Algebraic Multigrid for CFX Transient Analysis

vagantes · April 29, 2013, 12:14

Hi !

I am studying a transient thermal problem using CFX. The solution seems to converge. But I would like to try to apply multigrid technique in order to accelerate the solution process.
I am a bit confused with the solver control parameters.

Could someone explain me please (or give a hint) where can I find parameters that control the application of Multigrid method.
(I have looked into CFX Solver Modeling Guide, p. 450, 16.3.2 Linear Solver Parameters but I can not figure out how to use multigrid...)

Thank you in advance !

Lance · April 29, 2013, 12:29

CFX already use algebraic multigrid, and Im not sure that fiddeling with the parameters will speed up your simulation.
If you check the expert parameters you will find that
multigrid solver = true (by default)
and that
mg solver option = 5 (automatic)

ghorrocks · April 29, 2013, 20:04

This sort of comment seems to be a relic of SIMPLE based solvers where you continually had to tweak the under relaxation factor to get good convergence, and you also had to keep an eye on the linear solvers as well. With modern solvers like CFX it is very rare that you have to do this. These days 99% of the time the convergence issue is caused by a fundamental problem with the simulation setup and nothing to do with details of the linear solver.

oj.bulmer · April 30, 2013, 10:36

FLUENT has a coupled solver as well, though it uses the SIMPLE based solver as a default one. I wonder if anyone has done any comparison between these two coupled solvers!

OJ

April 29, 2013, 12:14	Algebraic Multigrid for CFX Transient Analysis	#1
vagantes New Member Join Date: Apr 2013 Posts: 5 Rep Power: 13	Hi ! I am studying a transient thermal problem using CFX. The solution seems to converge. But I would like to try to apply multigrid technique in order to accelerate the solution process. I am a bit confused with the solver control parameters. Could someone explain me please (or give a hint) where can I find parameters that control the application of Multigrid method. (I have looked into CFX Solver Modeling Guide, p. 450, 16.3.2 Linear Solver Parameters but I can not figure out how to use multigrid...) Thank you in advance !

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
FSI Transient Thermal Analysis	Abduri	CFX	1	January 1, 2017 23:54
algebraic multigrid vs geometric multigrid	Anna Tian	Main CFD Forum	5	June 18, 2013 06:03
vortex shedding, transient or steady state analysis type?	alfonsojurado	CFX	0	October 25, 2012 06:33
transient analysis _porous medium	shaswat	CFX	4	January 25, 2012 01:18
GAMG = geometric algebraic multigrid?	schmittp54	OpenFOAM Programming & Development	1	December 11, 2010 08:25

April 29, 2013, 12:29		#2
Lance Senior Member Lance Join Date: Mar 2009 Posts: 669 Rep Power: 22	CFX already use algebraic multigrid, and Im not sure that fiddeling with the parameters will speed up your simulation. If you check the expert parameters you will find that multigrid solver = true (by default) and that mg solver option = 5 (automatic)

April 29, 2013, 20:04		#3
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,870 Rep Power: 144	This sort of comment seems to be a relic of SIMPLE based solvers where you continually had to tweak the under relaxation factor to get good convergence, and you also had to keep an eye on the linear solvers as well. With modern solvers like CFX it is very rare that you have to do this. These days 99% of the time the convergence issue is caused by a fundamental problem with the simulation setup and nothing to do with details of the linear solver.

April 30, 2013, 10:36		#4
oj.bulmer Senior Member OJ Join Date: Apr 2012 Location: United Kindom Posts: 473 Rep Power: 20	FLUENT has a coupled solver as well, though it uses the SIMPLE based solver as a default one. I wonder if anyone has done any comparison between these two coupled solvers! OJ