open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello everyone,

I am trying to simulate a curved-channel lab setup with an submerged spur on the outer bank. My aim is to see how it interacts with the secondary helical flows that occur in rivers. However, due to only one element covering the whole vertical wall, velocity and water values obtained near the wall are very poor. Since the spur will have flow over it, I cannot make it a closed wall, as I have done always. I have previously had similar problems with telemac-2d when simulating over-topping of banks of a rectangular channel. I thought telemac-3d could handle problems including vertical walls better. But then I realized it uses a transformed moving mesh and thus the problem still remains.



As you can see, still only one prism covers the sharp gradient and hence flow's interaction with the spur is not obtained.
I wanted to know if there is any way to simulate such systems. Any advice or help in this matter will be extremely valuable.

Hi,

You should increase the resolution of your mesh over this singularity to improve its representation and the model results.

Regards,
Laurent

Yes, that would have been ideal, but as I said it is a vertical spur. No amount of refinement of mesh will increase the number of elements over the vertical face, until I use elements of size smaller than bathymetry resolution.

I actually wanted to know if perfectly vertical structures were impossible to be used in Telemac3d simulations. I know that vertical structures are impossible to properly simulate in Telemac2d, but what about Telemac3d?

Have a look at these papers:
hal.archives-ouvertes.fr/hal-01702813/document
henry.baw.de/bitstream/handle/20.500.119...quence=1&isAllowed=y

If you're looking for more accurate results, you should use a CFD model.

Regards,
Laurent

I thought of CFD but the domain is large enough to be unfeasible to run full cfd simulations. I will try to incorporate those papers though. Thank you very much for the links.