The amount of energy generated by wind sources is increasing all over the world. Brazil has a huge potential due to its geographic localization and the Brazilian Government has given a clear signal that is prone to invest in this kind of energy. Accurate wind speed forecasts are essential in the operation planning for the electrical wind power system. This work aims to make more accurate forecasts in the northeast of Brazil. To do so, we use a model that takes into account the daily characteristics and the long memory behavior. The model applied in this particular region proved to be more accurate than the persistence model and other models.

Dantas, Tiago Mendes; Oliveira, Fernando Luiz Cyrino (Advisor). Combining to succeed: A novel strategy to improve forecasts from Exponential Smoothing models.