[braennstroem]

Hi Niels,

forgot to say: thanks for the explanation!

Fabian

[ivan_cozza]

Hi Foamers!
I'm still struggling with the OF version of kOmegaSST:
it's not clear if the wallfunctions that are present in OF for this model require y+ near to 30 or less...
Any suggestion?
Thanks, Ivan

[ivan_cozza]

Hi Niels,
I'm managing to test your version of kOmegaSST on a smooth flat plate, as I need a lowRe version of kOmega, but I have some questions on your model:
1) It compiles on of 1.5?
2) How can I set d50 to work well on a smooth plate?
3) Have I to put fixedValue k = 0?
4) Have I to put zeroGradinet omega?
I will give you a report on how it works on a smooth plate!
Thank you!

[ngj]

Hi Ivan

Well, that is a lot of questions, thus I will try to answer them the best I can.

@1: I have not tried compiling it om 1.5, but one of my colleagues tried and as I recall was unsuccessfull due to the change in folder structure for the turbulence from 1.4.1 -> 1.5.

2. Wilcox (2006(7?)) states that it is sufficient to have k_plus < 5, but I have not been successfull in predicting the correct bed shear stresses in the smooth / lower-transitional range. Keep in mind, that it is mainly a _rough_ BC description.

3. No, you should be allowed to use either Dirichlet or Neumann conditions.

4. Omega is calculated and "hard-coded", thus it is changed every timestep as a function of bed shear stress according to the description in Wilcox (2007) / Roulund (2005?), though due to some consideration with respect to Reynolds independance as Re->inf, I would stick to the Wilcox formulation.

Best regards,

Niels

[ngj]

@(@1): That said, it should not be hard to take the 1.5 version of kOmegaSST and merge it with my kOmegaSSTRough from 1.4.1.

Good luck,

Niels

[ivan_cozza]

Ok Niels,
if I have understood, I have to try on my old 1.4.1 version.
Correct me if I'm wrong, but at a smooth wall (y+ less than 1) k = 0, so a dirichlet bc with 0 should work...
For omega, you mean that in the 0/omega file, in boundary definition on walls I can write whatever I want,as it calculates it by itself?
I have to apologies, I haven't read the wilcox(2006) paper, can you give me the reference?
Thanks, Ivan

[ngj]

Hi Ivan

Please bear with me, if I have underestimated you, but I believe you are confused about a couple of thing:

1. k_plus is the roughness height given in wall coordinates, and has nothing to do with the turbulent kinetic energy. k_plus is defined as k_N U_f / nu, where k_N equal 2.5d50.

2. When simulating a smooth surface, it is part of the definition of the viscous sublayer, that the molecular viscosity is larger than the turbulent one, thus the only physical meaningful value for k (turb. kinetic energy) is 0.

3. When resolving the boundary layer you do indeed need to resolve it having y+ in the order of magnitude 1.

4. This requirement is not explicitly coupled to the value of k_plus for smooth walls. The requirement for k_plus (less than 5), comes from the definition that the boundary is smooth if k_plus is _smaller_ than the thickness of the viscous sublayer, set to 11.7, thus if the roughness elements are fully covered by the viscous sub layer, then the surface is smooth. The "less than 5" requirement is probably some engineering rule of thump, which states that, then you are definitely below the limit Note that for k_plus (and d50) = 0, the definition for omega diverges, thus it is not at all recommended to go to the limit.

5. The reference to Wilcox is his book, I believe it is third edition. Search amazon or similar, and you will get the correct title.

Enjoy,

Niels

[ivan_cozza]

Hi Niels,
you're right, I'm confused about more than a couple of things, but I'm not a turbulenc modelling expert.
I work on Aeroacoustics, and I need a RANS calculation on a smooth flat plate just to begin my work, so I hoped to use simpleFoam as an user.
With the existing models (I tryed SST, lowRe launder sharma, lien cubic, etc) I didn't get sufficient results as in Star CCM+ with SST with hybrid wall treatment.
I hope that a low Re model for SST will help me, but I don't know anything about rough walls...
So, my questions are about the settings needed by your model to work, regardin k (tke), omega, and the roughness parameters to obtain a smooth wall.

Thank you for your clarifications, if I succeed to convert it in of 1.5 I will post it!

Regards,
Ivan

[ngj]

Hi Ivan

As I have said above, I do not have much experience with smooth walls, and the few trials I have carried out lead to error.

Thus I cannot help you much more, but I you have not read it already, you should read the posts from February 2008, which are mostly about turbulence and smooth walls, where I refer to Wilcox and cite some of the context.

Best regards,

Niels

[maddalena]

Hi all,

I am wondering if it is possible to apply a rough wall condition using the kEpsilon - realizableKE turbolence model in OF 1.5. Up to now, I have found references only to kOmega or Spalart-Allmaras model in this forum... And if so, how can I specify the dimensioned roughness height and the roughness constant as I would do in Fluent? Any idea? Any hint where to find more info?

Thanks in advance!

Mad

[braennstroem]

Hi,

I am trying to get some kind of enhanced or compound wall treatment into kOmegaSST. The switching between low-re and high-re bc is done similar to FLUENT's approach and Hanjalic/Popovac for the zeta-f model http://repository.tudelft.nl/file/354721/370010

Basically I just coupled the boundary conditions for omega, the turbulent kinetic energy production and viscosity for the low-Re and high-Re case using a function like exp(-Gamma) (like in FLUENT).
Right now, I set the values in the wall nearest cell (faceCelli), though I am not sure, if they have to get set at the boundary face (facei) as well!? A short test did not produce the desired result, so maybe I did something else wrong...

I would be glad, if some of you can test the model and make some suggestions for enhancement!


Fabian

[braennstroem]

Hi,

I am trying to get some kind of enhanced or compound wall treatment into kOmegaSST. The switching between low-re and high-re bc is done similar to FLUENT's approach and Hanjalic/Popovac for the zeta-f model http://repository.tudelft.nl/file/354721/370010

Basically I just coupled the boundary conditions for omega, the turbulent kinetic energy production and viscosity for the low-Re and high-Re case using a function like exp(-Gamma) (like in FLUENT).
Right now, I set the values in the wall nearest cell (faceCelli), though I am not sure, if they have to get set at the boundary face (facei) as well!? A short test did not produce the desired result, so maybe I did something else wrong...

I would be glad, if some of you can test the model and make some suggestions for enhancement!


Fabian

[braennstroem]

sorry,

[louisgag]

Dear Fabian,

I would love to try out your model as I run KOmegaSST turbulence model and have low and high Re regions.

You'll find my email by clicking on my user name.

Regards,

-Louis

[braennstroem]

Next try...
kOmegaSST_CWT.zip
Though, some more adjustments are probably needed.

Fabian

[philippebv]

Hi Fabian,

I know it's an old thread, but I think it is still very relevant to have a SST model with switching between low-re and high-re bc. Did you test further your model implementation since february?

I tested your implementation and I was wondering what you would suggest for the condition for k at walls? I see that you do not apply the checkkOmegaPatchFieldTypes.H verification. Is it because you set k=0 at walls? Or you still use zeroGradient?

Thank you,

Philippe

[ashish.svm]

I want to include frictional pressure drop calculation in rhoPisoFoam.

frictional pressure drop can be calculated by 0.5*ff*rho*U^2
where ff is fanning friction factor, which can be calculated by using standard correlations based on reynolds number.

I have modified UEqn as following.

fvVectorMatrix UEqn
(
    fvm::ddt(rho, U)
  + fvm::div(phi, U)
);

if(includeFrictionFactor)
{
  UEqn += fvm::div(phiFrictional, U);
}

phiFrictional is calculated as following.

phiFrictional =
    fvc::interpolate(fanningFrictionFactor)*
    (
        (fvc::interpolate(U) & mesh.Sf())
      + fvc::ddtPhiCorr(rUA, rho, U, phiFrictional)
    );

and added the following term in pressure equation

if(includeFrictionFactor)
    {
      pEqn += fvc::div(phiFrictional);
    }

but using this formulation leads to incorrect results.

Please let me know what is wrong in my formulation.

Or

Please suggest me some other solver which has the pressure drop calculation using fanning friction factor.

[fede32]

Originally Posted by gabriel

Hi Jason,

i have a similar test case. Do you have a method to calculate a mean pressure over a surface/cut. In paraview i can plot the pressure in the cut and estimate the averaged pressure over the cut, but do you have an idea how to get an exact value.

Bests Gabriel

You can do it in openfoam 6 from CLI:

postProcess -func "patchAverage(name=Inlet, p)"