# Length scales in LES jet

 Register Blogs Members List Search Today's Posts Mark Forums Read

 October 26, 2016, 13:19 Length scales in LES jet #1 New Member   UHGAR Join Date: Sep 2016 Posts: 12 Rep Power: 9 Dear Foamers, Can someone let me know, how to calculate Integral, Taylor and Kolmogrov length scales from oneEqEddy and Smagorinsky model? oneqEddy is outputting 'k' but not epsilon and Smagorinsky is outputting neither. 1) How can be able to get epsilon ? what relations I need to use between 'k' and 'epsilon' to get three length scales? 2) I heard, LES only models sub grid scales. Is it possible to get Integral length scales from LES simulation? How can I be able to do this? Thanks, UHGAR

October 27, 2016, 09:41
#2
Senior Member

Timofey Mukha
Join Date: Mar 2012
Location: Stockholm, Sweden
Posts: 118
Rep Power: 13
Quote:
 Originally Posted by UHGAR Dear Foamers, Can someone let me know, how to calculate Integral, Taylor and Kolmogrov length scales from oneEqEddy and Smagorinsky model? oneqEddy is outputting 'k' but not epsilon and Smagorinsky is outputting neither. 1) How can be able to get epsilon ? what relations I need to use between 'k' and 'epsilon' to get three length scales? 2) I heard, LES only models sub grid scales. Is it possible to get Integral length scales from LES simulation? How can I be able to do this? Thanks, UHGAR
Hi!

1) As you correctly point out the k for LES will only correspond to the energy of the subgrid scales, which you model. To get the the full k you should also add the resolved part of k, which you can get through the diagonal components of the Re-stresses (UPrime2Mean in OpenFOAM).
Dissipation is harder, since it involves derivatives.
The relations connecting k and eps to the Taylor microscale are given on page 2000 of Pope's "Turbulent flows".

Re_L = k^2/(eps*nu)

lambda = sqrt(10)*Re_L^(-1/2), where lambda is the microscale.

2) yes, it is possible, by computing two-point correlations of velocity in space, see the last appendix in
https://www.it.uu.se/research/public...4/2015-014.pdf

But there is no "neat" way to do this, I had to simply use probes to save the time-series of the velocity field at many points along the desired direction and compute the correlations using these huge datasets.

I realize my answer is perhaps to condensed to be very helpful, but hopefully better than nothing.

Best,
Timofey

 Tags length scales, les jet