Advanced State of Charge Estimation for Electric Vehicles Using Novel Pi and T Battery Equivalent Circuit Models

Accurate state of charge (SoC) estimation requires mathematical models that consider individual user usage patterns to ensure optimal performance of lithium-ion battery (LiB) systems. Accurate SoC estimation improves battery life, driving comfort and avoids failure. In this paper, Pi and T equivalent circuit models were developed and validated with the 1RC and 2RC Thevenin models. Simulations were carried out to test their behavior under different rates of charging and discharging (1C, 2C, and 3C). The proposed models were validated by using real data from battery tests. At high C-rates, the 1RC and 2RC models show large variations, but the Pi model provided the best agreement with the experimental data for all rates, and the T model provided the second-best fit. In this paper, the impact of battery aging and degradation on the accuracy of the Extended Kalman Filter (EKF) and Coulomb Counting (CC) methods for estimating the SoC was studied, emphasizing the need to include them in the models. The results emphasized the importance of advanced modeling approaches for efficient battery management and point to directions for future research.