open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,
I wonder if there is a pre-proc like matisse/blueKenue for Unix?
Because I have some problems to generate a mesh with matisse.
It seems that there are too many points (near 250 000 points). I tried your tricks, ie. for a XYZ-file, I put "Z=0" for every points. But it still crash, as you can see on this picture:

picasaweb.google.com/1041353125224904960...#5642461977174877154

Then I tried with BlueKenue, but the problem is that I can't specify the "size" of an element (refinement near the coast and larger in the ocean).

So I would like to try matisse on a Linux/Mac computer.

Thanks

Hi Kroptia

Long time ago, Matisse only exists on UNIX but I think none of the users still have a suchold version...
Blue Kenue could run on Linux using Wine.

About your problem

I think the Matisse crash doesn't come from the number of points (250 000 points without bathymetry information id OK for Matisse). In my opinion, I suppose you had a problem with your datas (double points or bad criterion on the boundary which generate a crossing).
Try to have a look into the emse_Defectivertices.txt to see the information.
You could also try with a smaller criterion to check if the crash still exist.

With Blue Kenue, you could specify the size of elements using the density layer of the mesh generator. Look into the manual you will see a small example of this capacity

Hoep this helps

Thank you for your (fast) answer. I'll try to look into my datas.
I'll keep in touch.

Hello,
I fixed all points, but it still get problem, as you can see on the following picture:

picasaweb.google.com/1041353125224904960...#5644140277136520418

Hi
hard to say exactly where the problem is...
At this step, I'm thinking there was a problem on your outline (or maybe on a constraint line).
It's often link to the mesh density if the value is too large in comparison with the outline density point. Sometimes this difference could generate a cross between 2 segments of the outline and the matisse crash.

Maybe you could zip your .MAT directory and join it.

Hope this helps

Hi,
Here you can find my .MAT directory zipped
grille1.MAT.zip
Thank you for your help

I can't join the file, so I upload it on dropbox, here the link:
dl.dropbox.com/u/32253414/grille1.MAT.zip

Hi
As I said previously in this topic, your problem comes from the density mesh you choose!!!
You had an island with a lot of small details (define by hard nodes). The size is about 10m.
Your smaller criterion is 4 km !!!
It's totally unrealistic to try to compute a mesh with a criterion of 4km with some constraint nodes only separated by 10m

Change your density or the definition of your island and it will works

Regards

Ok, thank you for your help.
I'll try to fix it.

we have just released a free limited version of our preprocessor janet
on www.smileconsult.de.
There is also a commercial version available for unix which allows to generate meshes with a million nodes or even more (depending on amount of memory available).

Best regards
Frank