[GiannisgR]

Hi OpenFOAMers,

Could anyone, please, send me the paper for the compressible k-omega SST, Sandy, Henry?. Additionally, for which y+ does this model work, meaning the applicability of the used wall function.


Menter, F., Esch, T.
"Elements of Industrial Heat Transfer Prediction"
16th Brazilian Congress of Mechanical Engineering (COBEM),
Nov. 2001

Thanks,
giannis

i.asproulias@gmail.com

[bastil]

Originally Posted by Peter85

Hey @ all!
But, If I have a mehs y+<1(low-re), and I'd like to use the k-omega-sst modell, what boundary conditions for the walls do I have to take? Zerogradient for omega and k with very low values (10^-8), and calculated for alphat and mut (or other values?)? Or do I have to take a wall function (like mutlowrewallfunction) for one of those?

Peter, all,

I am also wondering about that. Henry mentioned in his posts above ( #54, #56) that one has to use mutLowReWallFunction in case you use a Low-Re model but you do not have y+<1 at this specific wall. Otherwise, if you have y+<1 you have to use "no" wall function - but howe to define this in 0/nut?. I am also wondering how to define the value sin k/epsilon/omega-Files.
Regarding k-omega SST it seems to me like you can not use this approach at all since it is a Hi-Re model...

[ivan_cozza]

Originally Posted by bastil

Peter, all,

I am also wondering about that. Henry mentioned in his posts above ( #54, #56) that one has to use mutLowReWallFunction in case you use a Low-Re model but you do not have y+<1 at this specific wall. Otherwise, if you have y+<1 you have to use "no" wall function - but howe to define this in 0/nut?. I am also wondering how to define the value sin k/epsilon/omega-Files.
Regarding k-omega SST it seems to me like you can not use this approach at all since it is a Hi-Re model...

If I understood SST, it works quite fine at all y+ conditions (this is the reason why using SST instead of k-Omega).

If you have y+ < 1, then
k -> fixedValue, very small (1e-9)
Omega -> omegaWallFunction (is not really a wall function, is the b.c. defined by Menter for Omega, it works at all y+ I suppose).
mut -> fixedValue, very small (1e-9)

If you have y+ > 30, then classical wall function setup

If you have 1 < y+ < 30, then
k -> zeroGradient
Omega -> omegaWallFunction
mut -> mutSpalartAllmarasWallFunction (is a kind of all y+ mut wall function, sometime ago Henry said that this should be renamed in mutSpaldingWallFunction...)

Hope this can help...

Btw, is someone working on transition model for SST?

[alex_rubel]

Hi everyone, this thread is very intersting. I just wonder one thing, I'm using OF1.6 and I use nut in my 0 file is it mut or nut ?

[GiannisgR]

Thanks guys for the replies. I was thinking to switch off the wall functions, which are used automatically for every turbulence model in 1.6v (I don't know about 1.7v) and apply manually the BC for the turbulent quantities. This is because I have y+<1 everywhere. Thanks again

Giannis

[Arnoldinho]

Originally Posted by ivan_cozza

If I understood SST, it works quite fine at all y+ conditions (this is the reason why using SST instead of k-Omega).

If you have y+ < 1, then
k -> fixedValue, very small (1e-9)
Omega -> omegaWallFunction (is not really a wall function, is the b.c. defined by Menter for Omega, it works at all y+ I suppose).
mut -> fixedValue, very small (1e-9)

If you have y+ > 30, then classical wall function setup

If you have 1 < y+ < 30, then
k -> zeroGradient
Omega -> omegaWallFunction
mut -> mutSpalartAllmarasWallFunction (is a kind of all y+ mut wall function, sometime ago Henry said that this should be renamed in mutSpaldingWallFunction...)

What about the case when I have a combination of both in one model domain due to different grid resolutions and changing flow regimes? Is there an automatic switch betweeen low-RE and high-RE implemented, e.g. by setting 'wall functions' for all quantities (e.g. omegaWallFunction, kqRWallFunction, nutRoughWallFunction in incompressible case)?
... I also posted this question here in more details.

Arne

[mai]

Originally Posted by ivan_cozza

If I understood SST, it works quite fine at all y+ conditions (this is the reason why using SST instead of k-Omega).

If you have y+ < 1, then
k -> fixedValue, very small (1e-9)
Omega -> omegaWallFunction (is not really a wall function, is the b.c. defined by Menter for Omega, it works at all y+ I suppose).
mut -> fixedValue, very small (1e-9)

If you have y+ > 30, then classical wall function setup

If you have 1 < y+ < 30, then
k -> zeroGradient
Omega -> omegaWallFunction
mut -> mutSpalartAllmarasWallFunction (is a kind of all y+ mut wall function, sometime ago Henry said that this should be renamed in mutSpaldingWallFunction...)

Hope this can help...

Btw, is someone working on transition model for SST?

Hi Ivan,

what is the reason, that for the zeroGradient BC of k. Why is it not kRWallFunction?
Please see also the post under the following link:
http://www.openfoam.com/mantisbt/view.php?id=179
What is correct?

Martin

[arunraj]

Hi Guys,

First very sry to disturb u people...Bt I need some important help

I have a created a domain as Domain1 which is Fluid domain

Now I want to create a sub-domain for Domain1 which is a porous domain

Is it possible to create like this?

If so can u explain how?

[Aerospace]

Originally Posted by henry

If your near-wall y+<1 everywhere you can use a low-Re model; you don't need wall-functions at all.

H

In order to know the y+ when the grid is in the laminar region (Low-Re model, no wall function or nutLowReWallFunciton if y+>1), is it possible to employ the yPlusRAS utility provided by OpenFOAM? or it calculates the y*.

It is not cleary explained in other topics like this: http://www.cfd-online.com/Forums/ope...tml#post306730

Thank you very much in advance

[1/153]

Originally Posted by ivan_cozza

If I understood SST, it works quite fine at all y+ conditions (this is the reason why using SST instead of k-Omega).

If you have y+ < 1, then
k -> fixedValue, very small (1e-9)
Omega -> omegaWallFunction (is not really a wall function, is the b.c. defined by Menter for Omega, it works at all y+ I suppose).
mut -> fixedValue, very small (1e-9)

If you have y+ > 30, then classical wall function setup

If you have 1 < y+ < 30, then
k -> zeroGradient
Omega -> omegaWallFunction
mut -> mutSpalartAllmarasWallFunction (is a kind of all y+ mut wall function, sometime ago Henry said that this should be renamed in mutSpaldingWallFunction...)

Hope this can help...

Btw, is someone working on transition model for SST?

Dear all,


I've been testing the "hybrid" or "enhanced" or "automatic" wall treatment. One thing I found nutSpaldingWallFunction is said to be a continuous WF, however, I cant get a good result with it. I am wondering anyone of you have success with nutSpaldingWallFunction?

Model: kwsst
Test cases: channelflow, periodic hills.

I set omega as omegaWallFunction, k as kqRwallFunction, nut as nutSpaldingWallFunction. And the results are not good. Separation and reattachments pts are all not good.



Another question is that, in Menter's paper of improved wall treatment, it is clear that omega is written as sqrt(sqr(omega_vis)+sqr(omega_log)). But in OF-2.0, the utau is calculated with sqrt(k), is this correct?

And also I am wondering why do we need a nut wall function, since the nut can be calculated simply by k and omega, once we have omega already set with certain wall function, then nut is calculated, why need nut wall function then?

Thanks!

[TengWU]

Originally Posted by GiannisgR

Hi OpenFOAMers,

Could anyone, please, send me the paper for the compressible k-omega SST, Sandy, Henry?. Additionally, for which y+ does this model work, meaning the applicability of the used wall function.


Menter, F., Esch, T.
"Elements of Industrial Heat Transfer Prediction"
16th Brazilian Congress of Mechanical Engineering (COBEM),
Nov. 2001

Thanks,
giannis

i.asproulias@gmail.com

Hey Giannis,

Have you got the paper "Elements of Industrial Heat Transfer Prediction"? If so, could you kindly send a copy to me?
tengwu86@gmail.com

Thanks,

[TengWU]

Originally Posted by henry

Here is a corrected version of kOmegaSST.C
kOmegaSST.C
to replace the file in OpenFOAM-1.4.1/src/turbulenceModels/compressible/kOmegaSST

I have tested the model runs and give sensible results but I have not performed a rigorous validation.

Hey Henry,

Could you kindly send a copy of the paper "Elements of Industrial Heat Transfer Prediction" to me?
tengwu86@gmail.com

Thanks,

[dogan]

Originally Posted by henry

I am referring to nutSpalartAllmarasWallFunction for incompressible flow and mutSpalartAllmarasWallFunction for compressible flow. Even though these were created for use with the Spalart-Allmaras model they are not dependent on this particular model in any way being generic implementations of the Spalding continuous wall-function using U rather than k as the controlling variable and can be used with other turbulence models.

We tested the kOmegaSST model with the nutSpalartAllmarasWallFunction wall-function and obtained good results, as good as others have obtained with adaptive/continuous wall-functions.

H

Dear Henry,
i would like to ask a question about the nutUSpaldingWallFunction. I have used this wall function in a diffusor test case with kOmegaSST turbulence model, and i also made the same simulation with CFX, then i compared the results. since i know, CFX is using the so called automatic wall treatment, i would like to be sure about that if nutUSpaldingWallFunction is also an automatic wall treatment, or in OpenFOAM is there anouther way to make an automatic wall treatment?
Thanks

[Claudio87]

Hi foamers!

I find this old thread, very interesting for me in this moment, but I still can't find an answer to my problem...

Here there is a description of my case (it is 3D).
I have my geometry inside a box, but its faces (of the box) must be considered as patches, not as walls. The fluid moves from the right to the left. The flow comes from the pipe on the right, but also from the opening on high.
The picture attached is a slice (coming from the laminar case).
To have an idea about the dimensions, the pipe on the right side has D=0.0025 m.

Now my problem: which BCs I have to put for the simulation?!
Of course, as wrote in the title, the case is compressible.
I used, for a first attempt, wall functions for the geometry (for all k, omega, alphat and mut);
then, I'm not sure about which could be right BCs on
1) the "external" faces (of the box)
2) the inlet (that of the pipe), "ingressoOX" in the folder

I put "fixedValue" on both.
Maybe, could I use
- "turbulentMixingLengthFrequencyInlet" for omega
- "turbulentIntensityKineticEnergyInlet" for k
on inlet ("ingressoOX") patch?

No problem for the outlet of the geometry, on the left side, because it is "free"...

Moreover, have the values I used any sense? or which values should I use?!


I attach also a folder with these BCs I used for now.


Thank you in advance for your suggestions!

Best regards,
Claudio


PS: I'm using rhoSimplecFoam as solver

[MaLa]

HI henry

I needed a little help. I am trying to adapt kOmegaSST model to kOmegaSST -EARSM model and wish to show its performance for simple developed channel flow, to begin with.

To get on this track, I first thought about testing kOmegaSST for channel flow and seeing if the results I get compare well with what other researchers have found.

Here's the thread that talks about my attempts:

http://www.cfd-online.com/Forums/ope...nnel-flow.html

I know that you implemented kOmegaSST in openFoam and hence wanted to ask you if you ever validated it for channel flow and if yes, would you mind sharing the case files.

Have been stuck on this for a little while and don't really know whom to ask. Any help is appreciated. Thanks.

[user10600]

Hi everyone!

A very interesting is this topic!

I have also some questions of my own regarding the k-omega SST model... yes, one more Foamer who is left in the doubts.. ^^

To sum up quickly, as I lacked theoritical background in CFD (I am relatively new in this field), I started to investigate more about the whole process which is behind all CFD Simulations. Now, I'm putting a lot of interests in the turbulence models (mainly k epsilon, k omega and also k omega SST). I want to link the formula that we can read in the litterature with the actual formula used in the OpenFOAM source code.

My biggest doubt lies on the k omega sst model. My "theoritical" sources about the k omega sst formulation and its representation in PDE come from :

• Ten Years of industrial experience with the SST Turbulence Model, from F. R. Menter, M. Kuntz and R. LAngtry
• the website from the nasa dedicated to this model: http://turbmodels.larc.nasa.gov/sst.html
• CFD Online forums (maybe I miss an other interesting thread about this subject ?) and wiki
• CFD Books and Dissertations, but I have worked mainly with: The Finite Volume Method in CFD, from F. Moukalled, L. Mangani and M. Darwish

Now my questions :

• I went to the source code of the k omega SST model, first in the incompressible case. My first question is about the definition of arg1. It is usually written in the litterature like this way (I am simplificating):

arg1 = min ( max(term 1); term 2 )

but still, in the OpenFOAM source code, we can find it like this way:

arg1 = min [ min ( max(term 1); term 2 ) ; 10 ]

Why this change of formulation ? Does anyone would have an explanation or even better, a paper that relates about this.

• My second question is about the compressible case. In general, how do we "switch" from the incompressible formulation to the compressible formulation ? For instance, I always notice the apparition of a new term with a factor "2/3" on the RHS of the PDE. Where does it come from ? I read somewhere that it might come from the bulk viscosity... but honestly that doesn't mean anything to me ^^ It would be great if someone could give the big lines (or litterature) to go, for example, from the k epsilon in incompressible formulation to the k epsilon in compressible formulation.

I thank you all in advance!


Have a nice evening, I'm looking forward to read your answers!

[user10600]

Hi again!


I forget to mention that I was using OpenFOAM 2.0.x.

I rethink of that this morning, about the

 arg1 = min ( term 1 ; 10 )

The value of "10" might be here to modelize the upper limit of the viscous sublayer ?! A paper about this specific formulation of the k omega SST would be nice though.

EDIT:
About the modelling of the k omega SST formulation in a compressible case: I guess the term with a factor "2/3" on the RHS of the PDE comes from the formulation of the Boussinesq Approximation in a compressible case (2nd term):


Does anyone could confirm or infirm my thoughts ?

Thanks in advance!

[usv001]

Originally Posted by user10600

Hi again!


I forget to mention that I was using OpenFOAM 2.0.x.

I rethink of that this morning, about the

 arg1 = min ( term 1 ; 10 )

The value of "10" might be here to modelize the upper limit of the viscous sublayer ?! A paper about this specific formulation of the k omega SST would be nice though.

EDIT:
About the modelling of the k omega SST formulation in a compressible case: I guess the term with a factor "2/3" on the RHS of the PDE comes from the formulation of the Boussinesq Approximation in a compressible case (2nd term):


Does anyone could confirm or infirm my thoughts ?

Thanks in advance!

Hi there,

You are right about the "2/3" term. It indeed comes from Boussinesq's eddy viscosity hypothesis. The point is that we are replacing the Reynolds stress tensor with a model but we still want to enforce that the trace of the tensor remains at . Since the first part of the model is deviatoric (traceless), the enforcement must come from the second term. But note that this assumes that the turbulence is homogeneous, i.e. .

Now to my problem. I share the same concerns about the appearance of the limiting constants in 'arg1' and 'arg2'. Also, I found some inconsistencies in the 'omegaEqn' and 'kEqn' as well. I have modified the current kOmegaSST model (OF-2.4.0) to match with what has been given in Menter's paper exactly. I would like to know if anybody has a good validation case, such as compressible pipe flow or Couette flow, which I can use to test the model.

USV

[wc34071209]

Hi,

I would like to know whether the k-omega SST model implemented in the recent versions of OpenFOAM is still a high-Re model. Or low-Re damping features have been added?

Thank you.