open TELEMAC-MASCARET The mathematically superior suite of solvers

Good morning,

I use v8p5r0.
I'm tuning mesh generations of a very shallow domain (about 140 km x 40 km x 70 m).

My computations are based on the example 'Vent'
for which I use the default value for VERTICAL TURBULENCE MODEL rather than 2.

If MESH TRANSFORMATION = 1 and NUMBER OF HORIZONTAL LEVELS = 3 (or more),
the call of OV subroutine is required to achieve my computation,
with a value of C changing with the 2D mesh refinement.
For instance :
if the edge length is about 1000 m, C must be greater than -5.4 m
and if the edge length is about 7000 m, C must be greater than -1.0e-4 m

If MESH TRANSFORMATION = 1 and NUMBER OF HORIZONTAL LEVELS = 2,
my computation well achieve without OV.

I also checked MESH TRANSFORMATION = 3 to managed crushed elements and tuned :
NUMBER OF HORIZONTAL LEVELS = 3
MINIMUM DISTANCE BETWEEN PLANE CLOSE TO THE BOTTOM = 0.
and THRESHOLD HEIGHT BEFORE CRUSHED ELEMENTS = 0.1 , 1. or 10. ;
and the call of OV is still needed to achieve my computations
(value of C of the same order than MESH TRANSFORMATION = 1 tests).

I also checked 2D meshes with refinement at slope regions and OV is still needed.

Is there a way I tune these cases without OV (or with C = 0.) ?

The edge length I foresee for this domain won't exceed 7000 m.

Please find attached a folder containing my basic cases for edge length = 1000 m and 7000 m :
there are also views of the water depth and 2D meshes if needed (see /postel/xxxxxx_NL02_C-00).

Kind regards,
Audrey

Hi Audrey
OV is a kind of operator on Telemac array.
It's not clear what is the problem you encountered with OV in your post.

Regards

Hi Christian,

The OV subroutine works perfectly to me.

My difficulty is that I am not able to run a computation (or build a 3D domain) with the original bedrock data of my region of interest.

I noticed that the computation passes with adding a certain water column length (via C < 0 in the OV subroutine).

What I observed is that :
there's a critical value of C for which the computation passes or not ;
and this critical value significantly changes with the characteristic edge length of the 2D mesh ;
for instance :
ABS(C) must be greater than 5.4D0 for edge elements of 1000 m order,
and ABS(C) must be greater than 1.0D-4 for edge elements of 7000 m order.

The computation passes well - without OV - when NUMBER OF HORIZONTAL LEVELS = 2.

My difficulty occurs when NUMBER OF HORIZONTAL LEVELS >= 3.

I mentioned in my 1rst message what I attempted else to make the computation pass.

The case I'm attempting to run is based on the T3D example 'vent' and only aims to manage the 3D mesh generation.

I don't see what I can do else to build my 3D domain with the original bedrock data only (without calling OV).
Is there a way I can do it ? Did I miss something in the computation procedure ?

If needed, you'll find attached the files I used to run the cases I mentioned (R01000 for 2D edge length ~ 1000 m) (R07000 for 2D edge length ~7000 m).

Regards,
Audrey

The use of OV in the vent exemple (in user_t3d_corfon.f) is just a method to define the bottom level at -10m. this is because the bottom in the geometry file is just a mesh with no bottom defined...

the tar file is not downloadable so I definitly don't understand what you're trying to do...

Alright. Please excuse me for not being clear.

I wish to build a 3D domain based on the attached case.

On my side, the computation achieved well if number of levels = 2 but do not achieve if number of levels = 3.

Is it well possible to run this computation with number of levels greater or equal 3 ?

Kind regards,
Audrey

Hi Audrey
In your geometry file, you have bottom values between -66 and 5.
Your initial condition is a zero elevation which means some nodes are dry...
Then with the sigma mesh transformation, it seems there is a problem during the computation of elevation of mesh layer...
Just checked it works if the initial elevation is chosen in a way there is no dry nodes...
You probably should work on the mesh transformation...

Alright Christian,
Thank you very very much for this tip !
I'll try it and let you know if it's ok.
Thanks again

Hi Audrey,

As Christian pointed out the bottom has values between -66 and 5 m. So, you need to use dry/wet mechanism by allowing

TIDAL FLATS : YES

I run your model with small modifications with 30 sigma-layers for one day and it looks OK.

Best,
Alexander

Here is the log file