Parameter Identification of Fractional-order Model for Lithium-ion Batteries via a Neighborhood Differential Evolution Algorithm

Abstract

Lithium-ion batteries are currently one of the primary energy sources in the power system of new energy vehicles due to their high specific energy, long cycle life, no memory effect, and no pollution. To better use lithium-ion batteries, it is essential to establish a precise battery management system, which requires accurate parameter identification results. Therefore, a battery model with good robustness, accuracy, and low complexity is needed. In this paper, a fractional-order equivalent circuit model for lithiumion batteries is established based on the fractional-order calculus theory for the second-order RC integer-order equivalent circuit model. The parameters of the fractional-order equivalent circuit model are then identified using a neighborhood-based differential evolution algorithm. The proposed NDE algorithm is compared with several commonly used algorithms, and the experimental results are given. The results show that the NDE algorithm can better extract the optimal parameters of the fractional-order model of lithium batteries, and has higher accuracy.