An analogy between open boundary condition and infinite element domains

Affifa Adeeb1, Timothy J Flack1
1University of Cambridge, Cambridge, United Kingdom
Publié en 2024

A magnetic field associated with current carrying coil(s) extends towards infinity, but an analytical model is of finite size. In the finite element analysis (FEA) of electromagnetic systems, open space needs to be modelled as close to reality as possible to get reliable results without losing accuracy and saving computational cost. Using magnetic vector potential (A) formulation such problems can be modelled by open boundary condition (OBC), also called asymptotic boundary condition (ABC) in the literature. In 2D axisymmetric problems, at the boundary where OBC is imposed, A is a prescribed function of the azimuthal angle and expansion of this function shows the higher the number of harmonics, the faster their magnitude decays with increasing radial distance of the boundary. In COMSOL, there is no explicit asymptotic boundary condition, so the current density boundary condition is modified to model the open domain asymptotically. Amperes law along with the solution of A makes the surface current density boundary condition behave like ABC. A 2D axisymmetric problem is setup in COMSOL where two single-turn circular coplanar coils are modelled in a steady state and frequency domain. The coil’s self and mutual inductance are obtained using OBC and comparison is made with infinite element domains of COMSOL and Dirichlet boundary condition in terms of error between inductances values. In steady state analysis, results are compared against values obtained from an analytical solution of self and mutual inductance for adapted problem setting. It has been found the implemented OBC and infinite element domains of COMSOL are analogous and can be used interchangeably and placed closer to the FEA problem design space without losing accuracy of model and subsequent results. The adapted modelling technique can be used in analytical design of electromagnetic systems giving reliable results within confined design space and saving computational time and resources.

Téléchargement