부산대학교 도서관

In order to solve the problem that the current bearing fault diagnosis model based on voiceprint signal is not enough to extract time features, a bearing fault diagnosis model based on 3DCNN is proposed in this paper. First, the Mel-spectrogram is used to extract the voiceprint of the bearing. Then, the 3DCNN model proposed is used to diagnose the fault of the bearing to make better use of the timing information of the model. Finally, the model proposed in this paper has improved the precision and recall rate by 6.25% and 7.03% respectively compared with the current classical algorithm. The model has good accuracy and is important for engineering practice.