Solar irradiation forecasting through a hybrid method, considering zonda, north and south winds in San Juan, Argentina

This paper presents a hybrid method to predict a day-ahead solar irradiation curve, under extreme meteorological phenomena (Zonda, North, South winds), existing in the province of San Juan -Argentina. The proposed method is based on an Artificial Neuronal Network (ANN) which is trained with a data-set filtered by the environmental variables that characterize the mentioned phenomena. A previously calculated ideal solar irradiation curve is modified from the forecasts generated by the ANN. This is a blended method between statistical learning methods and Numerical Weather Predictions methods (NWP) typically used to improve upon the raw forecast of a NWP model. A reduction of the uncertainty on the planning of photovoltaic plants dispatch in the province of San Juan while the meteorological phenomena take place, can be achieved with the results of this forecasting method.