-
The first and simplest thing we want to support is 1st order lagrangian finite elments for arbitrary dimensions.
The Simplicial Lagrangian FE spaces in 2D from NumPDE should easily generalize to high…
-
Besides affine transformations of reference elements, we should consider also supporting more transformations, like quadratic ones.
This would require a more sophisticated geometry layer, that makes …
-
So much boiler plate here... We really have to refactor the tests somehow. I.e. it would be nice to test the FEs with some dummy Grid and dummy Basis. Here so much construction happens.
_Originally…
-
# Missed elements
While generating finite elements, some are occasionally missed. It causes unnecessary empty spaces inside the finite element models.
## Disconnected parts
Those missed fini…
-
In order to avoid locking problems for incompressible models (such as the rubber models implemented), the weak form needs to be adjusted. This can be done through the Q1P0 family of elements, where t…
-
After searching for shape functions for hexahedrals to add in MSolve i run into [CalculiX](http://www.dhondt.de/) . This code contains a lot of [elements types](http://web.mit.edu/calculix_v2.7/Calc…
-
This could be a great way to also share learning resources for the finite element method to get started on the basics.
-
- [ ] Example: the number of integration points can be specified by the user
- [x] #252
-
With the removal of finite elements and dofmaps from UFCx in #681, there isn't much left apart from the kernel interface. Maybe we can removed `ufcx.h` and simply document the interface?
-
First, thanks for the great work behind this library! I was giving a try swapping an `fs2-grpc`-based client with this library, but I encountered one major problem:
We are unable to retrieve the res…