagnose motor rolling bearings faults,the distribution of source domain and target domain is inconsistent,due to the change of operating conditions and the noisy signal.The fault diagnosis method,based on one-dimensional re-parameterization Visual Geometry Group(RepVGG)and domain adaptation,was proposed.RepVGG had the characteristics of high precision and fast speed,which was convenient for model deployment and acceleration.One-dimensional RepVGG was used to extract the features of signals.The method used the integrated optimization objective function in the top-level structure of the network to achieve domain adaptation and the fault diagnosis of motor rolling bearings.Based on the bearing dataset of Case Western Reserve University,the experimental results show that the improved method has better diagnostic performance than other existing diagnostic methods when the motor runs under variable conditions.

关 键 词：电机滚动轴承 故障诊断 一维RepVGG 领域自适应 变工况 

分 类 号：TM307.1[电气工程—电机]