A Single-sample Fault Diagnosis Method of a Wind Turbine Transmission Chain

Aiming at the problem of high fault similarity in the transmission chain of wind turbines, this study proposes a single-sample wind turbine bearing fault diagnosis method based on empirical mode decomposition and signal equalization processing. In this method, the signal components of different modes are obtained by empirical mode decomposition of the actual monitoring fault signal. The energy value and peak value of each modal component are calculated, and some modal components with large energy and high peak value are selected for signal reconstruction to obtain a new fault signal. The new fault signal decomposed by wavelet packet, and the wavelet packet is decomposed into the third layer component for signal reconstruction. The variance of the reconstructed signal is used as the eigenvalue of the fault diagnosis. Non-linear equalization is performed on the eigenvalues, which solves the problem of signal mutual submergence. The concept of discrimination is introduced to quantify the difference between different fault signals. The experimental results show that the fault diagnosis method proposed is effective, and the discrimination between the four faults on the transmission chain of the wind turbine before and after the treatment is significantly increased, which shows that the experimental method has a strong robustness. Compared with the improved fuzzy clustering method and the deep learning method based on the improved AlxeNet network, the method performs better. The fault diagnosis method only uses a single sample to realize the fault diagnosis of the transmission chain of wind turbines, which is in line with the characteristic of the low failure rate of wind turbines, and is of great significance for improving the troubleshooting efficiency of wind turbines in actual projects.