Fault diagnosis for wind turbine gearbox based on recurrent convolutional generative adversarial network
Wind turbine gearbox is a key part of the wind turbine transmission system,but its faults are relatively random and the number of fault samples is insufficient,which seriously affect the accuracy of fault diagnosis.To solve this problem,a fault diagnosis method based on recurrent convolutional generative adversarial network was proposed in this paper.First,a sample generation model based on recurrent convolutional generative adversarial network was constructed,in which convolutional networks and recurrent networks were used as generators to enhance the temporal correlation between samples.Wasserstein distance and gradient penalty terms were introduced to improve the objective function and the game confrontation mechanism was used to optimize the generator and discriminator so that the generalization ability of the model could be strengthened.Then a fault diagnosis method based on stacked denoising autoencoder was designed by combining real samples and generated samples to realize the fault diagnosis of gearbox.Finally,the performance of the proposed fault diagnosis method was verified by the data set from the wind turbine transmission system.The results show that the proposed method can effectively balance the fault sample data set,which further improves the fault diagnosis accuracy of the wind turbine gearbox.
万方数据
维普
