[Gmsh] Local referential for the hexahedron and prism element family
Geordie McBain
gdmcbain at freeshell.org
Tue Feb 16 05:26:56 CET 2010
On Mon, Feb 15, 2010 at 11:57 PM, Rui Maciel <rui.maciel at gmail.com> wrote:
> This isn't a node ordering problem per se. My only issue is that according to
> Gmsh's manual it appears there is an inconsistency regarding the local
> coordinate system for the prism family of elements. In fact it appears that,
> at least on the ASCII art representation presented in the manual, the
> tetrahedron and hexahedron element family follow a "right-hand" coordinate
> system while the prism and pyramid follow a "left-hand" one.
Yes, I see what you mean. The ASCII representations are a little
ambiguous. For the tet and hex I had thought the u-axis was behind
and to the right of the vw-plane, while for the prism the u- and
v-axes were forward and for the pyramid the u-axis was forward with
the vw-plane in the plane of the page. With those interpretations,
all four systems are right-handed, but I think there are other valid
interpretations; e.g. for the prism the u- and v-axes could be into
the page which would give a left-handed system.
Perhaps the figures would be clearer if where one edge crossed
another, the closer edge symbol was used, rather than using a
plus-symbol?
(Hoping these are rendered in monospace.)
Prism (base forward):
w
^
|
3
,/|`\
,/ | `\
,/ | `\
4-------------5
| | |
| ,/|`\ |
| ,/ | `\ |
|,/ | `\|
,| | |\
,/ | 0 | `\
u | ,/ `\ | v
| ,/ `\ |
|,/ `\|
1-------------2
Prism (ridge forward):
w
^
|
3
,/|`\
,/ | `\
,/ | `\
4------|------5
| | |
| ,/|`\ |
| ,/ | `\ |
|,/ | `\|
,/ | `\
,/ | 0 | `\
u | ,/ `\ | v
| ,/ `\ |
|,/ `\|
1-------------2