Extended kalman filter for accurate state of charge estimation of lithium-based batteries: a comparative analysis

Nowadays, Lithium-based batteries are widely adopted in Electric Vehicles (EVs). The battery State-Of-Charge (SOC) represents its residual energy. However, it cannot be measured directly with sensors and must be estimated with mathematical techniques. The main objective of this paper is both to present two different battery parameter identification methods and to determine an accurate SOC estimation of Lithium-based batteries based on Extended Kalman Filter (EKF). The different methods of battery SOC estimation are comprehensively discussed. Then, the enhanced closed loop EKF estimator is proposed. For this model-based approach, a suitable model of the cell behavior is required to reach the most precise SOC estimation. Therefore, an accurate model of cell dynamics is necessary for applicable SOC estimation particularly for different current ratings and battery technologies. The hysteresis influence is considered in the Open Circuit Voltage (OCV) as a part of the battery modeling. Two battery parameters identification methods are compared in charging/discharging modes. Through the voltage drop measurement, all the parameters of such electric cell model are identified. Then, the EKF is experimentally implemented using C programming language in micro-controller (PIC18F4585 and PIC18F4685). Finally, the SOC estimation result of EKF is adequately correlated to that of the coulomb counting method with acceptable errors. The experimental results verify that the proposed method of Bo et al. (2015) provides enhanced SOC determination compared to that of Watrin (2013).