Gauss point contour plotting of 10-node tetrahedra meshes

Moderator: GiD Team

Post Reply
skatulla
Posts: 5
Joined: Sun May 29, 2016 9:58 pm

Gauss point contour plotting of 10-node tetrahedra meshes

Post by skatulla »

Hello

I let my structural analysis software write out scalar, vector and matrix-valued results at the Gauss points. This works fine for hexahedral meshes and linear tetrahedral meshes. However, when using quadratic tetrahedral meshes (10 nodes) with a 3rd-order Gauss quadrature (5 Gauss points), the visualization at the Gauss points fails. GiD shows the correct values color-coded contour legend (as found in the RES file) but the contour plot of the actual structure is completely black. Also, the "smoothed" option in the contour plot menu is missing.

The header section in my RES files looks as follows:

GiD Post Results File 1.0
GaussPoints "Volume Gauss points" ElemType Tetrahedra
Number Of Gauss Points: 5
Natural Coordinates: given
2.500000000000e-01 2.500000000000e-01 2.500000000000e-01 2.500000000000e-01
5.000000000000e-01 1.666666666667e-01 1.666666666667e-01 1.666666666667e-01
1.666666666667e-01 5.000000000000e-01 1.666666666667e-01 1.666666666667e-01
1.666666666667e-01 1.666666666667e-01 5.000000000000e-01 1.666666666667e-01
1.666666666667e-01 1.666666666667e-01 1.666666666667e-01 5.000000000000e-01
end gausspoints

Please advise.

Regards,
Sebastian
User avatar
escolano
Posts: 1918
Joined: Sun Sep 05, 1982 10:51 pm

Re: Gauss point contour plotting of 10-node tetrahedra meshes

Post by escolano »

If you use tetrahedra of 10 nodes and gauss results on your user define 5 gauss points GiD is unable to extrapolate the result from the gauss points to the 10 nodes and then is unable to show a continous contour fill. In these cases GiD will represent the result as spheres on each gauss point coloured with the result color.

I recommend you to write the result for postprocess on the GiD internal 10 gauss points (althougth you use for your internal calculation 5 integration gauss points), or if the result is a continous property I recommend to write a continuous 'nodal result' instad of discontinous 'gaussian results'
(you should somehow average the result of the elements on the shared node)
skatulla
Posts: 5
Joined: Sun May 29, 2016 9:58 pm

Re: Gauss point contour plotting of 10-node tetrahedra meshes

Post by skatulla »

Ok, thanks.

Sorry for not responding earlier. I thought any email responses would also be sent to me via email.

I have meanwhile interpolated the Gauss point results at the particles using the Moving Least Square method. So, it is all fine now.

Regards,
Sebastian
Post Reply