Research Information System
Journal of Energy Storage, vol.131, 2025 (SCI-Expanded)
This paper presents the parameter extraction of lithium-ion battery (LiB) using the escape algorithm (ESC). Since LiBs are inherently nonlinear systems, estimating their equivalent circuit model (ECM) parameters is a challenging optimization problem. To solve the problem in an unparalleled way, the most recent metaheuristic algorithms that remain untested so far in parameter estimation of LiB are selected: ESC, arctic puffin optimization (APO), educational competition optimizer (ECO), eel and grouper optimizer (EGO), flood algorithm (FLA), horned lizard optimization algorithm (HLOA), hiking optimization algorithm (HOA), moss growth optimization (MGO) and polar lights optimizer (PLO). Additionally, weighted mean of vector (INFO), particle swarm optimization (PSO) and walrus optimizer (WO) used in literature, are included for validation. The algorithms are tested on 24 datasets from DST, BJDST, FUDS, and US06. The results of 30 runs reveal that ESC showed the best performance in parameter estimation regarding RMSE evaluation metric at mean (5.4648057E−03), maximum (6.4903342E−03) and standard deviation (3.8224940E−03). Especially, ESC has the lowest standard deviation value on 24 datasets, highlighting the tendency of the ESC to produce consistently low RMSE and its high accuracy. Although it could not provide the lowest value in terms of RMSE evaluation metric at minimum (4.2719688E−03), this does not affect the overall success of the algorithm. Moreover, the computational accuracy of the algorithms is evaluated by Friedman and Wilcoxon signed-rank tests, and the ESC is statistically verified to exhibit superior performance compared to others. These findings indicate that the ESC is a robust method for LiB parameter estimation.