Multiscale Convolutional Reservoir Computing for Remaining Useful Life Prediction of Li-ion Batteries
IEEE Transactions on Consumer Electronics · 2025-05 · 71 (2) · 3350-3359
DOI: 10.1109/tce.2025.3570594 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Consumer Electronics, 2025-05. Verfasst von Chuan Li und 2 weiteren Autor:innen. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 33 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Chongqing Municipal Special Fund of Technological Innovation and Application Development, Chongqing Municipal Education Commission Project.
Abstract (Originalsprache)
Li-ion batteries are widely employed as power sources for consumer electronics. Accurate prediction of battery remaining useful life (RUL) is crucial for preventing system malfunctions and enabling predictive maintenance. This remains a challenging task because the battery degradation is complex and contains temporal dependencies in different time-scales. In this work, a novel backbone network referred to as multiscale convolutional reservoir computing (MCRC) is proposed to learn features in multiple time-scales for battery RUL prediction based on only previous capacity measurements. To this end, convolutional filters with different sizes are first used to extract multiscale temporal features within a fragment of previous capacity measurements. Multiple reservoir layers are then stacked to capture the multiscale sustained dependence of the extracted convolutional features. The present MCRC was evaluated on both a benchmark battery degradation dataset and in-house experimental data. It achieved root mean square errors of 211.27 and 70.01, mean absolute relative errors of 0.73 and 0.59, symmetry indices of 2.76 and 2.90, and $\alpha $ - $\lambda $ metrics of 0.25 and 0.46 in benchmark and in-house experiments, respectively. The results indicate that it can accurately predict RUL of Li-ion batteries, outperforming other state-of-the-art backbone networks.
Förderung
- National Natural Science Foundation of China
- Chongqing Municipal Special Fund of Technological Innovation and Application Development
- Chongqing Municipal Education Commission Project
- CTBU Starting Project
- New Chongqing Youth Innovative Talent Project
Zitieren
APA
Li, C., Guo, G., & Pu, Z. (2025). Multiscale Convolutional Reservoir Computing for Remaining Useful Life Prediction of Li-ion Batteries. IEEE Transactions on Consumer Electronics, 71(2), 3350-3359. https://doi.org/10.1109/tce.2025.3570594
Vancouver
Li C, Guo G, Pu Z. Multiscale Convolutional Reservoir Computing for Remaining Useful Life Prediction of Li-ion Batteries. IEEE Transactions on Consumer Electronics. 2025;71(2):3350-3359. doi:10.1109/tce.2025.3570594
BibTeX
@article{Li2025,
author = {Li, Chuan and Guo, Gaifang and Pu, Ziqiang},
title = {Multiscale Convolutional Reservoir Computing for Remaining Useful Life Prediction of Li-ion Batteries},
journal = {IEEE Transactions on Consumer Electronics},
year = {2025},
volume = {71},
number = {2},
pages = {3350--3359},
doi = {10.1109/tce.2025.3570594}
}
EUVIMED