Numerical Wind Speed Correction Method Based on Multiple Factors

Accurate wind speed prediction plays an important role in the effective utilization of wind energy. However, the current numerical prediction methods always have errors in wind speed prediction, and the prediction of wind speed is only based on a single variable and does not take into account the influence of other meteorological factors on wind speed. Therefore, a wind speed correction method based on multi-element numerical prediction is proposed in this paper. Firstly, the meteorological elements with a high correlation with wind speed were selected by a random forest algorithm, and then these meteorological elements were input into the AU-Net model based on the U-Net model for wind speed correction. Finally, through the experimental verification and the test of site observation data, the effect of multi-factor numerical prediction and correction is better than that of single-factor numerical prediction.