Modelling of cells’ capacity distribution and fading for lithium-ion battery pack during the charge and discharge cycles

Accurate and efficient prediction of pack-level capacity distribution and fading within lithium-ion battery packs is critical for state of health (SOH) and remaining useful life (RUL) assessment. However, different aging mechanisms and rates arising from intrinsic cell-to-cell inconsistencies and interactions complicate reliable distribution forecasting. This study proposes a modelling method based on the generalized extreme value (GEV) distribution, which offers superior generalization capability compared to single-component distributions. Using single-cell models validated with the UCL dataset, we generate comprehensive simulation data under 35 distinct conditions. Subsequently, the dataset is utilized to verify the advantages of the GEV distribution in characterizing capacity distributions and to establish formulas of the distribution parameters. The validation results show that the proposed model achieves high accuracy (MAPE ≈ 2.91 %) and computational efficiency in describing the capacity distribution of battery packs. The derived formulas intuitively reveal the relationships between pack operational/initial state parameters and key capacity distribution metrics including skewness (asymmetry), dispersion, and overall degradation level, supporting optimized battery pack management.