Averaged Cell Expansion As a Predictor of State of Health of Lithium-Ion Pouch Cell

The rise of electric vehicles on the road and associated supply chain concerns draw researchers' attention to batteries' second life. For this purpose, non-destructive and economical methodologies for battery state of health (SoH) have been sought. Although it is known that LiB pouch cells expand as they degrade, how the expansion evolves over their lifecycle and how the location of measurement affects the discussion, are not well studied. Consequently, the method to accurately determine SoH from cell thickness is yet to be reported. This study uses digital image correlation to obtain cell expansion at every surface point and observe its non-uniformity. The evolution of the non-uniform expansion is observed during extended cycling until cells reach their end of life. We further find that the expansion averaged over the cell surface has a strong linear correlation with SoH, and develop a machine-learning-based method to estimate cell SoH accurately from the averaged expansion. Lastly, we verify that this methodology is effective when using the averaged expansion approximated only from several points of measurement, which can practically and economically be obtained by lining up displacement sensors, for example, at the battery repurposing factory.