ConvNeXt-BiGRU Rolling Bearing Fault Detection Based on Attention Mechanism

This paper proposes a ConvNeXt-BiGRU rolling bearing fault detection method based on the attention mechanism, which aims to address the issues of low feature selection accuracy and insufficient model generalization ability of traditional neural networks for rolling bearing fault detection in complex working environments. By using this strategy, the defect diagnosis algorithm becomes less complex while maintaining its accuracy and efficiency. The defect diagnosis model can extract spatial features and analyze temporal features thanks to the ConvNext-BiGRU network model with an additional attention mechanism. Tests indicate that the technique can effectively extract vibration signal features, enhance fault detection accuracy significantly, and attain a 99.55% fault diagnosis rate.