open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,

I have a very simple question and do apologise in advance if the answer is trivial.

Let NPOIN2 and NELEM2 be the number of nodes and elements of a mesh in 2D.
Let NPOIN3 and NELEM3 be the same in 3D and let NPLAN be the number of planes of the 3D mesh.

We can write:
(1) NPOIN3 = NPOIN2 * NPLAN

Then, as 3D elements are prisms defined by 6 vertices (the same 3 nodes at two neighbouring layers according to the theory and user manuals), I would tend to write:
(2) NELEM3 = NELEM2 * (NPLAN-1)
which altogether make up the whole volume of the fluid domain.

In reality, NELEM3 is much greater, see attachment for 2D and 3D characteristics of a same model with NPLAN = 5 as outputted by T3D. In that example, Eq. (1) is verified but (2) is not.
I can't really figure myself what these so many elements look like...

Does anybody have an explanation on how the elements are defined in Selafin files?

Thank you.
JBS.

Hello,

You are nearly always right.
I think your "In reality" is questionable.

If you run a TELEMAC-3D simulation and look at the listing file, you have some information about the mesh, number of nodes and elements both in 2D and 3D (see e.g. T3D gouttedo example):

---

2D MESH

---

2D ELEMENT TYPE : 10
NUMBER OF 2D NODES : 4624
NUMBER OF 2D ELEMENTS : 8978
NUMBER OF 2D BOUNDARY NODES : 268

3D MESH

---

3D ELEMENT TYPE : 40
NUMBER OF 3D NODES : 13872
NUMBER OF 3D ELEMENTS : 17956
NUMBER OF LEVELS : 3
NUMBER OF BOUNDARY ELEMENTS : 536
TOTAL NUMBER OF BOUNDARY NODES : 10052
INCLUDING LATERAL BOUNDARIES : 804
SURFACE : 4624
BOTTOM : 4624

---

Try to rerun your simulation, and you should check what you wrote before.

I wonder if the information you have in your figure are not due to your post-treatment software?

Hope this helps,

Chi-Tuan

Thank you Chi-Tuan,

I tried opening the files with another tool and my expectations were validated, as you mentionned. So it seems that this is a bug in Blue Kenue for 3D Selafin files... to be reported to Alan!
I do that right now.

Thanks again,
JBS.

Please try the new alpha that will be created after today (15 May 2024) and let me know how it goes...

Kindly,
Alan

PS See also #44726 and #44727