due to wall function, continue refining mesh until when

zandi · January 16, 2010, 15:50

Hi
I'm doing a free surface flow CFD modeling project that the velocity and pressure near the bottom (wall boundary) are important parameters. for evaluating the pressure result I have data and after refining mesh I saw that the pressur was not sensitive to it and I reached to a good answer.

but for velocity I do not have any data and just want to reach an answer that not depedent on mesh size.

I use k-epsilon standard in CFX code

so my limitation for y+ is 20 and 100 (CFX help)

now i'm reaching to y+ under 20 and still my velocity profile near the boundary changes due to make the meshes smaller. what can I do

please help

ghorrocks · January 16, 2010, 17:21

I would switch to the SST turbulence model then you can use the scalable automatic wall functions. Then it would matter when y+ goes under 11 and you can refine as much as you like.

Is there a reason you wish to use k-e?

zandi · January 17, 2010, 07:59

Dear G. Horrocks Salam
and thanks a lot
actually my masters insist on it maybe they're not so friendly with low Rey models. I'm not too!
they said the k-e is sufficient for my case.

by the way

if I change my turbulent model how much change result in my work. does my pressure or mass flow rate change. do I have compute any parameters before

best regards Zandi

ghorrocks · January 17, 2010, 17:20

SST is not a low Reynolds number turbulence model! It is a general purpose turbulence model and works well for a wider range of flows than k-e. If you are having problems with mesh refinement in k-e then I recommend you go to SST as the wall function issue is not a problem for SST.

As for whether the change in turbulence model will affect things - how can I possibly know that? You have not described at all what you are doing. It is straightforward to work that out for yourself - do a test run using SST and find out for yourself whether it makes a difference. For most flows it takes the same run time as k-e and converges the same so it should be a easy test to do.

zandi · January 18, 2010, 03:53

thank you very much from your advice

paulo · January 18, 2010, 10:16

Quote:

Originally Posted by ghorrocks

(...) For most flows it takes the same run time as k-e and converges the same so it should be a easy test to do.

it can be even faster, if you use the k-epsilon results to start your simulation.

Best Regards,

Paulo Rocha

zandi · January 18, 2010, 14:29

Quote:

Originally Posted by paulo

it can be even faster, if you use the k-epsilon results to start your simulation.

Best Regards,

Paulo Rocha

thank you friend but could you tell me how I could use the result
maybe when we want write the solver but I don't know which option should I use
thanks again

ghorrocks · January 18, 2010, 17:47

Why use k-e at all? Just use SST then you don't have any issues about wall functions with mesh refinement. The SST model has superseded the k-e model for most general applications.

zandi · January 23, 2010, 05:16

thank you, nice of you

January 16, 2010, 15:50	due to wall function, continue refining mesh until when	#1
zandi Senior Member Fatema Zandi Goharrizi Join Date: Mar 2009 Posts: 158 Rep Power: 17	Hi I'm doing a free surface flow CFD modeling project that the velocity and pressure near the bottom (wall boundary) are important parameters. for evaluating the pressure result I have data and after refining mesh I saw that the pressur was not sensitive to it and I reached to a good answer. but for velocity I do not have any data and just want to reach an answer that not depedent on mesh size. I use k-epsilon standard in CFX code so my limitation for y+ is 20 and 100 (CFX help) now i'm reaching to y+ under 20 and still my velocity profile near the boundary changes due to make the meshes smaller. what can I do please help

January 17, 2010, 07:59		#3
zandi Senior Member Fatema Zandi Goharrizi Join Date: Mar 2009 Posts: 158 Rep Power: 17	Dear G. Horrocks Salam and thanks a lot actually my masters insist on it maybe they're not so friendly with low Rey models. I'm not too! they said the k-e is sufficient for my case. by the way if I change my turbulent model how much change result in my work. does my pressure or mass flow rate change. do I have compute any parameters before best regards Zandi

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Moving mesh	Niklas Wikstrom (Wikstrom)	OpenFOAM Running, Solving & CFD	122	June 15, 2014 07:20
[blockMesh] BlockMesh FOAM warning	gaottino	OpenFOAM Meshing & Mesh Conversion	7	July 19, 2010 15:11
OpenFoam 14 installation problem	gfcoppola	OpenFOAM Installation	20	November 2, 2007 14:38
Wall function problem in Fluent	mefpz	FLUENT	1	October 10, 2007 14:43
3-D turbulence model with wall function	Henry	Main CFD Forum	7	May 19, 2006 11:40

January 16, 2010, 17:21		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,916 Rep Power: 145	I would switch to the SST turbulence model then you can use the scalable automatic wall functions. Then it would matter when y+ goes under 11 and you can refine as much as you like. Is there a reason you wish to use k-e?

January 17, 2010, 17:20		#4
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,916 Rep Power: 145	SST is not a low Reynolds number turbulence model! It is a general purpose turbulence model and works well for a wider range of flows than k-e. If you are having problems with mesh refinement in k-e then I recommend you go to SST as the wall function issue is not a problem for SST. As for whether the change in turbulence model will affect things - how can I possibly know that? You have not described at all what you are doing. It is straightforward to work that out for yourself - do a test run using SST and find out for yourself whether it makes a difference. For most flows it takes the same run time as k-e and converges the same so it should be a easy test to do.

January 18, 2010, 03:53		#5
zandi Senior Member Fatema Zandi Goharrizi Join Date: Mar 2009 Posts: 158 Rep Power: 17	thank you very much from your advice

January 18, 2010, 17:47		#8
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,916 Rep Power: 145	Why use k-e at all? Just use SST then you don't have any issues about wall functions with mesh refinement. The SST model has superseded the k-e model for most general applications.

January 23, 2010, 05:16		#9
zandi Senior Member Fatema Zandi Goharrizi Join Date: Mar 2009 Posts: 158 Rep Power: 17	thank you, nice of you