EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A Novel Wind Power Ramp Events Prediction Model Based on Vector Feature Forecasting and Data Transfer Learning

Ling Liu, Jujie Wang, Hongfang Song

IEEE Access · 2026 · 14 · 52826-52833

DOI: 10.1109/access.2026.3680651 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Access, 2026. Verfasst von Ling Liu und 2 weiteren Autor:innen. Die Arbeit nennt 30 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Major Humanities and Social Sciences Research Projects in Zhejiang Higher Education Institutions, National Natural Science Foundation of China.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Accurate prediction of wind power ramp events (WPRE) is essential for the safe and economic operation of wind farms. However, this task remains challenging due to the multitude of uncertain factors, such as meteorological conditions, wind turbine characteristics, and control strategies. To enhance prediction accuracy, this study introduces a novel WPRE vector prediction method. Unlike traditional research methodologies that utilize raw wind power data as model input, the proposed method first identifies and extracts WPRE vectors, which then serve as input features for prediction. This vector-based forecasting method effectively reduces the input data complexity while enhancing prediction accuracy. The proposed model integrates data transfer learning to augment training datasets and leverages BiLSTM neural networks for data forecasting. Experimental results demonstrate that the model achieves WPRE occurrence detection scores of 0.788, 0.468, and 0.583 in terms of Recall (RC), Precision ( $P$ ), and $F_{1}-Score$ , respectively. Furthermore, the prediction accuracy $R$ squared ( $R^{2}$ ) values of the four WPRE vector features $T_{min}$ , $P_{min}$ , $T_{max}$ , and $P_{max}$ are 0.770, 0.962, 0.769, and 0.950, respectively, outperforming six benchmark models.

Förderung

  • Major Humanities and Social Sciences Research Projects in Zhejiang Higher Education Institutions
  • National Natural Science Foundation of China

Zitieren

APA

Liu, L., Wang, J., & Song, H. (2026). A Novel Wind Power Ramp Events Prediction Model Based on Vector Feature Forecasting and Data Transfer Learning. IEEE Access, 14, 52826-52833. https://doi.org/10.1109/access.2026.3680651

Vancouver

Liu L, Wang J, Song H. A Novel Wind Power Ramp Events Prediction Model Based on Vector Feature Forecasting and Data Transfer Learning. IEEE Access. 2026;14:52826-52833. doi:10.1109/access.2026.3680651

BibTeX
@article{Liu2026,
  author = {Liu, Ling and Wang, Jujie and Song, Hongfang},
  title = {A Novel Wind Power Ramp Events Prediction Model Based on Vector Feature Forecasting and Data Transfer Learning},
  journal = {IEEE Access},
  year = {2026},
  volume = {14},
  pages = {52826--52833},
  doi = {10.1109/access.2026.3680651}
}
Kostenlos starten →