Long term wind production modelling and forecast

This work introduces a comprehensive approach for the long term wind farm production analysis and forecast. Initially, a customized model overcomes the usually limited set of measurements by recreating a long, “virtual” history based on available climatological records. This local “virtual” history series is then analyzed and decomposed by associated “explaining” physical variables. These “explanations” are finally used to extend the resulting series, yielding a prediction (next twelve months) and a set of future scenario forecasts (at least 20 years ahead). A real case study with an application to the Northeastern Brazil is presented and discussed.