EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Short-term power load forecasting based on the CEEMDAN-TCN-ESN model

Jiacheng Huang, Xiaowen Zhang, Xuchu Jiang iD

PLOS ONE · 26.10.2023 · 18 (10) · e0284604

DOI: 10.1371/journal.pone.0284604 ↗

Kurz gefasst

Fachartikel in der Zeitschrift PLOS ONE, online erschienen am 26.10.2023. Verfasst von Jiacheng Huang und 2 weiteren Autor:innen. Beteiligte Einrichtung: Zhongnan University of Economics and Law, Wuhan, China. Hauptthemen laut MEDLINE: Economic Development, Electric Power Supplies. Die Arbeit wurde bisher 12-mal zitiert (Crossref) und nennt 23 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: The Open-funding Project of State Key Laboratory of Intelligent Manufacturing Equipment and Technology.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Ensuring an adequate electric power supply while minimizing redundant generation is the main objective of power load forecasting, as this is essential for the power system to operate efficiently. Therefore, accurate power load forecasting is of great significance to save social resources and promote economic development. In the current study, a hybrid CEEMDAN-TCN-ESN forecasting model based on complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) and higher-frequency and lower-frequency component reconstruction is proposed for short-term load forecasting research. In this paper, we select the historical national electricity load data of Panama as the research subject and make hourly forecasts of its electricity load data. The results show that the RMSE and MAE predicted by the CEEMDAN-TCN-ESN model on this dataset are 15.081 and 10.944, respectively, and R 2 is 0.994. Compared to the second-best model (CEEMDAN-TCN), the RMSE is reduced by 9.52%, and the MAE is reduced by 17.39%. The hybrid model proposed in this paper effectively extracts the complex features of short-term power load data and successfully merges subseries according to certain similar features. It learns the complex and varying features of higher-frequency series and the obvious regularity of the lower-frequency-trend series well, which could be applicable to real-world short-term power load forecasting work.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Economic DevelopmentElectric Power SuppliesElectricityLearningPanamaForecasting

Autor:innen und Institutionen

  • Zhongnan University of Economics and Law, Wuhan, China.
  • State Key Laboratory of Intelligent Manufacturing Equipment and Technology, Wuhan, China.

ORCID-Profile: Xuchu Jiang

Förderung

  • The Open-funding Project of State Key Laboratory of Intelligent Manufacturing Equipment and Technology

Zitieren

APA

Huang, J., Zhang, X., & Jiang, X. (2023). Short-term power load forecasting based on the CEEMDAN-TCN-ESN model. PLOS ONE, 18(10), e0284604. https://doi.org/10.1371/journal.pone.0284604

Vancouver

Huang J, Zhang X, Jiang X. Short-term power load forecasting based on the CEEMDAN-TCN-ESN model. PLOS ONE. 2023;18(10):e0284604. doi:10.1371/journal.pone.0284604

BibTeX
@article{Huang2023,
  author = {Huang, Jiacheng and Zhang, Xiaowen and Jiang, Xuchu},
  title = {Short-term power load forecasting based on the CEEMDAN-TCN-ESN model},
  journal = {PLOS ONE},
  year = {2023},
  volume = {18},
  number = {10},
  pages = {e0284604},
  doi = {10.1371/journal.pone.0284604}
}
Kostenlos starten →