Fault diagnosis model based on STFT and multi-scale convolution channel attention

In recent years, deep learning has been widely used in rolling bearing fault diagnosis due to its powerful feature learning ability. However, traditional fault diagnosis models based on deep learning still have certain shortcomings. On the one hand, the single time domain or frequency domain analysis method has limited ability in processing non-stationary signals. On the other hand, the model’s ability to extract fault features is weak, resulting in low diagnostic accuracy. To address the above problems, this paper proposes a fault diagnosis model based on STFT and multi-scale convolution channel attention. Firstly, the original signal is converted from the time domain to the time-frequency domain by STFT. Secondly, convolutions of different scales are used to extract multi-scale features in the time-frequency domain, and feature enhancement is performed through channel attention. Finally, experimental verification is performed using the CWRU dataset. The results show that the proposed model outperforms other advanced fault diagnosis models in terms of diagnostic accuracy and generalization ability.