Forum Message

 

 

We have moved the forum to https://forum.ngsolve.org . This is an archived version of the topics until 05/05/23. All the topics were moved to the new forum and conversations can be continued there. This forum is just kept as legacy to not invalidate old links. If you want to continue a conversation just look for the topic in the new forum.

Notice

The forum is in read only mode.

second-order derivatives for vector finite elements

More
4 years 2 months ago #2422 by Guosheng Fu
Hello,

I am not sure how to get the second-order derivative information for the edge elements.
For H1 space, we can do v.Operator("hesse"), but it seems that this operator is not implemented for vector finite elements yet... Do you have any suggestions?

Best,
Guosheng
More
4 years 2 months ago #2423 by christopher
Hi Guosheng,
Do you need it for postprocessing? You could project the function on a VectorH1 of higher order and then take the gradient. If you need it in the form, I think you would need to implement and export it.

Best
Christopher
The following user(s) said Thank You: tommaso
More
4 years 2 months ago #2425 by Guosheng Fu
Christopher,

I actually need it for the form. It's a forth-order vector PDE.
I am a bit scared to dig into the c++ code myself. But maybe it worth a try.
As a first step, I just want to work on square/box elements, which might also makes the coding less intensive. Can you suggest where to start with?

Best,
Guosheng
More
4 years 1 month ago #2450 by mneunteufel
Hi Guosheng,

have you tried a mixed formulation to reduce the fourth-order problem to two second-order problems like the Hellan-Herrmann-Johnson method in this tutorial ?

Then you maybe could avoid the Hessian operator.

Best
Michael
More
4 years 1 month ago - 4 years 1 month ago #2499 by bl736
Hello Michael,

I am following the demo: The Hellan-Herrmann-Johnson Mixed Method . I would like to know the explicit basis functions for the bending moments, could you point out some accessible references on this for an engineering background user. Many thanks!!

Sincerely,

Bin
Last edit: 4 years 1 month ago by bl736.
More
4 years 1 month ago #2502 by joachim
Hi Bin,
have a look into the PhD thesis by Astrid Pechstein (aka Astrid Sinwel). Chapter 4.4, in particular page 79.
TDNNS and HHJ use the same finite element spaces for the stress (or moments).
available from here:
www.numa.uni-linz.ac.at/Teaching/PhD/Finished/sinwel-diss.pdf
best, Joachim
Time to create page: 0.112 seconds