Wavelets pre-filtering in wind speed prediction

Wind power is the fastest growing renewable energy technology and is becoming a significant component of the energy mix. The secure and reliable operation of the power system implies the need for scheduling in advance the energy sources that will produce, so that the power system is balanced. Therefore, the use and importance of the wind power is strictly dependent on the ability to predict the wind in advance. In this paper, ARMA models are used to forecast the wind speed in terms of a medium-term prediction. Furthermore, an investigation on the benefits of pre-filtering the wind speed time series using wavelets is carried out. Some simulations are done with the twofold purpose of evaluating the performance of ARMA models as compared with reference models and investigating whether the wavelet pre-filtering technique leads to an improvement of the forecast results.