Accelerated State-of-Health Estimation for Battery Recycling, using Neural Networks

The recycling of lithium-ion batteries gains more and more importance due to an increased amount of electric vehicles. Not all cells in a used battery have already reached their end of life criterion and can still be employed for second life applications. However, the characterization of returned cells is often more expensive than the economic benefit of further usage. This paper proposes an accelerated state-of-health estimation by employing neural networks in the recycling process. An overall model incorporates the aging prediction and automatically decides if a cell should still be operated in first- or second-life applications or already reached the end of its life cycle. In this case, the battery cell will be recycled and the recovered electrode material can be added to the cell manufacturing process. Therefore, this process reduces the initial production costs as well as the negative environmental impact.