Topology optimization of rotor magnetic circuits for axial flux electrodynamic thrust self-bearing machines

Details

Supervisors

Dehez, Bruno

Faculty

Ecole polytechnique de Louvain

Degree label

Master [120] : ingénieur civil électromécanicien, à finalité spécialisée: mécatronique

Abstract

enMagnetic bearings offer an alternative to conventional mechanical bearings by enabling contactless rotor support. Combined with electric motors, which likewise rely on electromagnetic principles, these two technologies naturally evolved toward their integration within a single structure called self-bearing machines. This thesis focuses specifically on axial flux electrodynamic thrust self-bearing machines (EDTSBMs), whose operating principle relies on electrodynamic effects to generate passive axial levitation forces. These machines combine two functions within a single winding system, namely, the driving function, responsible for torque generation, and the bearing function, responsible for the axial suspension of the rotor. This work aims to improve the performance of this machine through topology optimization of its rotor magnetic circuits. This method has already proven its effectiveness in the field of electric machines and makes it possible to explore innovative geometries that would be difficult to achieve through a conventional parametric approach. An electromagnetic model of the motor is first developed in order to define the performance criteria associated with the driving and bearing functions. The machine is then “unrolled” into a linear motor model in order to make 2D topology optimization computationally viable in COMSOL Multiphysics. A multi-objective approach is employed to explore the trade-offs between the two functions. Finally, the most promising topology is parameterized and further investigated within a complete 3D model, through a parametric study that accounts for magnetic saturation. The results reveal a set of geometries forming a Pareto front, characterized by a slot introduced into the rotor’s ferromagnetic yoke, forming a flux barrier. As an illustrative example, taking an arbitrary point in the middle of this front shows that, compared to a slotless ferromagnetic yoke, the bearing function can be improved by +28.93%, while the driving function is degraded by only 6.75%. This work therefore demonstrates that topology optimization is a relevant and effective tool for improving the performance of this type of machine, opening the way toward innovative rotor magnetic circuit architectures for EDTSBMs.