CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)
-   -   STRUCTURE OF CFD CODE (http://www.cfd-online.com/Forums/main/15479-structure-cfd-code.html)

MAB July 27, 2008 15:20

STRUCTURE OF CFD CODE
 
Hi i am trying to visualize structure of CFD code. the point where i get stuck is 'how these commercial codes solve the matrix of discritised equations'. the only methods i know are GUASS SEIDEL and SOR etc... but i trying to understand what methods are being used by e.g CFX, FLUENT, PHOENICS etc thanks in advance

Nabi July 28, 2008 01:06

Re: STRUCTURE OF CFD CODE
 
They use usually Multigrid. Fluent uses AMG (Algebraic Multigrid). Numeca uses Multigrid. Flow3D 8.0 uses SOR, but they improved their solver to GEMRES is version 9.0. But in anycase they use usually AMG for the new commercial solvers.


MAB July 28, 2008 04:38

Re: STRUCTURE OF CFD CODE
 
Thanks Nabi... what about CFX ?

Charles July 28, 2008 11:52

Re: STRUCTURE OF CFD CODE
 
Some codes also use conjugate gradient type of solvers.

jojojo July 29, 2008 10:43

Re: STRUCTURE OF CFD CODE
 
But note that CG is not a general purpose solver! Those codes might use it for diffusion only cases.

Anishtain4 August 10, 2008 03:47

Re: STRUCTURE OF CFD CODE
 
not only diffusion BICG or CGSTAB can be implied for a non symetric and even non positive definition matrices but they both break down in some cases as it happened for me. and not to mention lots of codes use a ADI method specially for a 3d code. I prefer it while you can use it because of simplisity and efficiency.


All times are GMT -4. The time now is 05:41.