Search

This paper presents an improved scheme for the state-of-health (SOH) estimation, applicable to battery management system (BMS) of electric vehicles. The original scheme requires the prior information for the estimation, which is the prior SOH identified from the last charging. This limit implies that if a battery or its BMS is replaced, the prior SOH stored within the BMS no longer matches with the actual SOH, resulting in a critical error. To avoid this potential but critical pitfall, we newly devise an improved SOH estimation scheme. The original scheme is revised by adopting the fixed-point iteration method into its parameter estimation. By removing dependencies on the prior information, the revised scheme can function regardless of such replacements. The revised scheme is experimentally validated and demonstrated that even without the prior information, it can satisfy the requirement of the SOH estimation (within 3%) thanks to its improved design. [ABSTRACT FROM AUTHOR]

Abstract: This paper presents a comparative study of twelve equivalent circuit models for Li-ion batteries. These twelve models were selected from state-of-the-art lumped models reported in the literature. The test data used is obtained from a battery test system with a climate chamber. The test schedule is designed to measure key cell attributes under highly dynamical excitations. The datasets were collected from two types of Li-ion cells under three different temperatures. The multi-swarm particle swarm optimization algorithm is used to identify the optimal model parameters for the two types of Li-ion cells. The usefulness of these models is then studied through a comprehensive evaluation by examining model complexity, model accuracy, and robustness of the model by applying the model to datasets obtained from other cells of the same chemistry type. [Copyright &y& Elsevier]