PI‐AF‐MISO‐MFAC for Ultra‐Short‐Term Solar–Wind Power Forecasting
IET Generation, Transmission & Distribution · 2026-01 · 20 (1)
DOI: 10.1049/gtd2.70364 ↗Kurz gefasst
Fachartikel in der Zeitschrift IET Generation, Transmission & Distribution, online erschienen am 22.07.2026. Verfasst von Zhiyi Zhang. Die Arbeit nennt 29 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0).
Abstract (Originalsprache)
Accurate short‐term forecasting of hybrid solar‐wind renewable power is essential for improving grid stability and renewable energy dispatch efficiency under highly variable meteorological conditions. However, conventional deep learning forecasting models often require extensive offline training and high computational resources, limiting their applicability in low‐latency edge‐side renewable forecasting environments. To address these challenges, a physics‐guided adaptive‐forgetting multi‐input single‐output model‐free adaptive control (PI‐AF‐MISO‐MFAC) framework is developed for ultra‐short‐term renewable power forecasting. The method combines recursive pseudo‐partial derivative estimation with a dual‐source adaptive forgetting mechanism and lightweight physics‐guided projection constraints to improve online prediction stability under rapidly changing renewable operating conditions. Unlike conventional MFAC control applications, the recursive formulation is interpreted here as an online forecasting update mechanism driven by historical meteorological observations. Experiments on the German 2016 renewable energy dataset demonstrate that PI‐AF‐MISO‐MFAC achieves lower one‐step‐ahead forecasting errors than several statistical, machine learning, and deep learning benchmark models, including ARIMA, XGBoost, Informer, TCN, and N‐BEATS. For hybrid renewable generation forecasting, the proposed method achieves an MAE of 23.6 kW while maintaining low computational complexity and stable parameter evolution across different seasonal and meteorological conditions. These results indicate that the developed approach provides a physically guided and computationally efficient recursive forecasting solution for ultra‐short‐term renewable energy prediction.
Zitieren
APA
Zhang, Z. (2026). PI‐AF‐MISO‐MFAC for Ultra‐Short‐Term Solar–Wind Power Forecasting. IET Generation, Transmission & Distribution, 20(1). https://doi.org/10.1049/gtd2.70364
Vancouver
Zhang Z. PI‐AF‐MISO‐MFAC for Ultra‐Short‐Term Solar–Wind Power Forecasting. IET Generation, Transmission & Distribution. 2026;20(1). doi:10.1049/gtd2.70364
BibTeX
@article{Zhang2026,
author = {Zhang, Zhiyi},
title = {PI‐AF‐MISO‐MFAC for Ultra‐Short‐Term Solar–Wind Power Forecasting},
journal = {IET Generation, Transmission & Distribution},
year = {2026},
volume = {20},
number = {1},
doi = {10.1049/gtd2.70364}
}
EUVIMED