|
[Sponsors] |
March 5, 2003, 04:07 |
numbering in unstructured grid
|
#1 |
Guest
Posts: n/a
|
Hi, when we use a usual structured grid (for example, rectangles in a 2D case), each node (or cell) receives its "natural" number. The unstructured grid looks as a chaotic set of nodes. How it is necessary to number its nodes and cells? Thank.
reader. |
|
March 5, 2003, 05:11 |
Re: numbering in unstructured grid
|
#2 |
Guest
Posts: n/a
|
Unstructured meshes will, usually, not show any "natural" numbering pattern.
If you are writing a CFD solver for unstructured meshes, you will have to keep track of the neighbors for each cell, as well as its faces. Madhukar Rao ACRI |
|
March 5, 2003, 08:26 |
Re: numbering in unstructured grid
|
#3 |
Guest
Posts: n/a
|
When you generate an unstructured grid, the grid generator will number the nodes somehow. You may want to renumber them for efficient cache utilization. See
R. Lohner: Applied CFD Techniques, Chapter 15 |
|
March 5, 2003, 10:05 |
Re: numbering in unstructured grid
|
#4 |
Guest
Posts: n/a
|
Here are some references:
R. Löhner - Some Useful Data Structures for the Generation of Unstructured Grids; Comm.Appl.Num.Meth. 4, 123-135 (1988). R. Löhner and M. Galle - Minimization of Indirect Addressing for Edge-Based Field Solvers; AIAA-02-0967 (2002). R. Löhner - Applied CFD Techniques; J. Wiley & Sons (2001). Regards, |
|
March 5, 2003, 10:36 |
Re: numbering in unstructured grid
|
#5 |
Guest
Posts: n/a
|
Lohner, Rainald, "Renumbering strategies for unstructured-grid solvers operating on shared-memory, cache-based parallel machines", Computer methods in applied mechanics and engineering, Elsevier, Vol 163, pp.95-109, 1998
|
|
March 5, 2003, 11:36 |
Re: numbering in unstructured grid
|
#6 |
Guest
Posts: n/a
|
I used GAMBIT to generate 10-node tetrahedral mesh. The node numbered well. Say, node 1,3,6,10 are always on the vertex, between 1 and 3 is always 2, between 1 and 6 is always 4, the others are the same.
|
|
March 6, 2003, 03:23 |
Re: numbering in unstructured grid
|
#7 |
Guest
Posts: n/a
|
When you generate an unstructured grid, an additional piece of information is required to supplement the node coordinates and numbers, namely a connectivity table. This table lists the node numbers of each cell (or element). For example, in 2D, if the quadrilateral cell No. k comprises of the corner nodes numbered { n1, n2, n3, n4 }, the connectivity table has { n1, n2, n3, n4 } as its k-th row (or coloumn).
For structured grid this information is redundant, since each cell indices are related to its node indices. For example, in 2D, cell (i+1/2,j+1/2) comprises of the nodes { (i,j), (i+1,j), (i,j+1), (i+1,j+1) }. |
|
March 6, 2003, 14:14 |
Re: numbering in unstructured grid
|
#8 |
Guest
Posts: n/a
|
Hi,
The numbering of cells/nodes in unstructured grid goes beyond efficient memory utilization & reducing band width of the coefficient matrix. An fully explicit will work well with un-ordered mesh. If you want to implement implicit/Semi-Implicit scheme your mesh (cells) must be re-ordered. If you don't do this the scheme reverts to explicit.... since updated value may not be available. The scheme is then at mercy of CFL condition. I was not aware of this until, I had to write a unstructured grid code for simulatin incompressible viscous flows. Intrestingly this is not stated in the literature that I came across. Probably it is implied or too "trivial" to be stated. Abhijit |
|
March 6, 2003, 14:58 |
Re: numbering in unstructured grid
|
#9 |
Guest
Posts: n/a
|
Some more refernces which might be of interest:
D. Sharov, H. Luo, J.D. Baum and R. Löhner - Implementation of Untructured Grid GMRES+LU-SGS Method on Shared-Memory, Cache-Based Parallel Computers; AIAA-00-0927 (2000). Sharov, D. and Nakahashi, K., "Reordering of Hybrid Unstructured Grids for Lower-Upper Symmetric Gauss-Seidel Computations," AIAA Journal, Vol. 36, No. 3, pp. 484-486, 1998. Regards |
|
December 31, 2010, 01:07 |
numbering in unstructured grid
|
#10 | |
New Member
Join Date: Dec 2010
Posts: 1
Rep Power: 0 |
Quote:
|
||
Thread Tools | Search this Thread |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Help! Validity of unstructured grid on S809 airfoil | didiean | FLUENT | 11 | April 3, 2011 03:10 |
2D unstructured Grid | prapanj | Main CFD Forum | 4 | March 15, 2008 18:01 |
unstructured mesh grid independence for Fluent | Shane Schouten | FLUENT | 0 | October 11, 2006 17:50 |
markers in unstructured grid | lavan | Main CFD Forum | 1 | July 12, 2004 12:34 |
Combustion Convergence problems | Art Stretton | Phoenics | 5 | April 2, 2002 06:59 |