Loading...
Decoupling method for parallel Delaunay two-dimensional mesh generation
Linardakis, Leonidas
Linardakis, Leonidas
Abstract
Parallel mesh generation procedures that are based on geometric domain decompositions require the permanent separators to be of good quality (in terms of their angles and length), in order to maintain the mesh quality. The Medial Axis Domain Decomposition, an innovative geometric domain decomposition procedure that addresses this problem, is introduced. The Medial Axis domain decomposition is of high quality in terms of the formed angles, and provides separators of small size, and also good work-load balance. It presents for the first time a decomposition method suitable for parallel meshing procedures that are based on geometric domain decompositions.;The Decoupling Method for parallel Delaunay 2D mesh generation is a highly efficient and effective parallel procedure, able to generate billions of elements in a few hundred of seconds, on distributed memory machines. Our mathematical formulation introduces the notion of the decoupling path, which guarantees the decoupling property, and also the quality and conformity of the Delaunay submeshes. The subdomains are meshed independently, and as a result, the method eliminates the communication and the synchronization during the parallel meshing. A method for shielding small angles is introduced, so that the decoupled parallel Delaunay algorithm can be applied on domains with small angles. Moreover, I present the construction of a sizing function, that encompasses an existing sizing function and also geometric features and small angles. The decoupling procedure can be used for parallel graded Delaunay mesh generation, controlled by the sizing function.
Description
Date
2007-01-01
Journal Title
Journal ISSN
Volume Title
Publisher
Collections
Download Dataset
Files
Loading...
3282512.pdf
Adobe PDF, 3.31 MB
Rights Holder
Usage License
Embargo
Research Projects
Organizational Units
Journal Issue
Keywords
Citation
Advisor
Department
Computer Science
DOI
https://dx.doi.org/doi:10.21220/s2-mqsk-5d79