Estimation of battery parameters of the equivalent circuit models using meta-heuristic techniques

This paper describes the parametrization and comparison of dynamic battery models for Li-ion battery. The estimation of battery parameters deploys experimental methods that are expensive, require high computational power and are time-consuming. Hence, two commonly used equivalent circuit battery models are tested using GA, PSO, ASMO and DE optimization techniques. Estimation has been done by the estimated voltage curves closeness to the known catalog voltage curve. Feasibility of various optimization techniques is evaluated by the accuracy of predicted model and the rate of convergence in predicting the model parameters. Investigation showed that DE algorithm has the best accuracy among meta-heuristic optimizers for battery parameter estimation for first order model while ASMO has best accuracy for the second order model. Further analysis showed that for both the models, DE algorithm was reliable as well as computationally less expensive compared to other optimization techniques. Also the second order RC model proved out to be more robust as for almost all scenarios the performance of optimization algorithm improved by use of this model.