Aggregated wind power prediction methods based on distances between weather forecasting vectors

This paper describes a method to make aggregated wind power predictions for a region with several wind farms distributed. As the number of wind farms in a region increases, the use of aggregated wind power prediction methods becomes faster than making predictions for each of the wind farms in the region and add them. The method described is based on the comparison of new wind-speed forecasting vectors with a set of historical wind-speed forecasting vectors and the corresponding historical wind power measured. Some distances are presented to make the comparison of vectors, and some selection methods of the nearest wind-speed vectors to calculate the wind power prediction. Finally, the study includes a comparison of the aggregated methods with another wind power prediction agency and with SIPREOLICO, a wind power prediction system used by the Spanish System Operator.