Optimal parameter identification of linear and non-linear models for Li-Ion Battery Cells

Abstract

This study proposes a reduced model based on the state space representation for identifying an accurate electric equivalent circuit of Lithium-Polymer Battery Cells. The parameter extraction process is formulated as non-linear optimization problem via three-stage procedure. The first stage estimates the state of charge (SoC) based on the non-linear characteristics associated with the battery current and the initial SoC condition. In the second stage, the open circuit voltage is estimated in terms of the resulted SoC that is employed in the first stage with varied linear and non-linear models. In the third stage, an Equilibrium Algorithm (EA), a recent optimizer, is developed for optimally identifying the battery parameters. The EA’s parameters are adjusted based on Taguchi’s design of experiment approach to reduce the computational time as well as the number of experiments that are required to get the optimum possible parameter arrangement Numerical simulations associated with experimental implementation are emulated on Li-Ion Battery to prove the high capability of the proposed EA an as efficient identification procedure. In Addition, the proposed EA is characterized with high accuracy compared with several recent optimization algorithms for ARTEMIS driving cycle profile. The solution quality improvement of the proposed reduced model is achieved with high closeness to the experimental measurements for battery voltage and SoC. Furthermore, 16 % less computational times 12 % more accuracy are obtained by the proposed reduced model compared with linear and non-linear models.