[nilesh]

Hi,
I am creating a 3D mesh using hexahedral elements. In some regions, the hexahedrals are supposed to actually convert into wedges due to overlapping of their sides and some rectangles will consequently get converted to triangles at the boundary surfaces (imagine meshing a cylinder with hexahedrals).
Instead of defining wedge elements seperately, if I declare these elements as hexahedrals with two pairs of distinct nodes having the same coordintes so that geometrically it is actually a wedge, will it cause any error in flow computation?
Similar is the case of defining a rectangle with one repeated coordinate.

Thanks.

[fpalacios]

Originally Posted by nilesh

Hi,
I am creating a 3D mesh using hexahedral elements. In some regions, the hexahedrals are supposed to actually convert into wedges due to overlapping of their sides and some rectangles will consequently get converted to triangles at the boundary surfaces (imagine meshing a cylinder with hexahedrals).
Instead of defining wedge elements seperately, if I declare these elements as hexahedrals with two pairs of distinct nodes having the same coordintes so that geometrically it is actually a wedge, will it cause any error in flow computation?
Similar is the case of defining a rectangle with one repeated coordinate.

Thanks.

Hi,

In principle, it should work. In the past we have found some issues with the gradient computation… our recommendation is to use Green Gauss.

Cheers,
Francisco

[nilesh]

Originally Posted by fpalacios

Hi,

In principle, it should work. In the past we have found some issues with the gradient computation… our recommendation is to use Green Gauss.

Cheers,
Francisco

Thank you Sir. Can you please point to a source/paper which talks about the effect of cell degeneration in finite volume computation?