Anisotropic surface mesh


I am trying to use MMGS to perform anisotropic remeshing. The mesh & metric I use are in the attached files
The solution is quite singular, so the computation of a clean metric is challenging.

Here is a view of what I (try to?) impose:

On the left are the directions of the metric field, scaled by 1/sqrt of the eigenvalue.
On the right is the adapted mesh. Is it normal that I get such small triangles close to the free edge?

Best regards,

aniso.meshb (134.0 KB) aniso.solb (112.5 KB)

Hello Xavier,

I have an issue with your metric file: I can’t vizualize it with Medit and Mmg complains about the tensor values at point 2:

point 2 -> eigenvalues : 0.000000e+00 2.455882e+225 -2.455882e+225).
           metric tensor: 7.098843e+29 -1.438243e-47 -2.455882e+225 -2.855411e-307 1.427582e-39 -1.096746e-282.

I am not sure but I think that you have small sizes near the sharp angle (the very little red arrows). If yes, it is probable that the Mmg gradation propagates these sizes up to the sharp angle and you end up with small elements.

You can confirm that by disabling the gradation using the -hgrad -1 option.
Note that your resulting mesh may have a poor quality…

If it works, you can try to relax the gradation value (by default it is of 1.3 but you can try values like 1.6 or 1.8).

I hope that it will help.