|
|
Research on Fault Diagnosis Algorithm of Rolling Bearings Based on Multi-scale Adaptive Convolution Transformer
2026, 46(4):
100-106.
Rolling bearing fault diagnosis is crucial for the reliable operation of industrial equipment, and the diagnostic accuracy and robustness of traditional methods under complex working conditions still need to be improved. In order to more effectively explore the fault characteristics of rolling bearings and improve the accuracy of fault diagnosis, a multi-scale adaptive convolution Transformer rolling bearing fault diagnosis method is proposed. Firstly, a rolling bearing fault experimental platform was established, and acceleration sensors were used to collect the vibration signals from rolling bearings, which come from a variety of operating conditions in different states at the fan end of the motor; secondly, the collected one-dimensional vibration signals were processed by a multi-scale convolutional neural network.This network were used to deal with the time-domain features and frequency-domain features, and to deeply excavate the multi-dimensional features.A channel-attention mechanism was introduced to adaptively weight the feature channels to improve the fusion effect; finally, the local convolution module were applied to further deepen the features. The features are further deepened by the local convolution module, the global feature extraction module was used to capture the global features of the vibration signal.An adaptive blending module was utilized to realize the dynamic fusion of different features, which enhances the model's ability to identify weak fault features. On this basis, the effectiveness of the method is verified by different rolling bearing experimental data, and the experimental comparative analysis with MCNN, CNN-LSTM, Transformer, and AdaMCT is conducted.the results showed that the proposed method outperforms the existing methods in terms of fault recognition accuracy, which reaches 99.8%. The model also has high diagnostic accuracy under different noise levels, which further proves the superiority of the method and provides a new method for rolling bearing diagnosis in industry.
Related Articles |
Metrics
|