A charge dynamics model for state-of-charge estimation by Kalman filter and lithium-ion battery parameters observation in real-vehicle

Abstract

State-of-charge estimation by Kalman filter requires a linear dynamics model of lithium-ion battery. This is particularly challenging for formulating a battery dynamics model by a set of linear differential equations, for which battery dynamic characteristics are nonlinear. While much of the research on battery dynamics models has focused on linearizing the battery model by some mathematical methods, comparatively little is known about how to directly model a linear differential equation containing some nonlinear characteristics. In this article, we are trying to build a linear charge dynamics model that consists of equivalent circuit model and voltage charge characteristics, based on which a Kalman filter is designed for state-of-charge estimation. The designed approach is deployed in a battery electric vehicle for some validations. According to some experimental results, it can be concluded that the charge dynamics model of lithium-ion battery is a good linear model for direct application of Kalman filter.