Efficient Parametrization of Electric Circuit-Based Battery Models: A Novel Approach Using Adaptive Polynomial Fitting

Battery systems are increasingly recognized as crucial components within both current and future energy systems. Thereby, a proper modelling of their dynamic behavior under different load conditions is of high importance. In this work, we present a workflow using the Electric Equivalent Circuit model with one RC-element (ECM), and a novel approach for its parametrization. As a first step, pulse discharge measurements on a set of LiFePO4-battery cells are conducted with a uniform discharge current and equal time-intervals. Based on the recorded voltage and current profile, discrete sets of the EEC components and corresponding State of Charge values are calculated for every single cell. An adaptive polynomial fitting method is applied to this discrete dataset, thereby generating a numerically optimal ECM model. The result is a mathematical model which is able to simulate the transient behavior of a random sample of a certain cell model. In comparison to other common parametrization approaches, the presented method has the advantage of having a low time effort for the generation of a large dataset, guaranteed numerical success, and applicability for an arbitrary battery type.