A hybrid squeeze excitation gate recurrent unit-autoregressive integrated moving average model for long-term state of health estimation of lithium-ion batteries with adaptive enhancement ability.

Wu, Wenjie; Wang, Shunli; Fan, Yongcun; Liu, Donglei; Long, Gang; Fernandez, Carlos

Abstract

The accurate estimation of the state of health (SOH) of lithium-ion batteries is crucial for ensuring their safe operation and effective management. Since the data acquired from the lithium-ion battery aging experiment is abundant in electrochemical and dynamic information, helpful health indicators can be extracted for SOH predictions through machine learning. This paper proposes an innovative squeeze excitation gate recurrent unit-autoregressive integrated moving average (SEGRU-ARIMA) composite model designed explicitly for high-precision SOH estimation of lithium-ion batteries. The model employs SENet to extract essential features from charging and discharging data, while the GRU network adeptly processes time-series data to provide an initial SOH estimate. Furthermore, the ARIMA model is utilized to predict the discrepancy between the estimated and actual SOH values, with the final SOH result refined through linear weighting. The results demonstrate that, compared with other models, the SEGRU-ARIMA model exhibits superior estimation performance, with the absolute error range consistently remaining within 2.5 %. The mean absolute error (MAE), mean absolute percentage error (MAPE), and root mean square error (RMSE) were all <0.738 %, 1.018 %, and 0.899 %, respectively.

Citation

WU, W., WANG, S., FAN, Y., LIU, D., LONG, G. and FERNANDEZ, C. 2025. A hybrid squeeze excitation gate recurrent unit-autoregressive integrated moving average model for long-term state of health estimation of lithium-ion batteries with adaptive enhancement ability. Journal of energy storage [online], 131(part B), article number 117600. Available from: https://doi.org/10.1016/j.est.2025.117600

• Journal Article Type: Article
• Acceptance Date: Jun 28, 2025
• Online Publication Date: Jul 4, 2025
• Publication Date: Sep 30, 2025
• Deposit Date: Aug 15, 2025
• Publicly Available Date: Jul 5, 2026
• Journal: Journal of energy storage
• Print ISSN: 2352-152X
• Electronic ISSN: 2352-1538
• Publisher: Elsevier
• Peer Reviewed: Peer Reviewed
• Volume: 131
• Issue: part B
• Article Number: 117600
• DOI: https://doi.org/10.1016/j.est.2025.117600
• Keywords: Lithium-ion batteries; State of health; Data-driven; SEGRU model; ARIMA model
• Public URL: https://rgu-repository.worktribe.com/output/2923216