Battery parameterisation based on differential evolution via a boundary evolution strategy

Attention has been given to the battery modelling in the electric engineering field following the current development of renewable energy and electrification of transportation. The establishment of the
equivalent circuit model of the battery requires data preparation and parameterisation. Besides, as the equivalent circuit model is an abstract map of the battery electric characteristics, the determination of the possible ranges of parameters can be a challenging task. In this paper, an efficient yet easy to implement method is proposed to parameterise the equivalent circuit model of batteries utilising the
advances of evolutionary algorithms (EAs). Differential evolution (DE) is selected and modified to parameterise an equivalent circuit model of lithium-ion batteries. A boundary evolution strategy (BES) is
developed and incorporated into the DE to update the parameter boundaries during the parameterisation. The method can parameterise the model without extensive data preparation. In addition, the
approach can also estimate the initial SOC and the available capacity. The efficiency of the approach is verified through two battery packs, one is an 8-cell battery module and one from an electrical vehicle.