Short Term Wind Speed Forecasting for Wind Farms Using an Improved Autoregression Method

A new method in wind speed prediction based on auto regression (AR) method is proposed. The new method not only takes actual range of predicted value into account but also combines AR with the mean filter of the wind speed waveform. The restriction on predicted value makes the prediction more conform to the fact, and the filtering varies the measured wind speed curve to become smoother, leaving the more effective data. Applying the method to analyse Anxi in China demonstrates that the proposed method provides a better wind speed prediction, and it is an excellent method for prediction of wind speed in wind farms.