A Lightweight Residual Network and Its Application in Bearing Fault Diagnosis

Fault diagnosis is crucial for the safe operation of rotating machinery. Existing rolling bearing fault diagnosis algorithms generally have low accuracy in the presence of noise, and there are issues with complex algorithm structures and high computational complexity. Therefore, this paper researches rolling bearing fault diagnosis using a lightweight residual network. First, the signals are preprocessed using the short-time Fourier transform to highlight the impact features of the vibration signals. Then, a lightweight residual network based on channel attention is designed to focus on the impact features of faults, effectively improving the algorithm's ability to extract nonlinear features. Finally, experimental results demonstrate that the proposed algorithm can effectively solve the problem of rolling bearing fault diagnosis in noisy and multi-load environments, with low computational complexity and model size.