A New Strategy for Wind Speed Forecasting Based on Autoregression and Wavelet Transform

A new strategy for wind speed prediction based on autoregression (AR) and wavelet transform is proposed.The model manages the original wind data with wavelet transform and utilizes AR model to forecast the different components of the wind speed signal in order to get the eventual forecasted wind results through wavelet restructure.Analysing the wind data of Jiuquan,China in April,2010 with the new method demonstrates that the the correlation coefficients between the predicted value and original data are 0.97,0.93 and 0.83 for 1 hour,3 hours and 7 hours ahead ,respectively.The experimental results demonstrate that the new model gives acceptable results,which offers a novel and available idea for prediction on wind speed.