Analysis of the number of replicates required for Li-ion battery degradation testing

Abstract

Aging prediction of Lithium Ion Batteries is of major importance for assessing both longevity and sustainability of any battery system. In addition to the aging itself, aging trajectories are also dependent on the cell-to-cell variability that is caused by production tolerances. To be able to accurately model and predict the aging of battery systems, researchers and manufacturers must thus take the cell-to-cell variability into account when modelling battery aging. This paper contributes to the methodology for including cell-to-cell variability in aging testing by generating empirical aging data for a large number of replicates of commercial battery cells and assessing prediction stability. The conclusion from several different methods of evaluation is that a minimum of 4 replicates is required to accurately capture cell-to-cell variability in aging testing and modelling. The typical variance for the tested cells was about 10% of the capacity lost at any given point in testing.