Second-life Lithium-ion Batteries: AI-based Prediction of Remaining Useful Life

Imagine a world where the journey of a lithium-ion battery doesn’t end when it leaves an electric vehicle. Instead, it initiates a second chapter—serving as a stationary storage backup in solar
systems, homes, and industries. These batteries, though no longer fit for high-demand tasks, still hold significant energy potential. But unlocking this potential safely and efficiently hinges on one critical question: How much life is left in them?

To address this, researchers have leveraged the power of AI/ML. In this study, advanced surrogate models were developed to predict the remaining useful second-life (RU2L) of repurposed batteries. These models act like digital twins, learning from publicly available degradation data to understand how batteries age over time. By capturing intricate patterns in charge and discharge cycles, the models can forecast how many cycles are left before a battery reaches its true end-of-life (EoL).

The results are promising. With high prediction accuracy, these machine learning tools offer a reliable way to manage second-life batteries—ensuring they’re used wisely, safely, and sustainably. It’s a step forward in reducing electronic waste, lowering energy storage costs, and making smarter use of the resources we already have.