EUVIMEDEUVIMED entdecken
Crossref

State of Health Estimation for Lithium‐Ion Batteries via Multisource Health Feature Fusion and Transformers

Pengbo Zhang, Ranjun Huang, Feiyu Ji, Zhikai Li, Jiangrong Bao, Changqing Du

Energy Technology · 2026-06 · 14 (6)

DOI: 10.1002/ente.70511 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Energy Technology, online erschienen am 12.06.2026. Verfasst von Pengbo Zhang und 5 weiteren Autor:innen. Die Arbeit nennt 41 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Accurate estimation of the state of health (SOH) of lithium‐ion batteries is crucial for enhancing the safety and reliability of battery management system (BMS). This article proposes a SOH estimation framework that integrates multisource feature fusion with a transformer‐based architecture to achieve effective redundancy reduction and accurate modeling of long‐term degradation dynamics. In practical BMS applications, the charging process provides a controlled and repeatable operating condition, under which degradation‐related characteristics can be more reliably observed. Accordingly, multiple health indicators are extracted from charging data, including voltage–time curves, incremental capacity (IC) curves, and differential temperature (DT) curves, enabling a comprehensive characterization of battery degradation from electrical, electrochemical, and thermal perspectives. To alleviate feature redundancy and enhance information utilization efficiency, an autoencoder (AE) is first employed to compress and refine the high‐dimensional feature space. The resulting compact latent representations are then fed into a transformer network to effectively capture the long‐term temporal dependencies inherent in battery degradation trajectories. Experimental results on both the Oxford and NASA datasets demonstrate that the proposed AE–transformer framework achieves superior SOH prediction accuracy, with mean absolute error (MAE) and root mean square error (RMSE) consistently below 1.1%. Compared with several benchmark models, the proposed method yields performance improvements exceeding 10% in terms of MAE and RMSE.

Zitieren

APA

Zhang, P., Huang, R., Ji, F., Li, Z., Bao, J., & Du, C. (2026). State of Health Estimation for Lithium‐Ion Batteries via Multisource Health Feature Fusion and Transformers. Energy Technology, 14(6). https://doi.org/10.1002/ente.70511

Vancouver

Zhang P, Huang R, Ji F, Li Z, Bao J, Du C. State of Health Estimation for Lithium‐Ion Batteries via Multisource Health Feature Fusion and Transformers. Energy Technology. 2026;14(6). doi:10.1002/ente.70511

BibTeX
@article{Zhang2026,
  author = {Zhang, Pengbo and Huang, Ranjun and Ji, Feiyu and Li, Zhikai and Bao, Jiangrong and Du, Changqing},
  title = {State of Health Estimation for Lithium‐Ion Batteries via Multisource Health Feature Fusion and Transformers},
  journal = {Energy Technology},
  year = {2026},
  volume = {14},
  number = {6},
  doi = {10.1002/ente.70511}
}
Kostenlos starten →