Unlocking the full potential of electric vehicle fast-charging over lifetime through model-based aging adaptation

Abstract

Fast-charging within 15min is increasingly demanded in electric vehicles to reduce downtime and increase customer acceptance. To ensure this, battery models depicting internal states can be used to safely operate along the batteries’ physical boundaries. This, however, poses an increasing risk of model error as the battery degrades over its lifetime, requiring aging-based derating and model updating. Conventional model-updating strategies based on capacity measurements are inaccurate, as they neglect the individual degradation paths of the electrodes and lithium inventory. In this study, we utilize half-cell open circuit potential measurements to update an electrochemical battery model using electrode and lithium inventory aging information over lifetime, showing that conventional methods may underestimate fast-charging potential. The average model error can be reduced by 66% for a publically available aging dataset compared to no model updating and by 33% compared to full cell capacity derating. Furthermore, we predict that charging speeds can be increased by up to 34% for strongly aged cells compared to full cell derating, as the latter neglects the loss of lithium inventory, which reduces the risk of critical anode potentials causing lithium plating.