State-of-Charge Estimation for Batteries Based on the Nonlinear Double-Capacitor Model and Extended Kalman Filter

State-of-charge (SOC) estimation plays a foundational role in advanced battery management systems (BMS), having attracted much attention in the past decade. It is widely acknowledged that the accuracy of SOC estimation largely depends on the accuracy of the selected model. In this work, we contribute a new SOC estimation method based on the nonlinear double-capacitor (NDC) model, a novel equivalent circuit model distinctly capable of simulating the charge diffusion inside an electrode of a battery and capturing the battery's nonlinear voltage behavior simultaneously. With improved predictive accuracy, the NDC model provides a new opportunity for enabling more accurate SOC estimation. With this motivation, this paper exploits the well-known extended Kalman filter (EKF) to perform SOC estimation based on the NDC model. The EKF is desirable here as it leads to efficient computation, straightforward implementation, and good convergence in its application to the NDC model, which is low-dimensional and governed by linear dynamics along with nonlinear output. The proposed SOC estimation method is validated through simulations and experimental data under various conditions, showing significant accuracy as well as robustness to different levels of initialization error and noises.