Denoising of wind speed data by wavelet thresholding

Completeness and reasonableness test of wind farm data is a critical prerequisite for wind speed prediction, meanwhile it plays an important role on the reliability and safety of the grids. The key to dealing with the unreasonable data lies in eliminating the noise. In this work, we perform an analysis of the noise source of wind speed data observed in wind farm and highlight using the wavelet analysis to eliminate the noise. To test the denoising effect, the equivalent wind speed model is then established. Experiment results show that the effective preprocessing of the raw data yield good performance for attenuating the influence of noises.