open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,
I imported a *.2dm mesh in BlueKenue which worked satisfying when using SMS / Hydro_AS-2D. When I display the mesh in the 2D-View the mesh is illutrated different than in SMS, especially around a pier. You can see the differences in the attached pictures (Attachment 1 and 2). The geometry of the pier and the mesh is completely different and not as expected or desired.
First I thought it is an display issue and ignored it.
But when I wanted to start a simulation with Telemac-2D the simulation can't start because there are Elements with negative Determinants.
When I searched for the affected Element it was an Element where all three points seem to be on one line(see Attachment 3, Screenshot made from FudaaPrepro, where the mesh looks the same like in BlueKenue). For comparison i added a screenshot of the affected element in SMS (See Attachment 4).
Now I'm not sure if this is still a display issue or if something went wrong during the Import.
I shifted my mesh next to the origin, because I thought maybe it would help, but there are still the same problems in displaying the mesh and the aborted simulation.
So now my questions are:
1) Is it really just a display issue or is the mesh defective?
2) What can I do to avoid negative determinants and how can I fix it when the mesh already exists?
3) Is there anything I need to keep in mnd when Import a *.2dm-mesh-file to avoid errors like that?

I would appreciate any kind of help.

Cheers
Sarah

PS: I had the same display problems and problems with negative determinants when I created my mesh completely new with BlueKenue.

Hi Sarah,

This is a problem of precision. (this has been discussed in other threads)
Floating point is constrained to 7 significant digits and you have a very fine mesh. If you shift your data to be closer to (0,0) all should be good.

I'm guessing it's about time I start supporting double precision SELAFIN files.

However... Please post your 2dm file here so I can investigate.

Cheers... Martin

Okay,
I found the hint to shift the mesh near to the origin in other threads. That's what I did in BlueKenue, but I still have the same problems.
In the Attachment I added the *.2dm-file with the original coordinates.
Do you think the mesh is too fine around the piers?

Cheers Sarah

Hi Sarah,

Yes, it is a resolution problem.
You have 10cm element edge lengths around the piers.

Try this.

1. Import the 2dm file into Blue Kenue.
2. Using "Edit->Shift in X/Y" shift the mesh (-711000, -9612000)
3. Now display your mesh and zoom in to the piers.
4. It should look as expected.

I don't know what you are trying to model however
10cm elements in a 5km domain seems a bit excessive.
Your timestep will certainly have to be very small.

Hope this helps... Martin

Hi Martin,
Of course I have to translate the mesh near to the origin before I import in BlueKenue. I did it afterwards in BlueKenue.
Now it looks fine and the simulation runs!

Thank you very much for your help!

Cheers
Sarah