AI-based methodology for estimating the remaining useful life (RUL) of second-life lithium-ion cells coming from EVs, to be reused in stationary ESS

This project aims to develop an AI-based methodology for the accelerated and accurate estimation of RUL in second-life lithium-ion cells. By using data-driven approaches, we intend to significantly reduce the time and cost of qualification processes and improve the efficiency of cell sorting and matching. Additionally, the project will explore modular ESS architectures with swappable cell units, enhancing the adaptability of second-life systems and maximizing the utility of available battery inventories.
The core of the solution will be machine learning models trained on operational and diagnostic data from used EV battery cells. These models will learn to identify patterns linked to degradation and performance decline, enabling precise RUL predictions without the need for long-term cycling tests. The approach will prioritize explainability and integration potential, so that outputs can support both automated decision-making and expert validation during repurposing. Over time, the system will improve its accuracy as more data becomes available from real-world deployments.