Data-driven Cycle-calendar Combined Battery Degradation Modeling for Grid Applications

Battery degradation is the main uncertainty that hedges the development of battery projects. In this work, we build a data-driven battery degradation model assessing the impact of the complex state of charge (SOC) operation condition. Both cycle life and calendar life are incorporated, based on the available lab testing data of battery cells. Various degradation modeling functions are compared to acquire the best fitting results under different depths of discharge (DOD) ranges calculated by rainflow counting algorithm. The statistical relation between shallow cycles and calendar time is used to address the calendar degradation in the cycling scope. A battery frequency regulation service case study is carried out based on a one-year frequency record in the Nordic synchronous area. Our work bridges the battery cell testing datasets to the battery degradation modeling in grid applications and proposes a new perspective to address the calendar life by the analysis of the shallow cycles for the state of health (SOH) estimation, which improves applicability and accuracy.