[Nek5000-users] turbulent 3d box

[nek5000-users at lists.mcs.anl.gov]

Personally, I'm partial to filtering (for lack of a better
alternative at present).

I'm running with lx1=8, say, I would set parameter 103 to 0.1
and parameter 101 to 1 in the .rea file.

This would imply:

   Nth mode is reduced by 0.1 at the end of each timestep

   N-1st mode is reduced by 0.1*(1/4) at each step.

The multimode filter (anytime p103>0 and p101>0) reduces
the Nth mode by p103 and additional modes N-k, k=1,...,p101,
by a quadratic fit.

For LES, you probably want to remove energy from ~25% of the modes,
with a smooth roll-off, as given by the quadratic fit.  Thus, for
lx1=12, I would take p101=2, which would result in modes N, N-1, and N-2
being reduced.

Paul


On Sat, 18 Jan 2014, nek5000-users at lists.mcs.anl.gov wrote:

> Dear Paul,
> Regarding the issue of DNS/LES. How I define what is the turbulence model is 
> used?
> Thanks,
> Barak
>
>
> _______________________________________________
> Nek5000-users mailing list
> Nek5000-users at lists.mcs.anl.gov
> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users
>