open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi everyone,

I'm attempting to model a thermal plume discharged into a bay using Telemac 3D. Since my vertical resolution is quite high (~0.4m) and my horizontal can be quite low (~30m in the far-field), I've decided to use k-epsilon in the vertical and experiment with other models in the horizontal (I'm leaning toward Smagorinsky at this point due to it's consideration for grid resolution).

First, however, I've decided to check sensitivity to the horizontal turbulence model in a simplified domain with a passive scalar discharged at a single boundary (uniform with depth).

Interestingly, I find no difference in tracer concentration (or velocity) between "Constant Viscosity" (at 1e-6m2/s), "k-epsilon", and "Smagorinsky" when k-epsilon in used in the vertical!

There are slight differences, however, between the three horizontal methods when constant viscosity is used in the vertical (instead of k-epsilon). Of these runs, the run with k-epsilon in the horizontal and constant viscosity in the vertical is actually the same as all the runs with k-epsilon in the vertical.

It seems that k-epsilon trumps all other models when it is used in either the horizontal or the vertical.

Does anyone have any insight on this?? More generally, does anyone have any input on consideration for horizontal resolution in choosing an appropriate closure model?

Thanks in advance and sorry for the long post!

Cheers,

Ben

Hello,

Actually when k-epsilon is asked in one dimension it is imposed on both, it explains your tests and you came to the right conclusion (congratulations...). This is something we should relax even if turbulence specialists do not like it (k-epsilon turbulence is isotropic...), because we have, especially in your case, an important aspect ratio. I would hint that doing this is only a few lines to cancel but that remains to check. It is on my todo list, I hope I can do it before retiring (i.e. very soon).

With best regards,

Jean-Michel Hervouet

Ah! Thanks for the reply! I thought I was loosing my mind there for a little bit