Improved pulse-fitting method for parameterization of a high-frequency equivalent circuit model for batteries

The importance of modelling is growing due to the ever-increasing integration of processing power and software into portable devices, EVs, and stationary energy storage systems. Modelling is a tool for efficiently operating and managing battery systems, including the simulation of electrical, thermal, and aging behavior. Nonetheless, state-of-the-art electrical modelling, particularly equivalent circuit models (ECMs) still commonly rely on 2nd order RC models, despite the fact that higher-order models offer increased fidelity and provide model parameters with physicial significance, granting them a meaningful physical interpretation that low-order models lack. Furthermore, the advancements in measurement devices over the years have significantly enhanced the ability to investigate high-frequency behavior. This paper introduces an advanced ECM that uses an improved pulse-fitting method for parameterization via an evolutionary algorithm. It provides an alternative to the conventional characterization approach, which uses electrochemical impedance spectroscopy for high-frequency behavior and pulse tests for long-term behavior, with a 2-in-1 solution that relies solely on pulse tests with sampling rates of up to 100 k H z to capture the complete behavior of the cell. Hence, the proposed method is a more cost- and time-efficient approach, particularly for product inspection and second-life qualification. This work emphasizes the significance of the sampling rate and corresponding model order, as both directly impact the projected bandwidth of the model. Although it seems intuitive that the sampling rate limits the fastest detectable process, this correlation is often overlooked. Therefore, this work derived a workflow for determining the optimal model order and showed that a sampling rate of at least 10 k H z is necessary to describe the complete high-frequency capacitive behavior and at least 100 k H z for the inductive behavior.