Rolling bearing fault diagnosis based on residual network and multi-scale feature fusion

Aiming at the problems of insufficient feature extraction, difficult classifier selection and low diagnostic accuracy in traditional bearing fault diagnosis under complex working conditions, a rolling bearing fault diagnosis model based on multi-scale feature fusion of residual neural network is proposed and used for fault diagnosis of motor bearings. Firstly, the wavelet transform is used to transform the bearing vibration signal into a two-dimensional time-frequency diagram as the input data set. Then, a multi-scale feature fusion module is constructed in the residual network to extract the features of fault samples at different scales. Finally, the bearing data set is input into the network to realize feature extraction and fault diagnosis. Experiment results show that the proposed fault diagnosis model based on multi-scale feature fusion of residual network can fully extract signal features and improve the accuracy of fault diagnosis.