A novel wind turbine de-noising method based on the Genetic Algorithm optimal Mexican hat wavelet

摘要

This paper proposed a new wind turbine de-noising method based on the Genetic Algorithm (GA) optimal Mexican hat wavelet. The filtering characteristic of the Mexican hat wavelet is studied first and then improved with two shape changing parameters. The GA is introduced to optimize the shape parameters by the selection, genetic and replacement process. The scale factor in the wavelet transform (WT) is then optimized using the same arithmetic. Finally the optimal Mexican hat wavelet and the scale factor are applied in the continue wavelet transform to extract the useful features of the analyzed signal. The experiment results using the real wind turbine gearbox vibration signal proved the effectiveness and validity of the new method.