Hybrid weather-based ANN for forecasting the production of a real wind power plant

In this paper we propose a hybrid approach for long-term and short-term wind power forecasting: this approach is based both on a suitably defined ANN model for wind speed forecasting, and a Numerical Weather Prediction (NWP), usually employed by a weather forecasting service. The simulations of real plant data show that this ensemble method allows to decrease the influence of the intrinsic limits of the two approaches considered individually and enhances their potential on a broader range of time-horizons.