科技创新与应用2025，Vol.15Issue(19)：1-4,4.DOI:10.19981/j.CN23-1581/G3.2025.19.001
基于ResNet50-CBAM模型的滚动轴承故障诊断研究
摘要
Abstract
Aiming at the shortcomings in traditional rolling bearing fault signal feature extraction,a rolling bearing fault diagnosis method based on Convolutional Block Attention Module(CBAM)and residual network(ResNet50)is proposed.The fault signals in the Case Western Reserve University data set were randomly and locally overlapped sampled,and the bearing fault signals were converted into two-dimensional time-frequency domain images using ICEEMDAN and Hilbert.The time-frequency domain images were then input into the ResNet50-CBAM network model.,training and testing the accuracy of the model.Convolutional neural networks and transfer learning are added to the network model to solve the problems of difficulty in data acquisition and long training time.Experiments have proved that ResNet50-CBAM has strong fault feature extraction capabilities.Compared with other network models,the accuracy rate is 8%～15%higher.Finally,rolling bearing signals are collected on a servo system experimental simulation platform,and the improved network model is used for diagnosis.The results prove that this diagnosis method has high accuracy in rolling bearing fault diagnosis.关键词
滚动轴承/故障诊断/ResNet50-CBAM/网络模型/数据Key words
rolling bearing/fault diagnosis/ResNet50-CBAM/network model/data分类
信息技术与安全科学引用本文复制引用
王鹏,邢高举,牛浩平..基于ResNet50-CBAM模型的滚动轴承故障诊断研究[J].科技创新与应用,2025,15(19):1-4,4.
