Bearing Fault Diagnosis Method Based on Morphological Feature and Time-Frequency Interactive Convolutional Network

Bearing fault diagnosis is of great significance for reliability analysis of rotating machinery. Aiming at the problem that it is difficult to extract fault features of rolling bearing signals in complex backgrounds, this paper proposes a novel bearing fault diagnosis method based on time-frequency morphological feature module (TFMFM) and time-frequency interactive convolutional network (TFICN). Firstly, TFMFM obtains time domain features containing pulse information and frequency domain features containing fault frequencies. Then the time-domain and frequency-domain features are fed to the TFICN, which focuses on both time-domain and frequency-domain information through interactive attention and self-attention. Finally, the classification results are obtained by linear mapping. TFMFM-TFICN is compared with the advanced techniques, and the effectiveness and advancement of the proposed method is demonstrated through experiments.