Wrong calculation of nut in the kOmegaSST turbulence model

timo_IHS · March 15, 2011, 09:38

Hello Felix,

the limiter is like your first mentioned one.

The correction to $\nu_t$ you recommended in post #1 leads from the red curve to the green curve. And this is what I mean by "instability".
Because of the results are normalized by the results of the CC, I said that it is "closer". But as you see, the old result is more than 10% away!

So because of the instability I was searching for a reason and I found the two definitions in the $\omega$ -equation:
$\gamma S^2$ (1)
and
$\gamma { {\omega}\over{k}}\nu_t S^2$ (2)

The result of this (2) is the blue curve, which is stable (guide vanes).
But testing this modified version (2) on a separating diffuser, I don't get the expected separation. I haven't shown here.

The grids aren't special, y+ of about 30.
The schemes shouldn't have an influence, because in the old version it worked well. Fluid is "just" water. And it's steady state.

Greetz

FelixL · March 15, 2011, 09:55

I can't see the graphs you're talking about. Could you try to reupload them?

Greetings,
Felix.

timo_IHS · March 15, 2011, 10:22

I talked about the graph in post #19, sorry I didn't mention again.
Or you also can't see this?

FelixL · March 15, 2011, 10:25

Sorry, no can do.

timo_IHS · March 15, 2011, 10:25

.. or here

FelixL · March 15, 2011, 12:52

Thanks for the plot.

Hm, well, it looks like when using (1) the simulation fails to reach a steady state. Is the guide vane in your test case possibly on the verge of separation? If so, your modification (2) probably removed the separation thus helping the simulation to converge. This would also explain why the separation point of your diffusor geometry isn't correctly predicted.

If it's pretty sure that the guide vane flow remains attached for a whole range of different turbulent characteristics inside the flow field, the only thing I could imagine causing your trouble would be the mesh. Could be the wall functions you are using, too, but since I usually avoid to use these, I can't really give a qualified comment on that topic.

Greetings,
Felix.

lakeat · January 24, 2012, 17:16

Hi Felix,

Hmm, thanks for sharing.

I am curious, have you used k-omega-SST for some bluff body flows. (Incompressible, 2D URANS)?

I can't get a correct vortex shedding freq with it. Any experience?

Generally, I found many models including the k-omega-SST can give a good prediction of Cd and Cl' for certain bluff bodies. But their vortex shedding freq are big different with each other.

PS: I am using the model on a low-Re mesh.

waynezw0618 · March 27, 2012, 09:02

Quote:

Originally Posted by lakeat

Hi Felix,

Hmm, thanks for sharing.

I am curious, have you used k-omega-SST for some bluff body flows. (Incompressible, 2D URANS)?

I can't get a correct vortex shedding freq with it. Any experience?

Generally, I found many models including the k-omega-SST can give a good prediction of Cd and Cl' for certain bluff bodies. But their vortex shedding freq are big different with each other.

PS: I am using the model on a low-Re mesh.

Hey Lao Wei`
How is going?
what you mean for low-Re mesh? how could you setup the boundary condition for all turbulent variables. in principle It should be zero for k at wall, while I guess most of us use zeroGradient, and for nut, it is confuse to use nut<>Wallfunction as the BC for wall as in the MotorBike example. for Omega wall BC, is there any literatures on the production term specified for sublayer like the implement in OF?

January 24, 2012, 17:16		#27
lakeat Senior Member Daniel WEI (老魏) Join Date: Mar 2009 Location: Beijing, China Posts: 689 Blog Entries: 9 Rep Power: 21	Hi Felix, Hmm, thanks for sharing. I am curious, have you used k-omega-SST for some bluff body flows. (Incompressible, 2D URANS)? I can't get a correct vortex shedding freq with it. Any experience? Generally, I found many models including the k-omega-SST can give a good prediction of Cd and Cl' for certain bluff bodies. But their vortex shedding freq are big different with each other. PS: I am using the model on a low-Re mesh. __________________ ~ Daniel WEI ------------- Boeing Research & Technology - China Beijing, China Email

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Use of k omega turbulence model	john_w	OpenFOAM Running, Solving & CFD	2	September 22, 2009 05:15
Discussion: Reason of Turbulence!!	Wen Long	Main CFD Forum	3	May 15, 2009 09:52
Fan heater model: what turbulence source to use?	andy20	Main CFD Forum	0	March 2, 2008 12:46
Advanced Turbulence Modeling in Fluent, Realizable k-epsilon Model	Jonas Larsson	FLUENT	5	March 13, 2000 03:27

March 15, 2011, 09:55		#22
FelixL Senior Member Felix L. Join Date: Feb 2010 Location: Hamburg Posts: 165 Rep Power: 18	I can't see the graphs you're talking about. Could you try to reupload them? Greetings, Felix.

March 15, 2011, 10:22		#23
timo_IHS Member Timo K. Join Date: Feb 2010 Location: University of Stuttgart Posts: 66 Rep Power: 16	I talked about the graph in post #19, sorry I didn't mention again. Or you also can't see this?

March 15, 2011, 10:25		#24
FelixL Senior Member Felix L. Join Date: Feb 2010 Location: Hamburg Posts: 165 Rep Power: 18	Sorry, no can do.

March 15, 2011, 12:52		#26
FelixL Senior Member Felix L. Join Date: Feb 2010 Location: Hamburg Posts: 165 Rep Power: 18	Thanks for the plot. Hm, well, it looks like when using (1) the simulation fails to reach a steady state. Is the guide vane in your test case possibly on the verge of separation? If so, your modification (2) probably removed the separation thus helping the simulation to converge. This would also explain why the separation point of your diffusor geometry isn't correctly predicted. If it's pretty sure that the guide vane flow remains attached for a whole range of different turbulent characteristics inside the flow field, the only thing I could imagine causing your trouble would be the mesh. Could be the wall functions you are using, too, but since I usually avoid to use these, I can't really give a qualified comment on that topic. Greetings, Felix.