基于CWT和优化Swin Transformer的风电齿轮箱故障诊断方法

Here, aiming at shortcomings of traditional fault diagnosis method applied in wind power gearbox operation fault diagnosis, a wind power gearbox fault diagnosis method based on wavelet transform and optimized Swin Transformer was proposed. This method could use wavelet transform to convert vibration signals of a wind turbine gearbox into a time-frequency map. Samples were expanded using SuperMix data augmentation algorithm. The transfer learning technique was used to train and optimize Swin Transformer model with pre-trained model parameters. The optimized and trained Swin Transformer model was applied in contrastive verification of actual wind field operation and maintenance data with a classification correct rate of 99. 61%. The verification results showed that this method can effectively realize fault diagnosis of wind power gearbox and improve the model’ s recognition correct rate.