A Deep Learning-Based Bearing Fault Diagnosis Model: IEMDWave-SwinTGAMNet

In the realm of mechanical equipment fault diagnosis, the identification of bearing faults using deep learning techniques has garnered significant attention. Many researchers aim to enhance the accuracy of bearing fault detection. This paper proposes an improved IEMDWave-SwinTGAMNet model, which integrates advanced signal noise reduction techniques with deep learning methodologies. The model features a low-noise time-frequency image generation method and employs a deep neural network that utilizes a fusion attention mechanism to extract both local and global features from time-frequency images. The model’s effectiveness was validated using a bearing fault dataset from Case Western Reserve University, USA. The results demonstrate that the proposed model achieves high recognition accuracy for bearing faults.