open TELEMAC-MASCARET The mathematically superior suite of solvers

When I run the model, it shows: NEGATIVE DETERMINANT
I have searched the same error of this issue in forum, the answers were the mesh was not good for calculation, or used channel meshs. Currently my mesh is no such problem, but I still get this kind of error.

Many thanks!!

Hello,

This is indeed a problem in your mesh. We expect that the triangles in a mesh are oriented (their 3 points anti-clockwise), so that computing (x2-x1)*(y3-y1)-(x3-x1)*(y2-y1) gives a positive number which is twice the area of the triangle. You cannot see this on a screen. What mesh generator did you use ? Another possibility is that a triangle has 3 points at the same point, thus having an area = 0.

With best regards,

Jean-Michel Hervouet

Thanks for this information.

I used Blue Kenue to generate the mesh.
How can I check if I have a 0 area triangle ?

Hello,

I presently work on a project related to river modelling with Telemac 2D (version v8p2).
I had to densify an existing model in order to integrate some piles of a temporary bridge.
I get a similar error (see enclosed log file) with negative determinant.
I do not understand the origin of the error as I corrected the meshes with Blue Kenue (make all elements CCW).
I also enclosed the set of modelling files.

Thanks in advance for your feedback.

Regards.

Hi
Try in sequential rather than in parallel
You could also use checking the mesh which could give you more information

For info *.log file couldn't be opened in the forum --> change to txt

Regards

Thanks for your answer.

I tried to run the .cas with 1 set to the number of parallel processors and here is the error which is reported:
GEOELT: ELEMENT 34589 : NEGATIVE DETERMINANT
application called MPI_Abort(MPI_COMM_WORLD, 2) - process 0
I checked the mesh, in particular this element, and did not notice any problem at this element.

Would you have any idea?
I don not have logfile anymore as previously but you will find enclosed a screeshot of the message in cmd dialog box.

Regards.

You should check element 34589
BK could give also some mesh informations
Look at the geometry file, not the mesh in BK format (rounding of coordinates could lead to such errors)

I checked this element and everything seems OK

Here the element at the basis of the problem.

The mesh analysis done in BK seems to be OK also as there are no over constrained nodes, no CW elements neither elements with more than 10 adjacent elements.
I really do not understand where it can come from.