How to assemble (bi)linear form using refined mesh?


164
views
0
5 months ago by
Dear all,

I am trying to improve the way I assemble an expression-based form over a mesh. For now I am using high order quadrature. Would it be possible to refine the mesh and integrate with low-order quadrature on the refined mesh, while keeping the functional space of the coarse mesh? Thanks for any suggestion!

Martin
Community: FEniCS Project
From the blurry description I understood you want to integrate over a mesh finer than the one the test (and trial) function lives on. This is not currently possible for general situations. It is nevertheless possible to hack this in for special cases like DP0. You can use a DP0 test function on fine mesh, assemble and take care of reductions to coarse mesh by yourself, see for example https://github.com/blechta/paper-norms-nonlin-code/blob/91369a45e771305790543b31156a57764026c0b5/main.py#L224.
written 5 months ago by Jan Blechta  
Thanks Jan! And sorry for my blurriness. I am realizing that in some cases what I try to do can be done by simply defining a custom quadrature rule, i.e., a custom quadrature finite element. However, I cannot find where it can be done in FEniCS. Can you give me some pointers? Thanks again! Martin
written 4 months ago by Martin Genet  
https://bitbucket.org/fenics-project/dolfin/issues/955/support-for-arbitrary-quadrature-rules#comment-41836403 describes a hack how to arbitrary (compile time) quadrature.

There is some facility to do custom runtime quadrature. It is employed by the multimesh functionality in DOLFIN. But I am not sure how it is customizable and usable from user (not multimesh) codes. Keyword is "custom" integral in FFC.
written 4 months ago by Jan Blechta  
Please login to add an answer/comment or follow this question.

Similar posts:
Search »