EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Explaining solar forecasts with generative AI: A two-stage framework combining transformers and LLMs

Ayesha Siddiqa, Nadim Rana, Wazir Zada Khan iD, Fathe Jeribi, Ali Tahir

PLOS One · 17.09.2025 · 20 (9) · e0331516

DOI: 10.1371/journal.pone.0331516 ↗

Kurz gefasst

Fachartikel in der Zeitschrift PLOS One, online erschienen am 17.09.2025. Verfasst von Ayesha Siddiqa und 4 weiteren Autor:innen. Beteiligte Einrichtung: Department of Computer Science, University of Wah, Wah, Pakistan. Hauptthemen laut MEDLINE: Solar Energy, Deep Learning. Die Arbeit wurde bisher 6-mal zitiert (Crossref) und nennt 19 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Accurate and interpretable solar power forecasting is critical for effectively integrating Photo-Voltaic (PV) systems into modern energy infrastructure. This paper introduces a novel two-stage hybrid framework that couples deep learning-based time series prediction with generative Large Language Models (LLMs) to enhance forecast accuracy and model interpretability. At its core, the proposed SolarTrans model leverages a lightweight Transformer-based encoder-decoder architecture tailored for short-term DC power prediction using multivariate inverter and weather data, including irradiance, ambient and module temperatures, and temporal features. Experiments conducted on publicly available datasets from two PV plants over 34 days demonstrate strong predictive performance. The SolarTrans model achieves a Mean Absolute Error (MAE) of 0.0782 and 0.1544, Root Mean Squared Error (RMSE) of 0.1760 and 0.4424, and R 2 scores of 0.9692 and 0.7956 on Plant 1 and Plant 2, respectively. On the combined dataset, the model yields an MAE of 0.1105, RMSE of 0.3189, and R 2 of 0.8967. To address the interpretability challenge, we fine-tuned the Flan-T5 model on structured prompts derived from domain-informed templates and forecast outputs. The resulting explanation module achieves ROUGE-1, ROUGE-2, ROUGE-L, and ROUGE-Lsum scores of 0.7889, 0.7211, 0.7759, and 0.7771, respectively, along with a BLEU score of 0.6558, indicating high-fidelity generation of domain-relevant natural language explanations.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Solar EnergyDeep LearningForecastingWeatherModels, Theoretical

Autor:innen und Institutionen

  • Department of Computer Science, University of Wah, Wah, Pakistan.
  • Department of Computer Science, College of Engineering and Computer Science, Jazan University, Jazan, Saudi Arabia.
  • Engineering and Technology Research Center, Jazan University, Jazan, Saudi Arabia.

ORCID-Profile: Wazir Zada Khan

Zitieren

APA

Siddiqa, A., Rana, N., Khan, W. Z., Jeribi, F., & Tahir, A. (2025). Explaining solar forecasts with generative AI: A two-stage framework combining transformers and LLMs. PLOS One, 20(9), e0331516. https://doi.org/10.1371/journal.pone.0331516

Vancouver

Siddiqa A, Rana N, Khan WZ, Jeribi F, Tahir A. Explaining solar forecasts with generative AI: A two-stage framework combining transformers and LLMs. PLOS One. 2025;20(9):e0331516. doi:10.1371/journal.pone.0331516

BibTeX
@article{Siddiqa2025,
  author = {Siddiqa, Ayesha and Rana, Nadim and Khan, Wazir Zada and Jeribi, Fathe and Tahir, Ali},
  title = {Explaining solar forecasts with generative AI: A two-stage framework combining transformers and LLMs},
  journal = {PLOS One},
  year = {2025},
  volume = {20},
  number = {9},
  pages = {e0331516},
  doi = {10.1371/journal.pone.0331516}
}
Kostenlos starten →