EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

An approach of state-of-charge prediction for lithium-ion batteries using hybrid Pyraformer-LSTM learning network

Xianbin Wang, Liping Zhang, Qihua Fan, Yiyin Hu, Jian Zhang, Huimin Chen iD

PLOS One · 24.07.2026 · 21 (7) · e0353298

DOI: 10.1371/journal.pone.0353298 ↗

Kurz gefasst

Fachartikel in der Zeitschrift PLOS One, online erschienen am 24.07.2026. Verfasst von Xianbin Wang und 5 weiteren Autor:innen. Beteiligte Einrichtung: Institute of General Aviation Industry, Fujian Chuanzheng Communications College, Fuzhou, China. Hauptthemen laut MEDLINE: Lithium, Electric Power Supplies. Die Arbeit nennt 35 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Xinjiang Key Laboratory of Intelligent Control and Safety Assurance Technology for Low-Altitude Airspace.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Accurate prediction of the State of Charge (SOC) of lithium-ion batteries remains difficult under complex operating conditions. This difficulty arises from strong nonlinearity and long-term temporal dependency in battery data. In this study, we developed a hybrid prediction model that combines a pyramidal attention network with a Long Short-Term Memory (LSTM) network. The key innovation of this work lies in the integration of a pyramidal attention mechanism with a multi-scale sliding window design, enabling hierarchical extraction of temporal features ranging from local dynamic variations to long-term evolutionary trends. Specifically, the pyramidal attention module captures multi-granularity temporal dependencies through dynamic weighted fusion of features from different window scales (8, 16, and 32), while the LSTM network captures nonlinear dynamics and preserves long-term dependencies through gated temporal modeling. We evaluated the proposed method using battery datasets collected under multiple temperatures. We compared its performance with six representative SOC estimation methods. The experimental results show that the proposed model achieves RMSE and MAE of 1.8554% and 1.3021% on the test set, significantly outperforming baselines, demonstrating high prediction accuracy and validating the effectiveness of multi-scale attention and recurrent structures in SOC prediction.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

LithiumElectric Power SuppliesLong Short Term MemoryPrediction AlgorithmsPredictive Learning ModelsNeural Networks, Computer

Substanzen

Lithium

Autor:innen und Institutionen

  • Institute of General Aviation Industry, Fujian Chuanzheng Communications College, Fuzhou, China.
  • School of Aerospace Engineering, Xiamen University, Xiamen, China.
  • School of Aeronautics, Changji University, Changji, China.

ORCID-Profile: Huimin Chen

Förderung

  • Xinjiang Key Laboratory of Intelligent Control and Safety Assurance Technology for Low-Altitude Airspace

Zitieren

APA

Wang, X., Zhang, L., Fan, Q., Hu, Y., Zhang, J., & Chen, H. (2026). An approach of state-of-charge prediction for lithium-ion batteries using hybrid Pyraformer-LSTM learning network. PLOS One, 21(7), e0353298. https://doi.org/10.1371/journal.pone.0353298

Vancouver

Wang X, Zhang L, Fan Q, Hu Y, Zhang J, Chen H. An approach of state-of-charge prediction for lithium-ion batteries using hybrid Pyraformer-LSTM learning network. PLOS One. 2026;21(7):e0353298. doi:10.1371/journal.pone.0353298

BibTeX
@article{Wang2026,
  author = {Wang, Xianbin and Zhang, Liping and Fan, Qihua and Hu, Yiyin and Zhang, Jian and Chen, Huimin},
  title = {An approach of state-of-charge prediction for lithium-ion batteries using hybrid Pyraformer-LSTM learning network},
  journal = {PLOS One},
  year = {2026},
  volume = {21},
  number = {7},
  pages = {e0353298},
  doi = {10.1371/journal.pone.0353298}
}
Kostenlos starten →