The performance and aging behavior of a lithium-ion battery (LIB) are strongly influenced by inner processes as well as outer influence factors like the thermal boundary conditions.
A valuable and established tool to describe and gain a better understanding of the cell behavior is modelling using equivalent circuit models (ECM). For a deeper understanding and more detailed insights into the processes taking place inside the cell, physically based models, resolving the complex interaction of charge, heat and mass transfer as well as the electrochemical reaction by physical equations (such as Fick’s law and Nernst-Einstein-Equation) are needed. An equally well-established characterization method, that can provide experimental data on impedance as a representation of the integral multiphysical behavior of the cell is Electrochemical Impedance Spectroscopy (EIS). At the same time EIS allows to resolve the individual processes taking place inside the cell according to their time constants.

In this work, a model consisting of the combination of an equivalent circuit and a physical pseudo-2D (p2D) model according to Doyle-Fuller-Newman (DFN) was developed and implemented in COMSOL Multiphysics. With the help of this model, impedance spectra based on the multiphysical processes can be predicted and the underlying parameters can be systematically analyzed. By comparison of the impedance spectra from simulation with experi-mental data, the physical parameters determining the overall cell behavior can be identified. This inverse-parameterization approach can be accelerated significantly by adding least square optimization tool, fitting the simulated impedance spectra to an experimental data set by automatically optimizing the values of the physical parameters.

A challenging task by applying a DFN modelling approach combined with an optimization algorithm is to identify the value range of the crucial effective parameters for mass and charge transport, which are influenced by the microstructure of an electrode. Based on previous works on the sensitivity influence of each model parameter and the usage of an established hetero¬geneous 3D microstructure model, the range of the effective transport parameters was narrowed down in the following way: a 3D microstructure based on spherical and non-spherical particles is being generated with a drop and roll-algorithm, in which the particle size can be adjusted to replicate the particle distribution of real battery electrodes. The developed microstructure simulation model is then applied on these microstructures to analyze the impedance and voltage behavior as well as to derive crucial effective transport parameters like the tortuosity of the porous electrode. In this contribution the results of variations of the microstructural parameters like particle shape and size distribution are shown and the transfer from a heterogeneous 3D microstructure to a homogenized p2D model is evaluated.