Systematic parameter studies are a powerful tool for gaining profound insights into the performance of microstructural cell designs. Loading simulations using a homogenized model in GeoDict 2023 enable this process efficiently.

Taking into account the influence that changes in microstructural design have on cell performance is crucial for the effective design of electrode materials at the microscale. Parameter studies in simulations prove to be an extremely valuable tool to accurately capture these effects. Various parameters, such as grain size and shape, are systematically varied during the process. Rapid simulations of the battery charging process are an indispensable requirement for such in-depth investigations.

Up until now, battery charging simulations with the BatteryDict module of GeoDict have relied on complete 3D resolution. Simulating an entire charging cycle using this method could take up to a week, depending on the size of the structure being analyzed. It was evident that for parameter studies, the speed of these simulations needed a substantial boost.

Homogenized charging simulations are the solution.

The solution to the runtime issue resides in the homogenized charging simulations that we have implemented in the BatteryDict module of the GeoDict 2023 release. In this groundbreaking advancement, homogenized simulations cut down the simulation time to an absolute minimum of just a few minutes. For this achievement, the BESTmeso solver from Fraunhofer ITWM is integrated with the automatic calculation of effective parameters for the homogenized model derived from the fully resolved microstructure.

What has changed in the GeoDict 2023 release?

BESTmicro solver in GeoDict 2021: Up until the GeoDict 2021 release, the primary solver used to simulate the charging of lithium-ion battery cells was the BESTmicro solver of  Fraunhofer ITWM. The solver used the Butler-Volmer equation in 3D to simulate cell or electrode charging, in addition to mapping ion and electron transport processes within the electrode. Results of the solver were charge curves and 3D fields for lithium-ion concentration and current density.

LIR solver in GeoDict 2022: The LIR solver of Math2Market was adapted to facilitate flow and diffusion processes, and also 3D-resolved charging simulations of lithium-ion batteries. The LIR solver provided the same results as BESTmicro while requiring only about 10% of the simulation time and consuming thirty times less RAM. This allowed simulating the charging of a cell throughout its entire thickness at the micrometer scale for the first time.

BEST(:er)-micro solver in GeoDict 2023: This new implementation of the BESTmicro solver significantly reduces RAM consumption, but increases runtime, compared to the old BESTmeso solver. The users can adjust concentration-dependent parameters for the electrolyte using the BESTer-micro solver in the simulation.

BESTmeso solver in GeoDict 2023: The proven BESTmeso solver from Fraunhofer ITWM is integrated. The solver uses a pseudo-2D Newman model, which inherently is much faster than fully resolved microstructure simulations. While the results of homogenized simulations are limited to effective charge curves, they are excellent for systematic parameter studies.

Fully resolved vs. homogenized charging simulations: Fully resolved charging simulations remain the method of choice when detailed local information about processes within the microstructure during charging is needed. Transport limitations, crucial to investigate degradation mechanisms, are identified thanks to the 3D fields at the nano- to micro-meter scale.

The integration of a homogenized charging simulation expands the capabilities of BatteryDict and provide an attractive complement to the BESTmicro and LIR solvers, depending on the specific requirements and levels of investigation.


The input parameters needed for a homogenized charging simulation consist of the electrochemical parameters of the materials used - similar to those in fully resolved microstructure simulations - and additional effective parameters derived from the electrode's microstructure. These parameters include the average particle size of the active materials and the extent of the active surface area. Typically, the user needs to either know these sizes or, more challengingly, estimate them when conducting a 1D charging simulation with the Newman model.

However, in GeoDict, it is possible to manually calculate these sizes and even to automate this process in BatteryDict, when selecting homogenized charging simulations. The user simply inputs the same parameters as for the fully resolved simulation and uploads an electrode microstructure in GeoDict. Within minutes, BatteryDict provides a result file with charge curves, illustrating, for instance, the potential across the state of charge.

Remarkably, when the microstructure is replaced, the rest of the input parameters remain unchanged, allowing simulations for a different microstructure under the same conditions. This significantly facilitates the transition between different solvers and enables straightforward automation of simulations, especially for parameter studies.