Question on courant number

danvica · March 18, 2012, 12:14

I'm running a pisofoam with k-OmegaSST turbolence model case with an average courant number of about 0.04 but a max value of 3.2.

The residual for all the variables are ok, so far the solver had calculated thousand steps without any problem, but...

How can I verify whether the results are trustable ?

Is courant number just an indication of the convergence or something more ?

Thanks for any comment.
Daniele

kumar · March 19, 2012, 03:48

HI,
From my experience to check if your solution is right or not you should also check the residuals, especially the Final residuals.

I think courant number is mostly used to check the stability and to keep your solver with in the limits of the deltaT, to achieve a reliable and stable solution.

regards
K.Suresh kumar

akidess · March 19, 2012, 03:55

The Courant number doesn't even indicate convergence. All it says is that for the PISO algorithm to work, you'll probably get into troubles if your maximum Courant number is larger than 1.

Does it make sense for your case that the maximum and the average velocity are so two orders of magnitude apart? Do you have any (experimental) data you can use as a guide?

- Anton

danvica · March 19, 2012, 07:19

Thanks for the feedback.

I'm simulation the flow of water into a valve. As you (Anton) said, two order of magnitude is a lot.
So far, from Paraview, all I can say is that the flow velocity in the valve is nowhere more that three times the one at the inlet.

Basing on the definition of Courant number this means there's some "problem" with the mesh. But is it a problem (if the solver converges) ?

Daniele

kmooney · March 19, 2012, 11:11

If your Co is getting that high I'm guessing that you don't have adjustable time stepping enabled and a maxCo defined. Look into the controlDict options available and try it out.

danvica · March 19, 2012, 11:32

No, your're right.
But I don't want to limit the timestep just because some cells are too small. I'm not worried about them... or should I ? That's the question.

In the meanwhile the solver is still parallel running, residuals are fine (convergence is reached within tollerance in 1-2 iteration), and mean/max Courant num are oscillating at about 0.04/3.0...

Is there any way to display Courant number in Parafoam ? As it was a physical field.

Daniele

kmooney · March 19, 2012, 11:38

You could instantiate a new volScalarField, populate it with the local Courant No and have it print out with the rest of your results. Take a look at CourantNo.H for some hints on how to calculate the Co field.

March 18, 2012, 12:14	Question on courant number	#1
danvica Senior Member Daniele Vicario Join Date: Mar 2009 Location: Novara, Italy Posts: 142 Rep Power: 17	I'm running a pisofoam with k-OmegaSST turbolence model case with an average courant number of about 0.04 but a max value of 3.2. The residual for all the variables are ok, so far the solver had calculated thousand steps without any problem, but... How can I verify whether the results are trustable ? Is courant number just an indication of the convergence or something more ? Thanks for any comment. Daniele

March 19, 2012, 03:55		#3
akidess Senior Member Anton Kidess Join Date: May 2009 Location: Germany Posts: 1,377 Rep Power: 30	The Courant number doesn't even indicate convergence. All it says is that for the PISO algorithm to work, you'll probably get into troubles if your maximum Courant number is larger than 1. Does it make sense for your case that the maximum and the average velocity are so two orders of magnitude apart? Do you have any (experimental) data you can use as a guide? - Anton __________________ On twitter @akidTwit Spend as much time formulating your questions as you expect people to spend on their answer.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Courant Number Problems	wschosta	OpenFOAM Running, Solving & CFD	5	February 28, 2020 04:45
[snappyHexMesh] external flow with snappyHexMesh	chelvistero	OpenFOAM Meshing & Mesh Conversion	11	January 15, 2010 20:43
Problems with Courant number (LaunderGibsonTurbulence Model)	sven	OpenFOAM	3	August 10, 2009 04:12
Courant number, patches, etc	oort	OpenFOAM	1	July 24, 2009 19:05
Fluent 6.4 courant number	Aris Nikolopoulos	FLUENT	0	May 6, 2008 09:52

March 19, 2012, 03:48		#2
kumar Senior Member Suresh kumar Kannan Join Date: Mar 2009 Location: Luxembourg, Luxembourg, Luxembourg Posts: 129 Rep Power: 17	HI, From my experience to check if your solution is right or not you should also check the residuals, especially the Final residuals. I think courant number is mostly used to check the stability and to keep your solver with in the limits of the deltaT, to achieve a reliable and stable solution. regards K.Suresh kumar

March 19, 2012, 07:19		#4
danvica Senior Member Daniele Vicario Join Date: Mar 2009 Location: Novara, Italy Posts: 142 Rep Power: 17	Thanks for the feedback. I'm simulation the flow of water into a valve. As you (Anton) said, two order of magnitude is a lot. So far, from Paraview, all I can say is that the flow velocity in the valve is nowhere more that three times the one at the inlet. Basing on the definition of Courant number this means there's some "problem" with the mesh. But is it a problem (if the solver converges) ? Daniele

March 19, 2012, 11:11		#5
kmooney Senior Member Kyle Mooney Join Date: Jul 2009 Location: San Francisco, CA USA Posts: 323 Rep Power: 18	If your Co is getting that high I'm guessing that you don't have adjustable time stepping enabled and a maxCo defined. Look into the controlDict options available and try it out.

March 19, 2012, 11:32		#6
danvica Senior Member Daniele Vicario Join Date: Mar 2009 Location: Novara, Italy Posts: 142 Rep Power: 17	No, your're right. But I don't want to limit the timestep just because some cells are too small. I'm not worried about them... or should I ? That's the question. In the meanwhile the solver is still parallel running, residuals are fine (convergence is reached within tollerance in 1-2 iteration), and mean/max Courant num are oscillating at about 0.04/3.0... Is there any way to display Courant number in Parafoam ? As it was a physical field. Daniele

March 19, 2012, 11:38		#7
kmooney Senior Member Kyle Mooney Join Date: Jul 2009 Location: San Francisco, CA USA Posts: 323 Rep Power: 18	You could instantiate a new volScalarField, populate it with the local Courant No and have it print out with the rest of your results. Take a look at CourantNo.H for some hints on how to calculate the Co field.