New semi-analytical model approach of the current distribution within parallel-connected lithium-ion cells

Abstract

Automotive batteries contain a high number of parallel- and serial-connected cells to fulfill the energy and power requirements of electric vehicles. Due to current, State of Charge (SoC) and temperature differences among the lithium-ion cells, caused by cell production and battery design tolerances, the power and energy of the battery is limited to the most stressed cell. In order to provide the best battery performance, regarding fast charging, recuperation and acceleration of the electric vehicle, the current distribution within the parallel-connected cells has to be estimated onboard. This article provides a semi-analytical model of the current distribution, with low parametrization and computational effort sufficient for an implementation in the Battery Management System (BMS). The model is validated by measurements of two parallel-connected cells with constant and dynamic current stress. The model output fit well to the experimental data with a related Root Mean Square Deviation (RMSD) of σRMSD≈0.5 % (constant stress) and of σRMSD≈1 % (dynamic stress). Further, the model mathematically describes the affects of the Open Circuit Voltage (OCV) bending on the current distribution among parallel cells, which closes an important lack of knowledge. This correlation explains the characteristic current cross at 0.2≤SoC≤0.45 and the peak of the current difference at the end of a constant discharge stress.