Computational & Technology Resources
an online resource for computational,
engineering & technology publications |
|
Civil-Comp Proceedings
ISSN 1759-3433 CCP: 38
ADVANCES IN COMPUTATIONAL STRUCTURES TECHNOLOGY Edited by: B.H.V. Topping
Paper VII.11
Parallel and Distributed Adaptive Tetrahedral Mesh Generation J.K. Wilson and B.H.V. Topping
Department of Mechanical and Chemical Engineering, Heriot-Watt University, Edinburgh, United Kingdom J.K. Wilson, B.H.V. Topping, "Parallel and Distributed Adaptive Tetrahedral Mesh Generation", in B.H.V. Topping, (Editor), "Advances in Computational Structures Technology", Civil-Comp Press, Edinburgh, UK, pp 327-342, 1996. doi:10.4203/ccp.38.7.11
Abstract
A parallel adaptive tetrahedral mesh generation program using
the advancing front method has been described. The problem domain is initially defined by a course
background mesh of tetrahedral elements which form the input
for the finite element analysis and from which adaptive
parameters are calculated. Four-noded, constant strain
tetrahedra are used for the examples presented in this paper.
The background mesh is then divided into a number of
subdomains which are comprised of background tetrahedral
elements. Parallel adaptive mesh generation is carried out
by remeshing each subdomain concurrently using the nodal
adaptive parameters previously calculated. Within each subdomain
mesh generation is undertaken by refining each background
element, contained within the subdomain, in turn.
The remeshing of these tetrahedra is undertaken using the
advancing front method. Separate code was written for
sequential applications to run on a PC and parallel applications
to run on a transputer network. The code was merged
in such a way that it may be compiled and run in four different
environments: the Borland Turbo C compiler to perform
sequential runs on a PC; the GNU C compiler to perform
sequential runs within a UNIX environment; the 3L Parallel
C compiler to run parallel applications on a transputer
network; and the Parallel Virtual Machine (PVM) for distributed
runs within the UNIX environment. The program
has also been enhanced to increase the robustness of the algorithm,
to increase the quality of the resulting meshes and
ensure compatibility of tetrahedral element faces that form
the boundary of the subdomains.
purchase the full-text of this paper (price £20)
go to the previous paper |
|