Geometry Imitations and Material Parameter Identifications for Soft Magnetic Composites

In this thesis, electromagnetic properties of soft magnetic composite (SMC) materials are studied computationally. The thesis introduces a methodology for automatic generation of SMC geometry imitations for numerical simulations based on the finite element method. The approach allows studying how certain features in the geometries, for example the thicknesses of the insulations between the conductive particles, affect the electromagnetic properties of the material.

A leading principle in this thesis is consistency with measurements. Effective properties of a sample, specifically a nonlinear B-H curve, a frequency sweep of losses, and a frequency sweep of effective low-field permeability, are measured. Then, localized material laws are estimated such that computations of the effective properties agree with the measurements.

The thesis argues that contacts between SMC material particles are not necessary in order to understand linear and nonlinear magnetostatic behavior of the material. Estimations of localized complex permeability and resistivity of material particles are carried out. It is argued that the localized quantities are not unique, but they are heavily dependent on the geometric properties of the material imitations. These dependencies are quantified. This thesis also argues that 2-D models of SMC materials are questionable in terms of simulated effective permeabilities and losses.