Design and realization of a magnetic suspension for a high performance flywheel energy storage system

Details

Supervisors

Dehez, Bruno

Faculty

Ecole polytechnique de Louvain

Degree label

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

Abstract

This master’s thesis aims at developing an analytical model of a particular topology of active magnetic bearing (AMB). This topology, called slotless homopolar hybrid AMB (SHH-AMB), presents the advantage of mitigating the iron losses. This makes it an interesting candidate for a flywheel energy storage system (FESS) application, that stores energy under a kinetic form. After a description of the layout and a proof of concept, the global parameters of the SHH-AMB topology are identified as the axial position stiffness kz, the radial position stiffness kϵ, the tilt angle stiffness kΨ and the current stiffness ki. These stiffnesses model the electrodynamic and detent forces acting upon the rotor of the bearing. From these considerations, a semi-analytical magnetic model is developed based on a Fourier approach also known as the subdomain method, which solves the Maxwell’s equations in every subdomain. This model allows to precisely compute the distribution of magnetic flux density when the rotor of the AMB is radially centered. The model is also extended to take a radial eccentricity of the rotor into account thanks to a modulation function. The values of the global parameters are retrieved thanks to these distributions. A comparison with a finite element model (FEM) demonstrates the validity of the model. A mechanical model is also developed to assess the performances of a given bearing by computing the self-discharging time as well as the maximum speed of the rotor. Based on the magnetic and mechanical model, an optimization routine is set up with a view to the building of a prototype. The objectives being to have a lightweight device with the highest self-discharging time, this optimization uses the geometrical and electromagnetic parameters of the bearing as variables. Constraints are added to ensure the feasibility of the device. These are of different natures: geometry, magnetic saturation in the ferromagnetic materials, axial stability and lift-off criteria, and maximum rotating speed. The output of the optimization consists in a Pareto fronts from which a device is selected based on an analysis of the parameters of the machines. A design of a prototype built on the basis of the optimization is proposed. The objective of the prototype is to validate the magnetic model by comparing the actual global parameters of the prototype with those predicted by the model. However, difficulties to lift off the rotor have been encountered, making the experimental evaluation of some of the characteristics of the bearing impossible. As it does not require the levitation of the rotor, the value of the axial stiffness is experimentally evaluated and corresponds to the one predicted by the model.