Under-relaxation factors

steventay · February 19, 2010, 10:53

Hello,

May I know how to set under-relaxation factors for the velocity components, pressure correction and thermal energy for CFX? Many thanks!

zboud · February 19, 2010, 11:22

Same as the other post, but "Basic settings" tab ... there you have the Timescale factor.

steventay · February 19, 2010, 11:48

Hi Zboud,

Thanks again!

ghorrocks · February 20, 2010, 05:14

One further comment - under relaxation factors are commonly adjusted in SIMPLE based solvers like Fluent. It is very common to have to adjust them to obtain convergence. CFX is not a simple based solver and while it does have under relaxation factors it is very rare to have to adjust them. As zboud says, in CFX a far preferable way of adjusting solver stability in most cases is by using the time step size.

steventay · February 20, 2010, 07:14

Hi Glen,

Thanks for your comments. This info definitely helps. I have been trying to adjust the time-steps to achieve convergence, however, having reducing to 0.1s and 200 iterations, the run still does not convergence. Do I need to keep reducing the time-step till I get convergence? Is there any other setting that I need to do to help in the convergence? Thanks.

feizaghaee · February 20, 2010, 12:54

what are you modeling?

steventay · February 21, 2010, 02:39

Hi,

I am modeling a tank filled with fluid 1 and 1 coil of tubes coiled in the tank with fluid 2 flowing in the tube. The temp of fluids 1 and 2 will be in two extremes. I will measure the outlet temp of the tube and some points in the tank to know the temp change in fluid 1 as fluid 2 is passing thru the tube.

Meshing of this model is very difficult as even I have set the span angle and relevance center to fine, I still get the skewness ratio of 0.99958.

I set the convergence criteria to 1e-6 at the basic setting but with the timestep set at 0.1s and 200 iterations, it still does not converge. Should I continue to reduce the timestep or increase the number of iterations?

I am trying to set individual convergence criteria as I no need all the criteria to be set at 1e-6. For my run, my understanding is that themal energy is more vital and the rest of the setting can be set at 1e-3.

Any advise and suggestion are welcome and I need to know whether I am moving in the right direction and what else I can do to get convergence. Thank you and appreciate all of your time for reading my post!

feizaghaee · February 21, 2010, 03:28

continueing may not result in convergence. you should consider the rate of residual reduction. if it's very bouncy it may not converge. if it's reducing with acceptable rate continue it. but if the residuals are unstable try another time scale.

steventay · February 21, 2010, 06:34

Thanks for your prompt reply. I will consider your suggestions.

ghorrocks · February 21, 2010, 18:01

Sounds like a pretty common problem:

http://www.cfd-online.com/Wiki/Ansys...gence_criteria

February 19, 2010, 10:53	Under-relaxation factors	#1
steventay New Member Steven Tay Join Date: Jan 2010 Posts: 28 Rep Power: 16	Hello, May I know how to set under-relaxation factors for the velocity components, pressure correction and thermal energy for CFX? Many thanks!

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Relaxation Factors for Transient solvers	philippose	OpenFOAM Running, Solving & CFD	19	March 20, 2014 04:39
relaxation factors and time accuracy	Mike	Main CFD Forum	7	May 21, 2005 12:41
Relaxation Factors	Tim	Phoenics	3	June 30, 2004 02:03
relaxation factors	zhujianguo	Phoenics	1	August 5, 2003 06:17
relaxation factors adjust	zhujianguo	Phoenics	1	July 15, 2003 11:11

February 19, 2010, 11:22		#2
zboud Member prout Join Date: May 2009 Posts: 32 Rep Power: 16	Same as the other post, but "Basic settings" tab ... there you have the Timescale factor.

February 19, 2010, 11:48		#3
steventay New Member Steven Tay Join Date: Jan 2010 Posts: 28 Rep Power: 16	Hi Zboud, Thanks again!

February 20, 2010, 05:14		#4
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	One further comment - under relaxation factors are commonly adjusted in SIMPLE based solvers like Fluent. It is very common to have to adjust them to obtain convergence. CFX is not a simple based solver and while it does have under relaxation factors it is very rare to have to adjust them. As zboud says, in CFX a far preferable way of adjusting solver stability in most cases is by using the time step size.

February 20, 2010, 07:14		#5
steventay New Member Steven Tay Join Date: Jan 2010 Posts: 28 Rep Power: 16	Hi Glen, Thanks for your comments. This info definitely helps. I have been trying to adjust the time-steps to achieve convergence, however, having reducing to 0.1s and 200 iterations, the run still does not convergence. Do I need to keep reducing the time-step till I get convergence? Is there any other setting that I need to do to help in the convergence? Thanks.

February 20, 2010, 12:54		#6
feizaghaee Senior Member moein Join Date: Dec 2009 Posts: 132 Rep Power: 16	what are you modeling?

February 21, 2010, 02:39		#7
steventay New Member Steven Tay Join Date: Jan 2010 Posts: 28 Rep Power: 16	Hi, I am modeling a tank filled with fluid 1 and 1 coil of tubes coiled in the tank with fluid 2 flowing in the tube. The temp of fluids 1 and 2 will be in two extremes. I will measure the outlet temp of the tube and some points in the tank to know the temp change in fluid 1 as fluid 2 is passing thru the tube. Meshing of this model is very difficult as even I have set the span angle and relevance center to fine, I still get the skewness ratio of 0.99958. I set the convergence criteria to 1e-6 at the basic setting but with the timestep set at 0.1s and 200 iterations, it still does not converge. Should I continue to reduce the timestep or increase the number of iterations? I am trying to set individual convergence criteria as I no need all the criteria to be set at 1e-6. For my run, my understanding is that themal energy is more vital and the rest of the setting can be set at 1e-3. Any advise and suggestion are welcome and I need to know whether I am moving in the right direction and what else I can do to get convergence. Thank you and appreciate all of your time for reading my post!

February 21, 2010, 03:28		#8
feizaghaee Senior Member moein Join Date: Dec 2009 Posts: 132 Rep Power: 16	continueing may not result in convergence. you should consider the rate of residual reduction. if it's very bouncy it may not converge. if it's reducing with acceptable rate continue it. but if the residuals are unstable try another time scale.

February 21, 2010, 06:34		#9
steventay New Member Steven Tay Join Date: Jan 2010 Posts: 28 Rep Power: 16	Thanks for your prompt reply. I will consider your suggestions.

February 21, 2010, 18:01		#10
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	Sounds like a pretty common problem: http://www.cfd-online.com/Wiki/Ansys...gence_criteria