EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A Hybrid CNN–BiLSTM–GRU Model for Electric Energy Consumption Forecasting

Luciano Roberto da Silva Leal, Byron Leite Dantas Bezerra, João Fausto Lorenzato de Oliveira, Carlos Eduardo Lima da Costa

International Transactions on Electrical Energy Systems · 2026-01 · 2026 (1)

DOI: 10.1155/etep/4831304 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Transactions on Electrical Energy Systems, online erschienen am 21.04.2026. Verfasst von Luciano Roberto da Silva Leal und 3 weiteren Autor:innen. Die Arbeit nennt 48 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Accurate forecasting of electricity consumption (PEC) is crucial for ensuring efficient energy system planning and maintaining a balance between supply and demand, particularly in countries with extensive continental dimensions and socioeconomic heterogeneity, such as Brazil. This study proposes a hybrid sequential combination of convolutional neural network, bidirectional long short‐term memory, and gated recurrent unit (CNN–BiLSTM–GRU) model to overcome forecasting challenges in regions with distinct consumption patterns. Using data shared by the National System Operator (ONS), the model’s performance was compared to that of the traditional statistical method (ARIMA) and other machine learning approaches, including random forest and support vector regression. Compared to competing models, the CNN–BiLSTM–GRU demonstrated superior accuracy, achieving the lowest mean absolute error (MAE) values across all analyzed regions: North (0.0368), Northeast (0.0348), South (0.0503), and Southeast (0.0336). Statistical validation using the Friedman–Nemenyi test confirmed its robustness against regional disparities—a critical limitation of competing models, which struggle to generalize across different patterns. The results highlight the unique ability of CNN–BiLSTM–GRU to capture spatiotemporal dependencies in complex time series, offering a scalable solution for energy system management in diverse scenarios.

Zitieren

APA

Leal, L. R. D. S., Bezerra, B. L. D., Oliveira, J. F. L. D., & Costa, C. E. L. D. (2026). A Hybrid CNN–BiLSTM–GRU Model for Electric Energy Consumption Forecasting. International Transactions on Electrical Energy Systems, 2026(1). https://doi.org/10.1155/etep/4831304

Vancouver

Leal LRDS, Bezerra BLD, Oliveira JFLD, Costa CELD. A Hybrid CNN–BiLSTM–GRU Model for Electric Energy Consumption Forecasting. International Transactions on Electrical Energy Systems. 2026;2026(1). doi:10.1155/etep/4831304

BibTeX
@article{Leal2026,
  author = {Leal, Luciano Roberto da Silva and Bezerra, Byron Leite Dantas and Oliveira, João Fausto Lorenzato de and Costa, Carlos Eduardo Lima da},
  title = {A Hybrid CNN–BiLSTM–GRU Model for Electric Energy Consumption Forecasting},
  journal = {International Transactions on Electrical Energy Systems},
  year = {2026},
  volume = {2026},
  number = {1},
  doi = {10.1155/etep/4831304}
}
Kostenlos starten →