Hysteresis Modeling of Lithium-Silicon Half Cells Using Prandtl-Ishlinskii Model

Abstract

State-of-charge (SOC)-estimation accuracy is critical to battery management in electric-vehicle (EV) applications. SOC-estimation error can arise from unmodeled nonlinear battery hysteresis; so to enhance performance, it is necessary to model hysteresis voltage. This paper presents results from applying three kinds of Prandtl-Ishlinskii (PI) model to a Li//Si half cell: classical PI (CPI), generalized PI (GPI), and modified generalized PI (mGPI). The performance comparison among these models and also comparisons to the Preisach and Krasnosel'skii-Pokrovskii (KP) models are made. It is found that the mGPI provides best performance among all models having the same level of calculation.