Ultra-Short-Term Wind Power Forecasting Based on Variational Mode Decomposition and XGBoost Ensemble
Ingegneria Sismica · 08.07.2026 · 42 (4) · 1-15
DOI: 10.65102/is20261299 ↗Kurz gefasst
Fachartikel in der Zeitschrift Ingegneria Sismica, online erschienen am 08.07.2026. Verfasst von Hui Yuan.
Abstract (Originalsprache)
A VMD-XGBoost ensemble method for enhancing the accuracy of ultra-short-term wind power forecasting at 2-4 hour horizons. The original power series is decomposed into intrinsic mode functions by variational mode decomposition (VMD), and separate XGBoost models, with lagged features, are built for each component, with final forecasts obtained by summation. Using 3.3 years of 15-min wind farm data, the proposed method achieves an R2 of 0.8922 and an RMSE of 21.815 MW at the 4-h horizon, outperforming raw XGBoost by 18.1% in R2 and 33.6% in RMSE. Inference time remains below 0.05 s, confirming real-time applicability.
Zitieren
APA
Yuan, H. (2026). Ultra-Short-Term Wind Power Forecasting Based on Variational Mode Decomposition and XGBoost Ensemble. Ingegneria Sismica, 42(4), 1-15. https://doi.org/10.65102/is20261299
Vancouver
Yuan H. Ultra-Short-Term Wind Power Forecasting Based on Variational Mode Decomposition and XGBoost Ensemble. Ingegneria Sismica. 2026;42(4):1-15. doi:10.65102/is20261299
BibTeX
@article{Yuan2026,
author = {Yuan, Hui},
title = {Ultra-Short-Term Wind Power Forecasting Based on Variational Mode Decomposition and XGBoost Ensemble},
journal = {Ingegneria Sismica},
year = {2026},
volume = {42},
number = {4},
pages = {1--15},
doi = {10.65102/is20261299}
}
EUVIMED