On Tue, May 5, 2015 at 4:08 PM, David Knezevic <david.kneze...@akselos.com>
wrote:
> On Tue, May 5, 2015 at 5:48 PM, John Peterson <jwpeter...@gmail.com>
> wrote:
>
>>
>>
>> On Tue, May 5, 2015 at 2:04 PM, David Knezevic <
>> david.kneze...@akselos.com> wrote:
>>
>>>
>>> Also, going back to my negative jacobian issue: I realized that this
>>> element is actually not affine. So it seems like John's comment could be
>>> right, i.e. that the jacobian is negative at at least one of the quadrature
>>> points, and that Cubit doesn't pick this up because it's not testing the
>>> jacobian at the same points.
>>>
>>
>> Try initializing a CONSTANT order quadrature rule in your test code and
>> see if that runs without hitting a negative Jacobian?
>>
>
> That still gives a negative Jacobian.
>
> I can't imagine there's anything wrong with the libMesh code for computing
> the jacobian, since that's used everywhere, so I guess this element must be
> invalid. I pasted the node locations earlier in this thread, and I was
> wondering if there is any "sanity check" I can run on this element to
> confirm that it is indeed bad?
>
I plotted your element, and it looks pretty bad, almost completely flat.
In the first image, you can see that the apex node and the mid-face node on
the base are very close together. In the second image, you can see a
reference Tet10, just to prove that Paraview displays these correctly. I
really don't know how Cubit can claim that element has a non-negative
Jacobian...
https://drive.google.com/file/d/0B9BK7pg8se_iVUVpSUs4Y3J2dFk/view?usp=sharing
https://drive.google.com/file/d/0B9BK7pg8se_ibUFKOHRDZ1FrTEE/view?usp=sharing
Test code: https://gist.github.com/jwpeterson/c6c2fc8f627c02a0546b
--
John
------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_______________________________________________
Libmesh-devel mailing list
Libmesh-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/libmesh-devel
