Citation

Guo, Weining and Tan, Ai Hui and Ong, Duu Sheng (2025) Optimized prediction of remaining useful life of lithium-ion batteries: A voltage-current behavior analysis for enhanced health monitoring. Journal of Energy Storage, 134. p. 118138. ISSN 2352-152X

Text Optimized prediction of remaining useful life of lithium-ion batteries_ A voltage-current behavior analysis for enhanced health monitoring.pdf - Published Version Restricted to Repository staff only Download (12MB)

Abstract

Accurate forecasting of the Remaining Useful Life (RUL) of lithium-ion batteries is essential for efficient battery management, cost-effective maintenance, and enhanced safety in electric vehicles, where reliable performance over the vehicle's lifetime is crucial. Because of the nonlinear and dynamic nature of battery aging, the devel opment of a reliable predictive model remains a significant challenge. To address these challenges, a novel Coyote-Badger Optimization Extended Long Short-Term Memory (CBO-xLSTM) method is proposed to enhance RUL prediction. A key aspect of this study is the utilization of xLSTM for RUL prediction, leveraging its enhanced memory gating mechanisms to better capture long-range dependencies in battery degradation data. Additionally, the CBO algorithm adaptively tunes the most critical model hyperparameters of the xLSTM model, which ensures adaptability in diverse battery degradation trends. The model was evaluated on eight battery cells from the Oxford dataset, where several features were extracted from voltage, current, temperature, and capacity profiles to represent degradation patterns. The results demonstrated that CBO-xLSTM consistently outperformed benchmark models with R2 values above 0.990. Moreover, the proposed model can generalize well to different battery chemistries and operational scenarios, based on the observed performance across NASA and MIT data sets. The findings validate the effectiveness of integrating sophisticated sequence modeling and adaptive opti mization techniques in improving the accuracy of RUL predictions. The CBO-xLSTM method can present an effective solution for battery health monitoring and energy management systems toward efficient battery use and sustainable energy storage systems

Item Type:	Article
Uncontrolled Keywords:	Electric vehicle
Subjects:	T Technology > TL Motor vehicles. Aeronautics. Astronautics > TL1-484 Motor vehicles. Cycles
Divisions:	Faculty of Artificial Intelligence & Engineering (FAIE)
Depositing User:	Ms Rosnani Abd Wahab
Date Deposited:	30 Sep 2025 07:27
Last Modified:	05 Oct 2025 11:24
URII:	http://shdl.mmu.edu.my/id/eprint/14595

Downloads

Edit (login required)