A Comparative Study on the Performance of Battery Impedance Models in State-of-Charge Estimation

Lithium-ion battery (LIB) modeling can enhance accurate state estimation, whether through model-based or data-driven methods. This article explores the performance of various integer- and fractional-order Equivalent Circuit Models (ECMs), parameterized using Electrochemical Impedance Spectroscopy (EIS), with a focus on State-of-Charge (SoC) estimation accuracy. In this investigation, the parameters of various ECMs are employed as features in machine learning-based SoC estimation, and the results are compared in terms of estimation accuracy. Advantages and disadvantages of the utilized models are discussed and compared, considering all implementation factors, and the overall best model is concluded.