Aiming at problems that traditional fault diagnosis methods need to extract features manually and the recognition rate is low

a VMD-CWT-CNN model based on variational modal decomposition (VMD) and continuous wavelet transform (CWT) combined with convolutional neural network (CNN) is proposed for rolling bearing fault diagnosis. Firstly

the bearing vibration signal is decomposed into multiple modal components with different center frequencies by VMD. Secondly

the modal components are calculated by CWT and transformed into two-dimensional time-frequency diagram. Finally

the time-frequency diagram is input into the EfficientNet convolution neural network after structure cutting

the features are automatically extracted

and the fault diagnosis of rolling bearing is completed. Using the method proposed

the average accuracy of multiple experiments on 10 types of bearing fault data from Case Western Reserve University is 99. 86%

which can effectively complete the feature extraction of rolling bearing signal and the accurate diagnosis of damage degree.