A Motor Bearing Fault Diagnosis Method Based on EMD-GAF-RepVGG Framework

To address the limitations of bearing fault diagnosis methods, such as insufficient extraction of fault features and inadequate capture of interactions between features, this paper proposes a fault detection model based on cross-modal feature extraction and a re-parameterized convolutional neural network (RepVGG). First, empirical mode decomposition (EMD) is used to denoise the signal. Then, Gramian Angular Fields (GAF) are applied to achieve cross-modal transformation of the signal's overall structure and instantaneous features. Finally, supervised learning is conducted using the RepVGG framework for fault detection and classification. Experimental results demonstrate that the proposed model achieves a classification accuracy of 99.21% on the Case Western Reserve University (CWRU) bearing fault dataset, proving its strong robustness and excellent detection accuracy in rotating machinery bearing fault diagnosis.