EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A Numerical–Textual Dual-Channel Forecasting Framework for Short-Term Load Forecasting

Xiaohu Xin, Jing Xu, Xin Wang, Weili Xu, Yang Yue, Chengshuai Wang, Junwei Yang, Bing Sun

Energies · 21.08.2026 · 19 (16) · 3930

DOI: 10.3390/en19163930 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Energies, online erschienen am 21.08.2026. Verfasst von Xiaohu Xin und 7 weiteren Autor:innen. Die Arbeit nennt 37 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Science and Technology Project of State Grid Tianjin Electric Power Company.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Short-term load forecasting underpins power-system dispatching, yet classical models degrade when historical samples are scarce. This paper proposes NTSForecast, a numerical–textual dual-channel framework: a trainable multi-layer perceptron encodes the numerical input, a frozen large language model (LLM) embeds the task instruction and the field labels, and a small head returns the load from the frozen decoder output. No backbone parameter is updated at any point. At each of six rolling origins training sees only a seven-day record of 94 samples, while the non-overlapping test windows supply 7824 held-out hourly points across four seasons, five seeds each. NTSForecast attains the lowest MAE, RMSE and MAPE of the methods compared (5.66 ± 0.04% against 5.68% for the strongest baseline), but that margin is small and sensitive to how serial dependence is handled; under scarce history the framework is competitive with that baseline, not superior to it. What the evidence does establish is the contribution of the frozen backbone. With architecture, trainable-parameter count, hyperparameters and data held identical, removing it costs 1.38 percentage points of MAPE. Backbone scale runs against a capacity argument; on a 0.5B-to-8B ladder the smallest pretrained backbone is the most accurate and becomes the reference configuration (1445 MB peak memory). Removing the textual channel costs 1.17 percentage points.

Förderung

  • Science and Technology Project of State Grid Tianjin Electric Power Company

Zitieren

APA

Xin, X., Xu, J., Wang, X., Xu, W., Yue, Y., Wang, C., Yang, J., & Sun, B. (2026). A Numerical–Textual Dual-Channel Forecasting Framework for Short-Term Load Forecasting. Energies, 19(16), 3930. https://doi.org/10.3390/en19163930

Vancouver

Xin X, Xu J, Wang X, Xu W, Yue Y, Wang C, et al. A Numerical–Textual Dual-Channel Forecasting Framework for Short-Term Load Forecasting. Energies. 2026;19(16):3930. doi:10.3390/en19163930

BibTeX
@article{Xin2026,
  author = {Xin, Xiaohu and Xu, Jing and Wang, Xin and Xu, Weili and Yue, Yang and Wang, Chengshuai and Yang, Junwei and Sun, Bing},
  title = {A Numerical–Textual Dual-Channel Forecasting Framework for Short-Term Load Forecasting},
  journal = {Energies},
  year = {2026},
  volume = {19},
  number = {16},
  pages = {3930},
  doi = {10.3390/en19163930}
}
Kostenlos starten →