[wikstrom]

Hi,

I think there is a serious bug in extrudeMesh. Using the linearNormal model, all cells created become inside out. Extruding with negative thickness produces a valid mesh. (that ofcourse can be moved into the desired place with transformPoints, but this workaraound feels a bit awkward.)

Attacheed is a test case describing this. Run the included run.sh with argument negative and positive.

Cheers
Niklas

[mattijs]

Hi Niklas,

the normal is indeed reversed. Agree it is probably not logical. I'd rather not change the behaviour in 1.5.x but will in future versions.

Thanks,

Mattijs

[sbence]

Dear Foamers,

I am trying to create a 2D mesh to be used in OpenFOAM with snappy. My method is the following:

1. create a mesh in snappy, thickness should be large enough to have a few nice layers in the middle
2. create a cellSet containing these nice layers in the middle
3. create new mesh based on the cellSet with subsetMesh; the new mesh has almost planar patches on the sides
4. autoPatch and select a face for extrusion
5. extrudMesh using the right patch

My problem is that the generated mesh contains hexa cells and pyramids as well. So it is not a real extrusion, because the extrusion would result in hexas and wedges only.

Some explanatory screenshots:
Result with snappy:

cellSet from the middle:

new mesh from cellSet:

patch to extrude (why the triangles are not visualized in paraview?):

result mesh with extrudeMesh:

the same visualized in engrid. different colors > different cell types:

pyramids visualized in engrid:

tetras visualized in engrid:


How can I avoid pyramids/tetras?

Thanks for any help!

Bence