ignore lower dimensional elements in postTimeStep
Sometimes it may be useful to leave lower-dimensional submeshes in the domain mesh, e.g. for postprocessing. Running a simulation on such a mixed-dimensional mesh generally works (primary variables are correct), however there seems to be inconsistencies with derived variables in the postTimeStep computations.
In an example of the small deformation process on a cube with a meshed plane, representing a geological fault, the following has been observed. When the boundary conditions are set to the trivial case of 1D-compression, the displacements match the analytical solutions, however the stresses (nodal values) near the fault plane are lower than the analytical solution of constant vertical stress elsewhere. So the assumption is that the lower-dimensional elements are still active in the postTimeStep and their zero-values enter the secondary variables.
If this evaluation is by purpose, than it would be nice to have a flag in the prj-file to ignore all lower dimensional elements in the domain.