PROGRAM SUMMARY
Title of program:
Aranea
Catalogue identifier:
ADOG
Ref. in CPC:
139(2001)172
Distribution format: gzip file
Operating system: Windows 95/98/2000/NT, Tru64 V4.0f, SGI Irix
High speed store required:
6.5MK words
Number of lines in distributed program, including test data, etc:
168822
Keywords:
General purpose, Utility, Unstructured mesh generation,
Triangular mesh, User Interface Graphics.
Programming language used: Java
Computer:
PC ,
Alpha workstation ,
SGI Origin 2000 ,
Solaris 7 SPARC ,
Solaris 8x86 ,
PowerMac G4/350 MHz ,
iMac G3/400 MHz .
Nature of problem:
Several physical problems can be described in terms of coupled partial
equations in two dimensions. This is the case, for example, with plasma
particle and energy transport in axisymmetric tokamaks, or with the
evolution of vorticity in non neutral electron cylinders. At the basis
of solution of such equations is a discretisation of the governing
equations on a given mesh. Aranea is a code that automatically
generates unstructured triangular meshes in two dimensional plane
geometry, with arbitrary boundary shapes and connectivity. The
resulting meshes are well suited for solutions with the finite element
discretization.
Method of solution:
The construction of a mesh in Aranea is made according to a
parametrization of boundaries, a definition of boundary properties, and
a definition of the metric. The parametrization of the boundaries
defines the shape and connectivity of the various boundaries that
determine the simulation domain. Boundary properties refer to the
ability of various boundaries to specify boundary conditions or not.
They also refer to the fact that some boundaries are used to delimit the
simulation domain, while other boundaries may be used for the sole
purpose of aligning the mesh in certain regions of the simulation
domain. Finally, the metric specifies the desired size of the elements
that will make the mesh. The grid is then constructed by adding mesh
points at desired locations and defining the connectivity with
neighbouring points with the Bowyer-Watson algorithm.
Restrictions:
The meshes generated by Aranea must be in plane geometry.
Typical running time:
The running time may vary considerably, depending on the problem being
considered. The test cases presented were run on a 450 MHz pentium II
processor with 128 MB of RAM, under Windows 98. The execution time
varied from one to four seconds.
Unusual features:
None.