EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Short-Term Wind Power Forecasting with Transformer-Based Models Enhanced by Time2Vec and Efficient Attention

Djayr Alves Bispo Junior, Gustavo de Novaes Pires Leite, Enrique Lopez Droguett, Othon Vinicius Cavalcanti de Souza, Lucas Albuquerque Lisboa, George Darmiton da Cunha Cavalcanti, Alvaro Antonio Villa Ochoa, Alexandre Carlos Araújo da Costa, Olga de Castro Vilela, Leonardo José de Petribú Brennand, Guilherme Ferretti Rissi, Giovanni Moura de Holanda und 1 weitere

Energies · 24.11.2025 · 18 (23) · 6162

DOI: 10.3390/en18236162 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Energies, online erschienen am 24.11.2025. Verfasst von Djayr Alves Bispo Junior und 12 weiteren Autor:innen. Die Arbeit wurde bisher 12-mal zitiert (Crossref) und nennt 65 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: CAPES, CNPq, FACEPE.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Accurate wind power forecasting is essential to optimize wind farm operations and ensure the stable integration of renewable energy into the grid. This study explores Transformer-based architectures to address the challenges of wind variability and temporal dependencies in short-term forecasting. A sensitivity analysis on model architecture is conducted, incorporating Time2Vec—a temporal encoding technique that captures complex temporal patterns. In addition, we replace the standard FullAttention mechanism with ProbSparse Attention, FlowAttention and FlashAttention, resulting in the Informer, Flowformer and Flashformer models, to improve computational efficiency while maintaining predictive accuracy. The novelty of this work lies in applying FlashAttention within the context of wind power forecasting and integrating Time2Vec into the Informer, Flowformer and Flashformer models. We propose four architectures—T2V-Transformer, T2V-Informer, T2V-Flowformer, and T2V-Flashformer—and compare them against benchmark models: Multi-Layer Perceptron (MLP), Long Short-Term Memory (LSTM), and DLinear. Real-world data from a wind farm in the Northeast of Brazil is used under two forecasting scenarios. In Scenario A, T2V-Transformer, T2V-Informer and T2V-Flashformer achieved Improvement over Reference RMSE (IoR-RMSE) scores of 17.73%, 17.59% and 16.67%, respectively. In Scenario B, T2V-Flowformer and T2V-Flashformer reached 27.84% and 27.45%, respectively. These results confirm the effectiveness of the proposed models in advancing short-term wind power forecasting.

Förderung

  • CAPES
  • CNPq
  • FACEPE
  • CPFL-ANEEL R&D Program

Zitieren

APA

Junior, D. A. B., Leite, G. D. N. P., Droguett, E. L., Souza, O. V. C. D., Lisboa, L. A., Cavalcanti, G. D. D. C., Ochoa, A. A. V., Costa, A. C. A. D., Vilela, O. D. C., Brennand, L. J. D. P., Rissi, G. F., Holanda, G. M. D., & Ren, T. I. (2025). Short-Term Wind Power Forecasting with Transformer-Based Models Enhanced by Time2Vec and Efficient Attention. Energies, 18(23), 6162. https://doi.org/10.3390/en18236162

Vancouver

Junior DAB, Leite GDNP, Droguett EL, Souza OVCD, Lisboa LA, Cavalcanti GDDC, et al. Short-Term Wind Power Forecasting with Transformer-Based Models Enhanced by Time2Vec and Efficient Attention. Energies. 2025;18(23):6162. doi:10.3390/en18236162

BibTeX
@article{Junior2025,
  author = {Junior, Djayr Alves Bispo and Leite, Gustavo de Novaes Pires and Droguett, Enrique Lopez and Souza, Othon Vinicius Cavalcanti de and Lisboa, Lucas Albuquerque and Cavalcanti, George Darmiton da Cunha and Ochoa, Alvaro Antonio Villa and Costa, Alexandre Carlos Araújo da and Vilela, Olga de Castro and Brennand, Leonardo José de Petribú and Rissi, Guilherme Ferretti and Holanda, Giovanni Moura de and Ren, Tsang Ing},
  title = {Short-Term Wind Power Forecasting with Transformer-Based Models Enhanced by Time2Vec and Efficient Attention},
  journal = {Energies},
  year = {2025},
  volume = {18},
  number = {23},
  pages = {6162},
  doi = {10.3390/en18236162}
}
Kostenlos starten →