Collaborative SOH Estimation for Lithium-Ion Batteries by Integrating Both Charging and Discharging Profiles
IEEE Transactions on Transportation Electrification · 2026-08 · 12 (4) · 6626-6636
DOI: 10.1109/tte.2026.3680995 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Transportation Electrification, 2026-08. Verfasst von Zhijun Liu und 4 weiteren Autor:innen. Die Arbeit nennt 36 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Natural Science Foundation of Hunan Province.
Abstract (Originalsprache)
Accurate estimation of the state of health (SOH) for lithium-ion batteries (LIBs) is crucial to ensure the safe and stable operation of electric vehicles. Existing studies have mostly relied solely on either charging or discharging data, with few attempts to integrate both charging and discharging data. This reliance limits the models’ generalization capability, making it difficult for them to adapt to the complex and variable operating conditions encountered in the real world. To address this limitation, this article proposes a multidomain collaborative learning method for battery SOH estimation by integrating both charging and discharging domains. First, a parallel feature extraction network is employed to extract battery aging features from the charging and discharging domains, respectively. Second, the maximum mean discrepancy (MMD) method is introduced to reduce feature distribution discrepancies between charging and discharging domains, thereby alleviating the feature drift problem caused by data distribution differences. Finally, a cross-domain feature fusion mechanism is designed to dynamically adjust the feature weights from different domains. The experimental results demonstrate that the proposed method achieves outstanding performance in real-world scenarios, with an $R^{2}$ reaching 0.9141. Compared to methods relying solely on the charging domain or discharging domain, the $R^{2}$ value is improved by 14.4% and 122.6%, respectively. Moreover, the mean absolute error (MAE) and root-mean-square error (RMSE) are as low as 0.0191 and 0.0248, respectively.
Förderung
- National Natural Science Foundation of China
- Natural Science Foundation of Hunan Province
Zitieren
APA
Liu, Z., Jiang, F., Wu, Y., Liu, W., & Li, H. (2026). Collaborative SOH Estimation for Lithium-Ion Batteries by Integrating Both Charging and Discharging Profiles. IEEE Transactions on Transportation Electrification, 12(4), 6626-6636. https://doi.org/10.1109/tte.2026.3680995
Vancouver
Liu Z, Jiang F, Wu Y, Liu W, Li H. Collaborative SOH Estimation for Lithium-Ion Batteries by Integrating Both Charging and Discharging Profiles. IEEE Transactions on Transportation Electrification. 2026;12(4):6626-6636. doi:10.1109/tte.2026.3680995
BibTeX
@article{Liu2026,
author = {Liu, Zhijun and Jiang, Fu and Wu, Yue and Liu, Weirong and Li, Heng},
title = {Collaborative SOH Estimation for Lithium-Ion Batteries by Integrating Both Charging and Discharging Profiles},
journal = {IEEE Transactions on Transportation Electrification},
year = {2026},
volume = {12},
number = {4},
pages = {6626--6636},
doi = {10.1109/tte.2026.3680995}
}
EUVIMED