A Research on Fault Diagnosis of Wind Turbine CMS Based on Bayesian-GAN-LSTM Neural Network

Abstract

Fault diagnosis of large components of wind turbines is of great significance in improving the reliability of wind turbines. In the actual fault diagnosis project, insufficient data labels and low recognition accuracy are two major problems. In order to make up for these two deficiencies, this paper proposes to combine the generative adversarial neural (GAN) network and the LSTM model and uses the Bayesian distribution to optimize the GAN and LSTM, respectively. GAN uses the generator to solve the problem of insufficient data labels, and the Bayesian optimized LSTM prediction accuracy is better. This paper uses the actual wind turbine bearing data to test the algorithm, and the accuracy of the test results reaches 97.6%, which shows the algorithm is accurate and robust, and the upgraded algorithm can be applied to the actual fault diagnosis of large components of wind turbines.