Which AMG Linear Solver

Zuixy · March 11, 2011, 10:42

Hey,

I have a problem with solid body and a fluid flow. I use a segregated Solid Energy Model for the solid body. My Problem is that the temperatur changes just very very slowly. There a no big temperature changes due to the fluid. Any suggestions how i can improve the solver?

Thx

alastormoody11 · March 14, 2011, 03:41

hi,

the AMG linear solver just solves the simultaneous algebraic equations obtained after discretization, changing it not going to affect ne thing.

If your simulation is giving wrong results you need to check the simulation setup and not be concerned the linear solver.

If you provide some more data regarding your simulation setup maybe someone can help u.

Zuixy · March 14, 2011, 04:43

Thx for the answer. I didn't get wrong reluts and i could fix the problem by doing more iterations. Due to the fact that i'll have to calculate some variations i thought i could improve the speed a bit.

But another thing:
I want to calculate the heat flux over the outer cover of a cylinder.
1. I was using a Area Averaged Report with the Boundary Heat Flux field function over the interface. The result was zero.
2. I tried the same but with a cylinder section which had a smaler radius as the cylinder. The result was also zero.
3. I tried a heat transfer report over the interface. It was also zero.

After that i thought about the reason and got two ideas.
1. It is to small to be calculated.
2. There is something wrong with the interface. But what? I impried the bodies as i designed and the In-place Interface was created by StarCCM.
Do you have any idea where the problem could be?

alastormoody11 · March 14, 2011, 04:51

To increase the speed with which your solution converges you might try by increasing your under-relaxation factors

willimanili · March 20, 2011, 11:24

Try to optimize your initial settings close to the solution you expect, that should improve your speed, too.

March 11, 2011, 10:42	Which AMG Linear Solver	#1
Zuixy New Member Tom Schmitz Join Date: Mar 2011 Posts: 19 Rep Power: 15	Hey, I have a problem with solid body and a fluid flow. I use a segregated Solid Energy Model for the solid body. My Problem is that the temperatur changes just very very slowly. There a no big temperature changes due to the fluid. Any suggestions how i can improve the solver? Thx

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Fluent IcePak - AMG Solver	Russ	FLUENT	2	December 31, 2018 10:57
AMG solver _divergence detected	Shamoon	FLUENT	6	March 11, 2009 08:59
Linear system solver	elias	FLUENT	0	November 19, 2008 10:30
Error: divergence detected in AMG solver: temperat	Rubén	FLUENT	3	December 11, 2007 12:48
AMG versus ICCG	msrinath80	OpenFOAM Running, Solving & CFD	2	November 7, 2006 16:15

March 14, 2011, 03:41		#2
alastormoody11 Senior Member Join Date: Jun 2009 Posts: 100 Rep Power: 17	hi, the AMG linear solver just solves the simultaneous algebraic equations obtained after discretization, changing it not going to affect ne thing. If your simulation is giving wrong results you need to check the simulation setup and not be concerned the linear solver. If you provide some more data regarding your simulation setup maybe someone can help u.

March 14, 2011, 04:43		#3
Zuixy New Member Tom Schmitz Join Date: Mar 2011 Posts: 19 Rep Power: 15	Thx for the answer. I didn't get wrong reluts and i could fix the problem by doing more iterations. Due to the fact that i'll have to calculate some variations i thought i could improve the speed a bit. But another thing: I want to calculate the heat flux over the outer cover of a cylinder. 1. I was using a Area Averaged Report with the Boundary Heat Flux field function over the interface. The result was zero. 2. I tried the same but with a cylinder section which had a smaler radius as the cylinder. The result was also zero. 3. I tried a heat transfer report over the interface. It was also zero. After that i thought about the reason and got two ideas. 1. It is to small to be calculated. 2. There is something wrong with the interface. But what? I impried the bodies as i designed and the In-place Interface was created by StarCCM. Do you have any idea where the problem could be?

March 14, 2011, 04:51		#4
alastormoody11 Senior Member Join Date: Jun 2009 Posts: 100 Rep Power: 17	To increase the speed with which your solution converges you might try by increasing your under-relaxation factors

March 20, 2011, 11:24		#5
willimanili Member Join Date: Feb 2011 Location: DE-PB Posts: 56 Rep Power: 15	Try to optimize your initial settings close to the solution you expect, that should improve your speed, too.