To read this content please select one of the options below:

Post processing for the vector finite element method: from edge to nodal values

Christian Vollaire (Centre de Génie Électrique de Lyon (CEGELY), Ecole Centrale de Lyon, Ecully, France)
François Musy (Laboratoire de Mathématiques Appliquées de Lyon (MAPLY), Ecole Centrale de Lyon, Ecully, France)
Ronan Perrussel (Laboratoire de Mathématiques Appliquées de Lyon (MAPLY), Ecole Centrale de Lyon, Ecully, France)
309

Abstract

Purpose

Propose post processing methods for the edge finite element (FE) method on a tetrahedral mesh. They make it possible to deduce vector values on the vertices from scalar values defined on the edges of the tetrahedra.

Design/methodology/approach

The new proposed techniques are based on a least squares formulation leading to a sparse matrix system to be solved. They are compared in terms of accuracy and CPU time on a FEs formulation for open boundary – frequency domain problems.

Findings

A significant improvement of vector values accuracy on the vertices of the tetrahedra is obtained compared to a classical approach with a very small additional computation time.

Originality/value

This work presents techniques: to obtain the values at the initial nodes of the mesh and not inside the tetrahedra; and to take into account the discontinuity to the interface between two media of different electromagnetic properties.

Keywords

Citation

Vollaire, C., Musy, F. and Perrussel, R. (2005), "Post processing for the vector finite element method: from edge to nodal values", COMPEL - The international journal for computation and mathematics in electrical and electronic engineering, Vol. 24 No. 4, pp. 1274-1283. https://doi.org/10.1108/03321640510615607

Publisher

:

Emerald Group Publishing Limited

Copyright © 2005, Emerald Group Publishing Limited

Related articles