open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi!

I am modelling an outlet composed of 2 jets and diffusing in the surrounding water.
The outlet water is warmer than the surrounding water (+2 degrees).
I have ADCP measurements that show that the jets are first flowing in the upper part of the water column and then progressively plunging in depth (where the water temperature also increases, approx. +0,2 degree / m).

I set up a T3D model for this case that I run in hydrostatic mode, with k-epsilon and assuming a constant water temperature as the layering is not stable (excepted for the outlet). As you can see in the attached document the model reproduces roughly the measured flow conditions but with a higher (vertical) diffusion of velocities.

I tried to change a couple of parameters (turbulence models, non-hydrostatic) but the results remain roughly the same. My cas file is attached.

I was wondering if someone has some experience or advice to share that would allow a better reproduction of the plunging phenomena.

Thank you in advance!

Best regards
PL

Hello Pierre-Louis

Sorry for this very late reply.

I'm wondering if your problem is improvable again or not, at least with the last release of Telemac.

We know that in these cases, obtained results are very sensitive to the mesh transformation, time step, the number of vertical layers, AMR ...
As for a calibration exercise, you need to find the optimal value for all these parameters.

Moreover, for the next release, the new advection schemes will be added and they promise a remarkable decrease in numerical diffusion.

with my best regards

Riadh ATA

Hello Riadh,

Thank you for your reply.
I think I did test more or less all the configurations that could have helped but I did not notice any dramatic improvements and was unfortunately limited in time and budget (HPC is fast but also expensive!).

I was suspecting some numerical diffusion to take place here, as these results are similar to what other colleagues also observed in their models (with strong diffusion).

Anyway, this is absolutely great news that new improved advection schemes are being added to the next version! I will give it a try.

Best regards
PL