Analysis of ECM-based Li-Ion Battery State and Parameter Estimation Accuracy in the Presence of OCV and Polarization Dynamics Modeling Errors

Abstract

The paper presents an analysis of a dual Kalman filter-based estimator of Li-Ion battery equivalent circuit model (ECM) states and parameters, which include state of charge (SoC), internal resistances and open-circuit voltage (OCV) vs. SoC curve parameters. Different estimator structures are analyzed in order to identify the causes and magnitudes of related estimation errors. First, a standard dual estimator relying on known, invariant OCV vs. SoC characteristic is analyzed. The OCV vs. SoC characteristic is then modeled using a parametric model whose parameters are identified off-line in one case and estimated in the other. Next, the ECM is modified to include the OCV hysteresis effect, while the estimator is upgraded with an adaptation mechanism to capture the hysteresis effect without compromising the overall estimation accuracy. Finally, the estimator sensitivity with respect to inaccurate parameterization of the polarization capacitance/time constant is analyzed and compared with the sensitivity to OCV modeling errors.