End-of-Life Prediction Models for Lithium-ion Batteries in Electric Vehicles: Approaches, Challenges and Future Directions

As the global transition toward electrification accelerates across the transportation and stationary energy storage sectors, the critical need for accurate end-of-life (EoL) prediction of lithium-ion batteries (LIBs) has become increasingly apparent. Current battery failures impose substantial costs on manufacturers through warranty claims, while creating significant safety risks that threaten both electric vehicle (EV) adoption and grid-scale energy storage deployment. This paper examines the modeling approaches to predict the EoL and the remaining useful life (RUL) of LIBs in EVs. The paper includes data-driven models, physics-based approaches, and hybrid frameworks. Through systematic analysis of recent advances, the paper identifies that hybrid models demonstrate superior performance compared to single-approach methods, effectively addressing the inherent limitations of individual methodologies across diverse operating conditions. Key challenges remain in Battery Management System (BMS) integration complexity, data quality constraints, and real-time computational requirements. The proposed review establishes that next-generation prediction systems and incorporates transfer learning, digital twin technologies, and second-life battery strategies to support sustainable EV adoption and circular economy principles.