A Fault Diagnosis Method of Rolling Bearing Based on Complex Morlet CWT and CNN

In view of some shortcomings of traditional rolling bearing fault diagnosis, for instance, feature extraction relies heavily on subjective experience of people and the extracted features do not have high recognition rate for rolling element faults, a new fault type intelligent diagnosis method transforming signal recognition into image recognition based on time frequency diagram and Convolution Neural Networks (CNN) is proposed in this paper. Firstly, the Joint Time-Frequency Analysis (JTFA) with continuous wavelet transform (CWT) of complex Morlet wavelet is used to obtain the time frequency diagram features of the vibration signal, and the inputs of CNN is obtained through normalizing them. Then, the CNN is trained by the time frequency diagram with labels. Finally, the trained model is used to diagnose the fault type of the unknown data. The effectiveness of the proposed method is validated by fault simulation experiment.