Nan on LES dynamic models

sampaio · March 8, 2005, 13:52

Hi all,
I am trying to use a dynamic model for LES, but since the first iteration I got "nan" in all solvers...
Has anyone succeeded in using dyn models?
Do I need to do anything else, other than switch the name of the model in turbulenceProperties?
(I started with the "static" model, in my case 1Eq, and than after several timesteps, switched to dyn model)
Thanks,
Luiz

hjasak · March 8, 2005, 13:56

Do you have a zero turbulence initial condition? That would do it - have a look at the files in zero directory and make sure they make sense.

Hrv

henry · March 8, 2005, 13:59

Which dynamic model are you trying? I have just tried dynSmagorinsky on the tutorial case and it ran fine.

There may be an issue with the dynamic models starting from a zero or even uniform U field. Try running with a non-dynamic model first and then restart with a dynamic model.

eugene · March 8, 2005, 14:03

I dont think you can yet switch turbulence models in mid-calculation.

henry · March 8, 2005, 14:04

You can restart with a different turbulence model, I do it frequently.

sampaio · March 8, 2005, 14:07

Hi Henry,
It ran fine with dynSmagorinski, but it blows with dynOneEq...
I did run first with a static (non-dyn) model.
Thanks a lot

henry · March 8, 2005, 14:10

OK, I see what is going on. The problem is the dynamic models do ynot currently have stabilisation in the terms average(LL && MM)/average(MM && MM) where MM may be locally (or even globally) zero. I will look at the dynOneEq... and tell you what you need to change.

henry · March 8, 2005, 14:51

Try this

locDynOneEqEddy.C

sampaio · March 8, 2005, 16:41

Hi Henry,
I tried but it did not worked.
Did it work for you?
Thanks.
Luiz

sampaio · March 8, 2005, 16:44

Hi Henry,
I tried but it did not work.
Did it work for you?
Thanks.
Luiz

henry · March 8, 2005, 16:44

Yes, it worked fine for me. Put some print statements in it to check you are calling the new functions, the ones with the limiters in are at the top of the file.

March 8, 2005, 13:52	Hi all, I am trying to use a	#1
sampaio Member diablo80@web.de Join Date: Mar 2009 Posts: 93 Rep Power: 17	Hi all, I am trying to use a dynamic model for LES, but since the first iteration I got "nan" in all solvers... Has anyone succeeded in using dyn models? Do I need to do anything else, other than switch the name of the model in turbulenceProperties? (I started with the "static" model, in my case 1Eq, and than after several timesteps, switched to dyn model) Thanks, Luiz

March 8, 2005, 13:56	Do you have a zero turbulence	#2
hjasak Senior Member Hrvoje Jasak Join Date: Mar 2009 Location: London, England Posts: 1,907 Rep Power: 33	Do you have a zero turbulence initial condition? That would do it - have a look at the files in zero directory and make sure they make sense. Hrv __________________ Hrvoje Jasak Providing commercial FOAM/OpenFOAM and CFD Consulting: http://wikki.co.uk

March 8, 2005, 13:59	Which dynamic model are you tr	#3
henry Senior Member Join Date: Mar 2009 Posts: 854 Rep Power: 22	Which dynamic model are you trying? I have just tried dynSmagorinsky on the tutorial case and it ran fine. There may be an issue with the dynamic models starting from a zero or even uniform U field. Try running with a non-dynamic model first and then restart with a dynamic model.

March 8, 2005, 14:03	I dont think you can yet switc	#4
eugene Senior Member Eugene de Villiers Join Date: Mar 2009 Posts: 725 Rep Power: 21	I dont think you can yet switch turbulence models in mid-calculation.

March 8, 2005, 14:04	You can restart with a differe	#5
henry Senior Member Join Date: Mar 2009 Posts: 854 Rep Power: 22	You can restart with a different turbulence model, I do it frequently.

March 8, 2005, 14:07	Hi Henry, It ran fine with dy	#6
sampaio Member diablo80@web.de Join Date: Mar 2009 Posts: 93 Rep Power: 17	Hi Henry, It ran fine with dynSmagorinski, but it blows with dynOneEq... I did run first with a static (non-dyn) model. Thanks a lot

March 8, 2005, 14:10	OK, I see what is going on. T	#7
henry Senior Member Join Date: Mar 2009 Posts: 854 Rep Power: 22	OK, I see what is going on. The problem is the dynamic models do ynot currently have stabilisation in the terms average(LL && MM)/average(MM && MM) where MM may be locally (or even globally) zero. I will look at the dynOneEq... and tell you what you need to change. hua1015 likes this.

March 8, 2005, 14:51	Try this http://www.cfd-on	#8
henry Senior Member Join Date: Mar 2009 Posts: 854 Rep Power: 22	Try this locDynOneEqEddy.C

March 8, 2005, 16:41	Hi Henry, I tried but it did	#9
sampaio Member diablo80@web.de Join Date: Mar 2009 Posts: 93 Rep Power: 17	Hi Henry, I tried but it did not worked. Did it work for you? Thanks. Luiz

March 8, 2005, 16:44	Hi Henry, I tried but it did	#10
sampaio Member diablo80@web.de Join Date: Mar 2009 Posts: 93 Rep Power: 17	Hi Henry, I tried but it did not work. Did it work for you? Thanks. Luiz

March 8, 2005, 16:44	Yes, it worked fine for me. P	#11
henry Senior Member Join Date: Mar 2009 Posts: 854 Rep Power: 22	Yes, it worked fine for me. Put some print statements in it to check you are calling the new functions, the ones with the limiters in are at the top of the file.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Two-fluid models vs mixture models for bubbly flows	Hansong Tang	Main CFD Forum	6	December 8, 2009 04:21
Failed dimensional check with dynamic LES models	alberto	OpenFOAM Bugs	9	September 5, 2007 23:57
LES dynamic models fail dimensional check	alberto	OpenFOAM Running, Solving & CFD	0	June 11, 2007 14:23
Dynamic Mesh Models	Marc	Main CFD Forum	1	November 29, 2003 13:21
Dynamic Subgrid Models	Bono	Main CFD Forum	2	November 27, 2001 17:50