You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Consider a 3x1 unit square mesh with quadrilateral, where the cells are marked 1,0,2, from left to right.
Assume that we have a boundary condition on the left and right hand side of the unit square.
Assembling each of the domains 1 and 2 separately, and applying the boundary condition can only reproduce the expected behavior if there is only one cell integral present. See the minimal example below:
Here, we would expect that the L array, which used the bc for both domains and the matrix for both domains, should reproduce the sum of each of the integrals, L1 and L2.
This is not the case:
garth-wells
changed the title
Lifting for problems with domain markers has unclear behavior
Boundary condition lifting does not respect domain makers
Oct 20, 2020
Lifting for problems with subdomains have unclear behavior, as it only uses the kernel for the first integral:
https://github.com/FEniCS/dolfinx/blob/master/cpp/dolfinx/fem/assemble_vector_impl.h#L174
Consider a 3x1 unit square mesh with quadrilateral, where the cells are marked 1,0,2, from left to right.
Assume that we have a boundary condition on the left and right hand side of the unit square.
Assembling each of the domains 1 and 2 separately, and applying the boundary condition can only reproduce the expected behavior if there is only one cell integral present. See the minimal example below:
Here, we would expect that the
L
array, which used the bc for both domains and the matrix for both domains, should reproduce the sum of each of the integrals,L1
andL2
.This is not the case:
The text was updated successfully, but these errors were encountered: