[enricodl]

Hi everyone,

I am running a Homogeneous Isotropic Turbulence (HIT) case with LES (Smagorinsky) on OF2306 .
The case is a tutorial in:
$FOAM_TUTORIALS/incompressible/pimpleFoam/LES/decayIsoTurb

The case is already set up to export the energy density spectrum.
The problem is that the result spectrum decays in time much quicker than the reference case ( in 1 simulation second it decays 2 orders of magnitude more than CBC).
The reference is from Comte-Bellot and Corrsin wind tunnel(1971).

Things that I have tried:

1) remove LES model. by running the simulation in "laminar" mode, the inertial range decays the same quick, but there is a pile up of energy at the unresolved modes (the spectrum, instead of being monothonic decreasing, presents a local maximum after the filter cutoff mode), as expected.

2) reduce viscosity from 1e-5 (default value in the tutorial) to 1e-6 (value close to the one obtained from a dimensionless analysis to fix the same Re_lambda as CBC ( Re_lambda=71.6 , see Pope, Chp. 6 Sec. 3 ).
It didn't almost change the result at all. It make look like there is a "residual" (sorry for the mis-use of notation) viscosity that is much bigger than the Nu. Though, I'm not using upwind or any diffusive scheme ( I guess? Can you recognise the source of this diffusion in the fvSchemes and fvSolution case files.)

3) increase mesh from 64^3 to 128^3 : didn't change much, since 64^3 is already documented to be resolving this case ( classical HIT with LES )

4) (other less relevant things, like trying v.2206 and 2312)

To reproduce the behavior just:
-copy case from OF tutorials to your directory
-remember to change:
endTime= 2
writeFormat=ascii
-you can use Allrun (it's a small case and using the cluster isn't necessary)