86. メソアンサンブル予報システム（MEPS）予報データを入力とした機械学習モデルの日射予測大外しの分析の基礎検討

抄録

Transmission system operators need to reserve the adjustment power conservatively in the previous day's phase in preparation for the serious forecast error of variable renewable energy sources. Therefore, reduction of the extreme error of the solar power forecasting is required to reduce the cost of grid operation. In this study, for the aim of reducing the serious error of solar power forecasting, we applied the machine learning (ML) models to the meso-ensemble prediction system (MEPS) data, and the analysis was examined for the cases which our constructed ML model occurs serious errors.