Nos tutelles

CNRS UPS Enseeiht


Accueil > Actualités

3D Hybrid Models for Electrical Machines

par Laurence Laffont - publié le , mis à jour le

Théo CARPI’s thesis defense intitled "Modèles 3D hybrides pour les machines électriques" (3D hybrid models for electrical machines) will be taking place on Thursday, 1st of April 2021, from 2 pm.

The work thesis has been supervised by Yvan LEFÈVRE and Jean-François LLIBRE.

Jury :

Daniel MATT
Christophe ESPANET
Jean François LLIBRE

Zoom link :

Meeting ID : 952 8648 2821
Passcode : 926464

Abstract :

In the field of electric actuators dedicated to embedded systems, increasing power density has become a major challenge. In terms of torque mass ratio, the electrical machines offering the best performance are permanent magnet machines, which are the focus of this manuscript. However, on all existing machines, the presence of heads winding on which no torque is produced can be noticed, in addition, the production of heat over them is associated with a loss in the overall power balance.

Among the possible areas for improvement of electrical machines, increasing the interaction area between the rotor magnetic field and the stator current density enables a significant gain in terms of torque mass ratio. To this end, we have thought of a new machine structure, which theoretically draws 100% of the copper present in the machine to generate torque. This machine has a toroidal structure and is the subject of this research. Thus, we will focus on the elaboration of the model of this new machine structure in order to study its performances and to compare them to the already existing machine structures. In order to achieve this goal, we will first investigate what is done in terms of analytical modeling on existing machines. Thus, we will seek to develop a modeling method common to all these machines including the proposed new structure in order to be able to compare them under the same mathematical assumptions.

For this purpose, a hybrid modeling method linking the analytical method to the finite difference method has been developed to model in three dimensions electric machines. The goal of this modeling is to remain relatively simple in terms of complexity and computation time in order to integrate them into sizing/optimization algorithms. This method will be developed and validated by finite elements and experimentally on the case of an axial flux machine before being used to model the new machine structure proposed in this thesis.