open TELEMAC-MASCARET The mathematically superior suite of solvers

Dear All,

This is in regard with Telemac3D simulation results. When I run simulation in parallel and when try to combine things with the utility gretel_autop, I have the option of combining T3DRES and T3DHYD.

My question is: is the value in T3DHYD would be same as what you will get from using Telemac-2D for the same case. I think not.

Please commnet.

Thanks,
Sumit

Hello,

In T3DHYD you have the depth-integrated results, it means that the 3D velocities, tracers, etc. are depth-integrated. This is different from what Telemac-2D would get.

Best regards,

Jean-Michel Hervouet

Thanks a lot, I thought so.

Hello Jean-Michel,

Let me bother you with another question. I was wondering, is it possible for the same set of boundary conditions (discharge + stage) the free surface elevation results that you get from Telemac-3D and Telemac-2D can be entirely different ?

Thanks a lot for all your inputs.

Best regards,
Sumit

Hello,

Yes it is possible, e.g. the solution for an hydraulic jump is totally different, or the wave in a square basin is totally different if you try hydrostatic and non hydrostatic. Even a flow over a bump will give a different discharge (3D is right). However you can test the sensitivity to numerical parameters, which may have also an influence.

With best regards,

Jean-Michel Hervouet

Hi Jean-Michel,

I had this question in the first place because, I was simulating the flow in a river confluence. I had two inflow boundary conditions and one downstream stage boundary condtion.

There was one validation point with stage observation inside the domain. Initially I used Telemac-2D for this problem and by varying the bottom roughness (Manning's n) I was able to match the observed and predicted stage in a perfect manner.

But after this when I used the Telemac-3D with same setting and roughness, the stage-comaparison at the observation point was not at all good. It makes me wonder how can I calibrate Telemac-3D to give the desired results and is the calibration process for Telemac-3D entirely different from the Telemac-2D, your comments.

Best regards,
Sumit

Hello,

The basic friction laws, such as Chézy and Strickler, have been programmed to give the same effect in 2D and in 3D, in simple cases (see e.g. the test case called canal). If you use a Nikuradse friction law, it is a different process than these, it gives probably a better result as soon as there is enough refining on the bottom. But then you have the 3D and non hydrostatic effects.

JMH

Hello Jean-Michel,

Thanks a lot for your reply, I have been reading a little more on this, especially from the book: "Hydrodynamics of Free Surface Flows: Modeling with the finite element method". In the section, "Equivalence of free coefficients in 2D and 3D, page 48-49 Chapter-2" it is explained that friction coefficent are incorporated in such a manner that it has the same effect in 2D and 3D.

Now my question is that for the same boundary conditions as I was mentioning before and same manning's n why is free surface elevation at a point inside the computational domain different in Telemac-2D and Telemac-3D. Is there a keyword that I need to switch on in 3D for the friction coefficient to have the same effect as in 2D ?

I very much appreciate all your inputs.

With best regards,
Sumit

Hello,

No, there is no specific keyword, so the differences you observe could be either due non hydrostatic effects (you could also compare hydrostatic and non hydrostatic option in Telemac-3D) or to numerical parameters.

JMH

Hello,

I'm running a 3D model in non-hydrostatic mode. I'm using Nicuradse's law and I was wondering whether the paragraph "Equivalence of friction coefficients in 2D and 3D" (pp. 52/53) is also valid for Nikuradse or if there is a different relation?

Best regards,

Constantin