A parallel domain decomposition algorithm for the adaptive finite element solution of 3-D convection-diffusion problems

Jimack, P.K. and Nadeem, S.A. (2002) A parallel domain decomposition algorithm for the adaptive finite element solution of 3-D convection-diffusion problems. In: Computational science - ICCS 2002 : international conference, Amsterdam, The Netherlands, April 21-24, 2002 : proceedings. Lecture Notes in Computer Science, 2 (2330). Springer , Berlin , pp. 797-805. ISBN 354043593X

Full text available as:

Preview

Text jimackp41_JN02.pdf Available under License : See the attached licence file. Download (201Kb)

Abstract

In this paper we extend our previous work on the use of domain decomposition (DD) preconditioning for the parallel finite element (FE) solution of three-dimensional elliptic problems [3,6] and convection-dominated problems [7,8] to include the use of local mesh refinement. The preconditioner that we use is based upon a hierarchical finite element mesh that is partitioned at the coarsest level. The individual subdomain problems are then solved on meshes that have a single layer of overlap at each level of refinement in the mesh. Results are presented to demonstrate that this preconditioner leads to iteration counts that appear to be independent of the level of refinement in the final mesh,even in the case where this refinement is local in nature: as produced by an adaptive finite element solver for example.

Item Type:	Book Section
Copyright, Publisher and Additional Information:	Copyright © 2002 Springer-Verlag Berlin Heidelberg. This is an author produced version of a chapter published in Computational science - ICCS 2002 : international conference, Amsterdam, The Netherlands, April 21-24, 2002 : proceedings'.
Academic Units:	The University of Leeds > Faculty of Engineering (Leeds) > School of Computing (Leeds)
Depositing User:	Repository Officer
Date Deposited:	20 Dec 2006
Last Modified:	08 Feb 2013 17:03
Published Version:	http://www.springerlink.com/content/xvqrvvdk3jgq/
Status:	Published
Publisher:	Springer
Refereed:	No
Related URLs:	http://www.comp.leeds.ac.uk/pkj/publicat...
URI:	http://eprints.whiterose.ac.uk/id/eprint/1867

Actions (login required)

View Item