Simplified modeling approach for predicting the remaining useful life of EV batteries in second-life applications
Type of Publication
Year of Publication
Publisher
Publication Link
https://doi.org/10.1016/j.fub.2026.100204Repository Link
https://zenodo.org/records/22091352Authors
Citation
Montaru, M., Lonardoni, L., Serra, L., & Guillet, N. (2026). 'Simplified modeling approach for predicting the remaining useful life of EV batteries in second-life applications'. Future Batteries. doi:https://doi.org/10.1016/j.fub.2026.100204.
Short Summary
Two retired 23 kWh Renault Zoe battery packs, with initial states of health of 84% and 89%, were tested at module level under calendar aging (room temperature and 35°C, at 30% and 90% state of charge) and cycling aging (shallow, frequent cycling and deep, full cycling) conditions. The results show that storage at low charge and room temperature caused negligible degradation, while storage at high charge and elevated temperature drove capacity loss above 10% per year; degradation remained linear across all tested conditions, with no accelerated "knee point" behavior observed. Moreover, a four-parameter semi-empirical model combining calendar and cycling contributions was calibrated from this data and used to estimate Remaining Useful Life (RUL) for representative second-life applications, including frequency regulation and solar (PV) smoothing. The verification of these conclusions came from RUL estimates ranging from 7.6 to 14.6 years at 25°C depending on use case, a 40–58% reduction in RUL when operating temperature rose to 35°C, and a demonstrated 35% overestimation of lifespan when calendar aging effects are excluded from the model.
