EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Autonomous Classification and Prediction of Degradation Transitions in Lithium-Ion Batteries

Yixin Nie, Weiyu Li

Journal of The Electrochemical Society · 12.06.2026 · 173 (11) · 110504

DOI: 10.1149/1945-7111/ae7311 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of The Electrochemical Society, online erschienen am 03.06.2026. Verfasst von Yixin Nie und 1 weiteren Autor:innen. Die Arbeit nennt 41 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Lithium-ion batteries (LIBs) often exhibit a two-stage degradation behavior, transitioning from gradual capacity fade to accelerated aging at the degradation knee point. Accurate and timely assessment of a battery’s degradation stage relative to this knee is critical for battery management. However, most existing degradation-stage classification and knee-proximity prediction approaches rely on multi-cycle historical data, high-dimensional time-series inputs, or complex deep learning architectures, and are typically evaluated under a single knee definition, limiting interpretability and practical deployment. We propose a unified and practically deployable framework for degradation-stage classification and knee-distance prediction using information from a single charging cycle. The framework extracts a compact set of physically meaningful current-based features from the constant-voltage charging segment and employs gradient-boosted decision and regression tree models to determine whether a battery operates in the pre-knee or post-knee regime and to estimate its remaining cycle distance to the knee point. The use of low-dimensional, physically motivated features enables degradation assessment without reliance on long-term cycling history. The proposed approach is evaluated on a benchmark LIB aging dataset under multiple representative knee identification methods and compared against state-of-the-art classification and prediction models. Results demonstrate competitive classification accuracy and low knee-distance prediction error across knee definitions, while requiring substantially reduced data, sensing, and computational complexity, making the framework well suited for real-time and scalable battery management applications.

Zitieren

APA

Nie, Y., & Li, W. (2026). Autonomous Classification and Prediction of Degradation Transitions in Lithium-Ion Batteries. Journal of The Electrochemical Society, 173(11), 110504. https://doi.org/10.1149/1945-7111/ae7311

Vancouver

Nie Y, Li W. Autonomous Classification and Prediction of Degradation Transitions in Lithium-Ion Batteries. Journal of The Electrochemical Society. 2026;173(11):110504. doi:10.1149/1945-7111/ae7311

BibTeX
@article{Nie2026,
  author = {Nie, Yixin and Li, Weiyu},
  title = {Autonomous Classification and Prediction of Degradation Transitions in Lithium-Ion Batteries},
  journal = {Journal of The Electrochemical Society},
  year = {2026},
  volume = {173},
  number = {11},
  pages = {110504},
  doi = {10.1149/1945-7111/ae7311}
}
Kostenlos starten →