Minimum-cost numerical prediction system for wind power in Uruguay, with an assessment of the diurnal and seasonal cycles of its quality

Abstract

This work presents the results of a numerical forecast system of minimum cost for the electric power generated by wind farms in Uruguay. By keeping at minimum levels both the computational costs and the complexity of the empirical corrections of the numerical results, we obtain a benchmark for the forecast skill of more complex forecast systems, that is easily available during their calibration stages and operative functioning. The work also aims to explore the diurnal and seasonal cycle of the forecasts quality. It is found that this simple forecast system produces very good results, albeit the dependencies of the forecast skill and errors respect to the season of the year and the time of the day are distinguishable. It is also found that it is necessary to take into account the diurnal and the seasonal cycles during the calibration of the empirical corrections. The good results of this simple technique might had been possible due to the relative smooth topography of Uruguay.