[hz283]

Dear All,

After finishing a LES case using OPenfoam, I have the thinking to check how much turbulent kinetic energy (for instance 70% or 80%) is resolved based on this case's mesh and results. This can give me some feedback about the mesh or some numerical schemes. Does anybody know how to do this for the Openfoam LES results?

best regards,
H

[julien.decharentenay]

Hi H,

In my opinion, you can do this in two ways:

1) From the LES analysis you will have access to the mean and rms of the resolved velocity - you will want to check that the rms of velocity are converged. You can then calculate the resolved tke based on the rms of velocity. The unresolved can be calculated from the contribution of the SGS turbulence model.

2) Alternatively you could also look at the spectral distribution of the resolved turbulent structure using spatial FFT. This may have issues when coherent structures.

[hz283]

Originally Posted by julien.decharentenay

Hi H,

In my opinion, you can do this in two ways:

1) From the LES analysis you will have access to the mean and rms of the resolved velocity - you will want to check that the rms of velocity are converged. You can then calculate the resolved tke based on the rms of velocity. The unresolved can be calculated from the contribution of the SGS turbulence model.

2) Alternatively you could also look at the spectral distribution of the resolved turbulent structure using spatial FFT. This may have issues when coherent structures.

Hi Julien,

Thank you so much for your help. For the method 1 you mentioned, I need to obtain the mean and rms quantities on a run-time basis. I think I can these quantities from UPrime2Mean. But I found that even if I defined the fieldAverage1 in controlDict, I can not use these quantities in, e.g. createFields.H files.

So how can I access to the these mean and variance during the simulation?

Thank you.

[julien.decharentenay]

Hi H,

You may not need to do it during the simulation (i.e runtime), but just as a post-processing.

From a practical perspective, I am not able to help... I have done it in the past but with a different solver (star-cd) at post-processing stage only. I can't help with openFoam side...

Julien

[wenxu]

Dear Julien,

Could you give some more details or some references how to calculate the resolved kinetic energy and unresolved kinetic energy?

The resolved kinetic energy is equal to k? K=0.5*sqrt(U-Umean)?

which variable can be characterize the unresolved kinetic energy? Could you give the formulation of them?


Thank you in advance!

Best,
Xu

[openfoammaofnepo]

Pls see below:

https://tcg.mae.cornell.edu/pubs/Pope_NJP_04.pdf

[wenxu]

Originally Posted by julien.decharentenay

Hi H,

In my opinion, you can do this in two ways:

1) From the LES analysis you will have access to the mean and rms of the resolved velocity - you will want to check that the rms of velocity are converged. You can then calculate the resolved tke based on the rms of velocity. The unresolved can be calculated from the contribution of the SGS turbulence model.
.

I have read the paper by Pope, thank you very much. But could you give some more details how to calculate the unresolved part? e.g. Could you give the formulation of it?

What do you mean " calculated from the contribution of the SGS turbulence model."??

Thank you again!

Best,
Xu

[openfoammaofnepo]

probably you can think about how the k is calculated for the LES model?

[anishtain4]

I know this might be a little late for you, but in an LES solution the unresolved energy is simply:

autoPtr<incompressible::turbulenceModel> model
(
incompressible::turbulenceModel::New(U, phi, laminarTransport)
);
volScalarField k(model->k());

[wenxu]

Originally Posted by anishtain4

I know this might be a little late for you, but in an LES solution the unresolved energy is simply:

Thank you! Yes, I have solved it!

Best regards,
XU

[arone]

How to calculate resolved and unresolved kinetic energy in paraview from openfoam data?

[Bashar]

Hi Wenxu

Can you please share how to practically solve it? i.e. the step to do that? if you have time.
Thank you

[wenxu]

It is not very difficult. You can output this field, volScalarField& kToBeUsed = turbulence->k(); Then do what you want to deal with the kToBeUsed field. If you do not want to ruin the original field of k, use: volScalarField kToBeUsed = turbulence->k();