A Review and Comparison of Lithium-Ion Battery SOC Estimation Methods for Electric Vehicles

In this research, a Thevenin battery model is proposed considering the hysteresis effect. A general comparison of model-based methods is followed by simulations with different State-of-Charge (SOC) estimation methods i.e., Coulomb Counting, Open Circuit Voltage, Extended Kalman Filter and Adaptive Extended Kalman Filter, Luenberger Observer and Neural Networks based on the same dataset. The results show that the proposed battery model can accurately provide SOC estimation and Adaptive Extended Kalman Filter can provide SOC estimation even with false initialization. Neural Networks deliver a slightly more accurate prediction. However, the requirement of large memory storage can be restricted for some applications.