关键词: 风电齿轮箱, 故障诊断, 支持向量机, 参数优化

Abstract:

In  a  wind turbine gearbox,gear  teeth  produced  frequently the  problems  such as broken,pitting and wear failure.This paper will extract non-stationary  nonlinear vibration  signals of  wind power gear box and enhance wavelet packet energy entropy with using  SVM to fault diagnosis processing.To  improve the classification  accuracy  of  the  algorithm,genetic algorithms was used to optimize of parameters. Tests show that the best parameters can improve the prediction accuracy of the SVM to test samples.

Key words: wind , power gearbox;fault , diagnosis;support , vector , machine(SVM);parameter optimization