Resumen de Advancing wind energy forecasting by integrating power curves into SARIMAX models
L.Ricardo De la Rosa, Alejandro gómez, Matilde Santos Peñas, Lía García Perez
Effective forecasting of wind energy is crucial tooptimizing the integration of renewable energy into power grids.Traditional forecasting models often rely on historical windspeed data, which may not fully capture the complexities ofwind energy generation. This study examines the enhancementof wind energy forecasts by incorporating power-wind speedcurves as an exogenous variable within a SARIMAX model.Data preprocessing was performed using a public dataset fromKelmarsh, UK, which involved handling missing values andnormalization, followed by an in-depth analysis of stationarityand seasonality. A power-wind speed curve was fitted to historicaldata and compared with the use of raw wind speed as anexogenous input in the SARIMAX model. The analysis indicatesthat integrating the power-wind speed curve leads to improvedforecast performance, as evaluated by metrics such as RMSE,MAE, and MAPE, compared to using wind speed alone. Thesefindings highlight the potential of incorporating domain-specificfeatures to improve wind energy forecasting, contributing to moreefficient management of renewable energy resources.
