Coupled problems in electromagnetic systems

Research project P4/20 (Research action P4)

Persons :

Description :

Previous research has led to development of a high-performance software for computing electromagnetic fields. Based on finite element techniques as well as integral methods, this powerful numerical computation tool is original and innovative in several respects: an object-oriented structure, use of many formulations and specially adapted finite elements, automatic adaptive meshing based on error estimators, the ability to take nonlinearities and anisotropy into account,... With this modelling tool it is possible to solve most problems encountered in electromagnetism by computing not only electric and magnetic fields but also derived physical quantities such as eddy currents, Joule losses, electromagnetic forces and torque, ...

The aim of current research is to progress further in modelling. In real situations, electromagnetic phenomena are generally coupled with other physical phenomena, e.g., of thermal or mechanical nature. To study such coupled problems it is necessary, on the one hand, to solve simultaneously all the equations governing the problem and, on the other hand, to know how the properties of materials depend on the various physical quantities. For instance, electric conductivity and magnetic permeability are temperature dependent.

Research is thus conducted along two lines:

- development of the computing tool so that it can solve all the equations of coupled problems simultaneously. Thanks to its object-oriented structure, the software is completely modular and constitutes an ideal starting point for such developments;

- characterising the properties of materials (nonlinearity, anisotropy, magnetic and electric hysteresis, ...) and their dependence on external parameters. In some cases, these properties are well known but in others, the dependences must be determined experimentally prior to introduction into numerical models.

Within the partnership, tasks are distributed among the three laboratories (ULg, KUL, RUG), expertise pooled, and technical means combined so as to make the most of the complementarity between numerical and experimental research.

These studies will make it possible to study a wide range of important phenomena that are not yet fully mastered:

- coupling with power circuits will make it possible to study how the harmonics of power supplies affect the behaviour, e.g., of electric machines, and will lead to improved design procedures for passive components in power electronics;

- electromechanical coupling will notably make it possible to study magnetostrictive phenomena. These are often considered parasitic effects (generating noise and losses in magnetic cores) but might also be exploited as the basis of practical applications. Piezoelectric motors are a promising example of electromechanical coupling;

- examples of electrothermal coupling are many. They are encountered in all electric machines. In many such machines, the balance between losses (hysteresis, eddy currents, Joule effect) and cooling constitutes one of the major dimensioning requirements. Another important field of application is induction heating. Here, in-depth analysis of inductor design is the only way to ensure the as homogeneous as possible temperature distribution required by high technology metallurgic processes.