An upscaling minute-level regional photovoltaic power forecasting scheme

• DOI: 10.1016/j.ijepes.2023.109609
• Publication Date: 2023-11-05T16:22:03Z
• Authors: Xiangjian Meng; Xinyu Shi; Weiqi Wang; Yumin Zhang; Feng Gao

Along with the increasing penetration of photovoltaic (PV) power generation, regional forecasting becomes more and critical for stable economical operation system. The key challenge PV technology is lack complete accurate historical data since not all plants are equipped precise real-time output monitoring Besides, computation burden will be heavy when number in target region large. This paper therefore proposes an upscaling minute-level scheme using selected reference plants. In this paper, a novel method selection proposed by comprehensively considering prediction accuracy artificial neural network (ANN) as well Pearson correlation coefficient. plant coefficient μ introduced comprehensive indicator selection, which incorporates MAPE. addition, correction assumed to guarantee proper forecasting. flexible approach effectively decrease accumulated error rolling integrating results under different temporal resolutions. specific, resolutions 1 min, 5 min 15 simultaneously derived performance between traditional compared. validity finally verified collected installed city Eastern China. For time resolution mins 10 mins, corresponding RMSE 6.56, 5.73 4.85 MAPE 4.04%, 3.45% 2.86%.