[GiDlist] quadratic Prism elements

[Jose Antonio Fernandez]

Thank you!

-----Mensaje original-----
De: gidlist-admin at gatxan.cimne.upc.edu
[mailto:gidlist-admin at gatxan.cimne.upc.edu]En nombre de Miguel Pasenau
Enviado el: jueves, 03 de noviembre de 2005 16:43
Para: gidlist at gatxan.cimne.upc.edu
Asunto: RE: [GiDlist] quadratic Prism elements



Hi,

Well at the beginning there was no support for quadrilateral prismas, but
more and more we’ve added support for them, and the same happened for the 6
gauss points on those prismas.

The bugs have been corrected and will be available on the next beta version.

Note that when giving the results only for 6 gp on a 15node prismas, gid
will extrapolate the result to the 6 vertex of the prisma, and then a linear
interpolation will be done to get the results for the middle (quadratic)
nodes.

regards



miguel




_____


De: gidlist-admin at gatxan.cimne.upc.edu
[mailto:gidlist-admin at gatxan.cimne.upc.edu] En nombre de Enrique Escolano
Enviado el: miércoles, 02 de noviembre de 2005 21:16
Para: gidlist at gatxan.cimne.upc.es
Asunto: Re: [GiDlist] quadratic Prism elements



Hello,



It seems that exists some problem when mixing different gauss point
definitions.



About your sample, the Given Natural Coordinates are wrong (they define 6
gauss points, but are provided 12 coordinates)



In other hand, some features are currently missing for prisms.



- The "internal" gauss points was not documented in the help (we added now
for the next beta version)



For linear prisms, the 6 gauss points GiD internal normalized coordinates
are:

( a, a, b)

( b, a, c)

( a, b, c)
( a, a, d)

( b, a, d)

( a, b, d)

a=0.166666666666666, b=0.666666666666666
c=0.211324865405187, d=0.788675134594812

- The quadratic prism, with 15 gauss points is not implemented in
postprocess, it will added in the related next beta.



Note: for a given number of gauss points different of the number of nodes
(for example 6 gauss points with quadratic 15-noded prism), the results
cannot be extrapolated to nodes (they are less equations than unknown
values), and the results will be shown as balls on gauss points, instead
contour filled.

Maybe it's better to write, for postprocess pourposes, an alternative linear
mesh (6-noded prisms).



Thanks for your example, it's very interesting to detect and correct this
bug.



Regards

Enrique Escolano



----- Original Message -----

From: "Jose Antonio Fernandez" Jose.A.Fernandez at cedex.es

To: gidlist at gatxan.cimne.upc.es

Sent: Friday, October 28, 2005 3:42 PM

Subject: [GiDlist] quadratic Prism elements



Hi,

Are you sure that results at gauss points is working properly for
quadratic
Prism elements?

Linear Prism with only 1gauss point is ok but I think that quadratic Prism
with 6 gauss points doesn t work for "Natural Coordinates: Internal"
neither
for "Natural Coordinates: Given".
I send you a simple example.
Thank you




-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://listas.cimne.upc.edu/pipermail/gidlist/attachments/20051103/2afc6720/attachment.htm