Fault Diagnosis of Rolling Bearings Based on CNN-BiLSTM-SE Model

This study proposes a rolling bearing fault diagnosis model called CNN-BiLSTM-SE, which integrates Convolutional Neural Networks (CNN), Bidirectional Long Short-Term Memory Networks (BiLSTM), and Squeeze-and-Excitation (SE) attention mechanism. The method is validated using the CWRU dataset. Experimental results show that CNN-BiLSTM-SE model achieves a recognition accuracy of 100% for fault recognition, with a recognition accuracy of 96% for different fault types. The results demonstrate that this approach possesses strong capabilities in rolling bearing fault diagnosis and holds significant application value in industrial settings.