CFD Online Discussion Forums

CFD Online Discussion Forums (
-   OpenFOAM Running, Solving & CFD (
-   -   DNS kolmogorov Spectrum (

sampaio July 26, 2006 09:20

Hi All, Has anyone be able
Hi All,

Has anyone be able to get a kolmogorov spectrum with dnsFoam (32x32x32 turbulence in a box )?

Please, if someone can help, my thesis deadline is approaching and I need help ...

The spectrum does not look right for me. Maybe I am doing something wrong with the output Ek.xy...

I assumed what is written in this file (Ek.xy) is just the Ek's, and not K's. I suppose the K's are not written because they are simply the sequency of numbers 0 1 2 3 4 ...

Otherwise, the spectrum does not make any sense...

Any help is appreciated.
Thanks in advance

jros July 26, 2006 13:58

I'm remembering some dns of ho
I'm remembering some dns of homeogeneous turbulence using spectral methods on 64^3 mesh in wave space, those where made in the 90ties. The kolmogorov spectrum was difficult to see, just starting.

So I think 32^3 using finite volume is too little to see a developed cascade (even 64^3).

Javier Ros

sampaio July 27, 2006 09:09

First of all, thanks a lot for
First of all, thanks a lot for the reply.

I am actually preforming LES in the 32^3 box, with almost no viscosity (high Re).

The resulting spectra is a straight line, which is good, but with the wrong slope, which is bad.

To make things worst, I am not sure wether the first Ek (in the file Ek.xy) corresponds to K=delta_K=Kmax/number_of_Ks or to K=0. My gess (after reading what the code does) is that it corresponds to K=delta_K, but that makes the slope even worst...

Can anybody tell me what the code is outputing in Ek.xy?

Javier, can you give me references to papers? I have a bunch of them, but not exactly what I need.

Thanks a lot,

jros July 27, 2006 09:58

The reference is V. Eswaran an
The reference is V. Eswaran and S.B. Pope, Direc numerical simulations of the turbulent mixing of a pasive scalar, Phys. Fluids 31, 506 (1989).

As you can see is quite old and is focused in the scalar fiel, though I thik it deal with velocity spectrum.

Nevertheless I'm sure there are much better simulations done after this.

I abandoned the subjet in 97.


All times are GMT -4. The time now is 12:59.