Hi all,
as seen in the figures below, some of the hexes generated by
reset
brick x 9 y 9 z 9
move vol 1 x 4.5 y 4.5 z 4.5
webcut volume 1 with plane zplane offset 3
webcut volume 1 with plane zplane offset 6
imprint all
merge all
volume all size 3
mesh volume all
refine vol 3 using plane zplane
have more than 6 faces
.
Can you please explain why they are still recognized as hex?
Kind regards,
-Christos
Hi Christos,
the hexes you are showing are still hexes with 6 faces. That is just a bad visualization from the graphics kernel.
If you view the hex from a different angle its displayed as it should be.
Thank you, Norberbt.
I also thought that it is a visualization artifact, but then I checked if the points on the suspicious faces are coplaner, and it turns out that this is not the case.
Consider, for example, the nodes of the two surfaces shown below:
Using the online tool Coplanar points node 371 is located at a distance of 0.32 from the plane 1.5x + 1.5y - z = 0 defined by nodes 119, 362 and 369.
Similarly, node 408 lies at a distance of 0.38 from the plane 1.5x + 0.75y + z - 15.75 = 0 defined by nodes 126, 395, 406.
While in some cases faces on linear hexahedral elements might happen to be planar, this is not generally the case. This can be seen by looking at the basis for a linear quadrilateral elements:
This refinement scheme generates hex elements on the interior with non-planar faces. What you’re seeing are the render facets, which are two triangles that subdivide the quadrilateral face, using a “flat” shader.
Here’s an example of this on a linear hex mesh in ParaView using flat shading:
and the same mesh using “Gouraud” shading:
