Research on Bearing Fault Diagnosis Method Based on Wavelet Packet Transfer Learning
Aiming at the problem that traditional bearing fault diagnosis not only requires human prior knowledge but also has low bearing classification accuracy under variable working conditions,a bearing fault diagnosis method based on wavelet packet transfer learning(WPT-1DCNNMM)is proposed.The bearing data(source domain and target domain)are decomposed into signal components of different frequency scales through four-layer wavelet packets.The decomposed signal components are sent to a one-dimensional convolutional neural network(1DCNN)for extraction deep fault features.The distance between the source domain and the target domain is measured by multi kerner-maximum mean discrepancy(MK-MMD)to complete bearing fault classification.The proposed method is verified on the bearing dataset of Case Western Reserve University.The experi-mental results show that the method can not only effectively extract the bearing fault features,but also has a higher classifica-tion accuracy of bearings under variable working conditions than other classification models.
failure diagnosiswavelet packet decompositiontransfer learningmulti kerner-maximum mean discrepancy
万方数据
维普
