Systematic Comparison of Active Balancing: A Model-Based Quantitative Analysis

Abstract

A quantitative model-based comparison of ten active balancing circuits is presented for equalizing imbalanced cell energies of lithium-ion batteries. A mean current model approach for active balancing is used to describe energy transfers between battery cells for long charge/discharge operations. A Thevenin model for 96 battery cells is used as part of a mid-size electric vehicle. In quasi-static simulations, FTP75 driving cycles are continuously repeated until the first cell is discharged from initial full charge. Simulation results are compared for cases of new and aged cells with different capacity distributions concerning equalization speed and battery efficiency. The influence of control algorithms on results is investigated by comparing control approaches with voltage, state of charge, and charge capacities as control variables. The results of state of charge and charge capacities show their advantage over voltage-based balancing. The results of two detailed performance analysis for two specific cell distributions are confirmed by simulating two sets of 100 random cell charge capacity distributions for each balancing method and each control variable.