计算机与数字工程2025，Vol.53Issue(2)：587-592,6.DOI:10.3969/j.issn.1672-9722.2025.02.049
结合稀疏结构的域自适应轴承故障诊断
Domain Adaptive Bearing Fault Diagnosis Combining Sparse Structure
摘要
Abstract
A domain adaptive bearing fault diagnosis method with sparse structure is proposed for the problem that traditional deep learning uses computationally intensive methods,resulting in too many redundant parameters and low recognition ability under variable load conditions.Firstly,a large size convolutional kernel is used for feature extraction of the signal,then the network struc-ture is optimized by sparse structure to improve the sparsity of the network and reduce the excessive redundant parameters caused by the deepening of the network,and adaptive batch normalization is added to overcome the difference of sample distribution between the source and target domains to enhance the variable load diagnosis capability of the model.Finally,experiments are conducted on the bearing dataset of Case Western Reserve University and the actual collected data,and the experimental results show that the av-erage accuracy can reach over 99%under both variable load and actual fault conditions,which are higher than the comparison meth-ods,indicating the effectiveness and superiority of the method.关键词
计算密集型/稀疏结构/自适应批标准化Key words
computationally intensive/sparse structure/adaptive batch normalization分类
信息技术与安全科学引用本文复制引用
薛祥凯,孙双,叶蕾..结合稀疏结构的域自适应轴承故障诊断[J].计算机与数字工程,2025,53(2):587-592,6.基金项目
黑龙江自然科学基金项目(编号:LH2021F008) （编号:LH2021F008）
海南省科技专项(编号:ZDYF2022SHFZ047)资助. （编号:ZDYF2022SHFZ047）
