An adaptive solver for large-scale hierarchic finite element systems has been developed. A decision-making methodology aimed at selecting an optimal solution strategy on the basis of estimated conditioning, sparsity and memory requirements for a given problem has been devised. Numerical experiments
The efficiency of the p-version finite element method in a distributed computing environment
✍ Scribed by Ernst Rank; Martin Rücker; Alexander Düster; Henrike Bröker
- Publisher
- John Wiley and Sons
- Year
- 2001
- Tongue
- English
- Weight
- 783 KB
- Volume
- 52
- Category
- Article
- ISSN
- 0029-5981
- DOI
- 10.1002/nme.297
No coin nor oath required. For personal study only.
✦ Synopsis
Abstract
The p‐version of the finite element method is now commonly considered to be a very accurate discretization method for linear elliptic partial differential equations, but many researchers still doubt the efficiency of this method, when compared to the classical h‐version and applied to more complex problems. This paper will first discuss some general considerations about the efficiency of a numerical method and then present results on an evaluation of the p‐version. It will be shown, that there are many special techniques being applicable to the p‐version, yielding a well‐performing and robust method. This will be demonstrated on several examples, including non‐linear problems and a parallel implementation on a workstation cluster. A client–server software structure for an efficient integration of CAD and FEA using a strict separation of geometric and non‐geometric aspects will also be outlined. Copyright © 2001 John Wiley & Sons, Ltd.
📜 SIMILAR VOLUMES
A multilevel iteration scheme is utilized in the solution of equations resulting from the p-version of the finite element method. The advantage of this approach is that it is possible to attain the speed of multigrid techniques in the solution of systems of equations without generating a sequence of