# Re: Bug in SMESH?

Hi,

I am still working on the twisted beam I was speaking about in previous posts.

I could finally create the geometry, and mesh it with 6 x 1 x 1 solid elements.

Successful? Not really sure...

I would like to change the mesh density, but this leads to impossibility to re-compute the mesh. For instance:

- modifying hypothesis "Nb. Segments_length" from 6 to any other value (let's consider 12) make it impossible to get solid elements again

- even re-modifying hypothesis "Nb. Segments_length"from 12 to 6 returns impossibility to get solid elements again

I succeeded once to mesh this geometry, but modifying hypothesis values to change mesh density seems now systematically to lead to impossibility to create a correct mesh...

Is it a bug or is it related to the way I created the twisted beam?I thank you in advance for your help.

Bests,

Pierre

Ps: using Salome 6.5.0

Hi again,

Well, I do think there is somewhere a bug because I don't think it is normal to be unable to get back to a mesh that was correctly computed once, however this seems related to the 3D algorithm consideed: "extrusion 3D".

Trying with "3D: Automatic hexahedralization" works like a charm!

Have a good evening.

Bests,

Pierre

Hi Pierre

There is certainly a bug. A right behavior is that at re-computing after hyp modification there is a warning saying that the re-compute may fail + proposition to make a full re-compute instead of partial one.

Problem is that when one modifies a local hypothesis, only a part of mesh is cleared and then re-computed. In your case a remaining part is quadratic and newly computed part is linear that leads to failure of 'Extrusion 3D'.

I observe a right behavior if I modify "Number of segments" just after conversing a mesh to quadratic but not after loading a study with a ready mesh. So the bug is somewhere in persistence.

Thank your for your post.

Best regards

Edward