Mimicking soft magnetic composite geometries algorithmically

In this paper we describe an algorithmic method to imitate geometries of soft magnetic composites. We use microscope images to determine some of the most important geometric characteristics of the material and optimize the algorithm such that the measured and simulated characteristics are in agreement. We compute the static magnetization curves using the geometries obtained from the images and the algorithm comparing the results. The aim is to produce realistic geometries to be used in homogenization schemes as well as to study the effects of magnetic contacts in general.