CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   OpenFOAM (http://www.cfd-online.com/Forums/openfoam/)
-   -   Mathematical Basics for GAMG (http://www.cfd-online.com/Forums/openfoam/60721-mathematical-basics-gamg.html)

joern April 14, 2008 10:53

Hi, I'm searching for the m
 
Hi,

I'm searching for the mathematical basics of the GAMG Solver.
In my literature i only find GMG and AMG. I want to know how the GAMG works and whats the difference to GMG and AMG.


Joern

mattijs April 14, 2008 14:11

GAMG is geometric/algebraic MG
 
GAMG is geometric/algebraic MG. Instead of the agglomeration being calculated from the matrix coefficients (and thus differ every 'solve') this one can use an agglomeration based on geometry (which is usually constant and can be cached).

See the 1.4 release notes. http://www.opencfd.co.uk/openfoam/doc/ReleaseNotes-1.4

braennstroem April 15, 2008 04:15

Hi Mattijs, how can I adjus
 
Hi Mattijs,

how can I adjust the agglomaration? Is setting cacheAgglomeration to true enough for the geometric and setting it to false ennough for the algebraic version?

Thanks!
Fabian

mattijs April 15, 2008 15:53

enter an invalid name for the
 
enter an invalid name for the agglomeration and you'll see the possible agglomeration methods. You can make your own one. Look e.g. at pairGAMGAgglomeration.H. Cache the agglomeration if you don't mind the extra storage and it does not change.

braennstroem April 17, 2008 11:01

Ah, I see, the choices are alg
 
Ah, I see, the choices are algebraicPair and
faceAreaPair; I assume that one would mostly use the faceAreaPair as long as there is no mesh motion and no automatic mesh adaption!?

Thanks!
Fabian

braennstroem December 11, 2008 09:45

Hi, I wonder, if there is a
 
Hi,

I wonder, if there is an option to write out the agglomerated coarse mesh for every level?

Thanks in advance!
Fabian

braennstroem December 18, 2008 03:17

Hi, me again... do you hav
 
Hi,

me again...
do you have a hint, if the use of an agglomerated coarse mesh is the good approach to get a coarse surface and volume mesh!? I would like to use a real coarse mesh for thermal calculations of a room divided in a couple of regions.

Fabian


All times are GMT -4. The time now is 22:02.