Hello everyone,
I am working on a multi-physics problem (fluid and solid domains) using an
hp::FECollection similar to the layout in step-46. My collection is defined
as:
fluid_fe(FE_Q<dim>(degree), FE_RaviartThomasNodal<dim>(degree),
FE_DGQ<dim>(degree)),
wall_fe(FE_Nothing<dim>(), FE_RaviartThomasNodal<dim>(degree),
FE_DGQ<dim>(degree)),
I want to constrain the first finite element (FE_Q, component 0) to zero
along the fluid-solid interface. Because my system contains a non-primitive
vector element (FE_RaviartThomasNodal), standard primitive functions like
face_system_to_component_index cause assertion crashes if called directly
on all face DoFs.I have thought of two potential strategies to isolate and
constrain only the primitive FE_Q dofs on the interface face.
First approach is to use is_primitive to safely filter out the
non-primitive Raviart-Thomas face shape functions before querying the
components:
// ... [Interface detection loops from step-46]
... if(face_is_on_solid_boundary)
{
// v = 0 on boundary condition
cell->face(face_no)->get_dof_indices(local_face_dof_indices,0);
for(unsigned int i = 0; i < local_face_dof_indices.size();++i)
{
if(fluid_fe.is_primitive(fluid_fe.face_to_cell_index(i, face_no)))
{
if(fluid_fe.face_system_to_component_index(i).first == 0)
constraints.constrain_dof_to_zero(local_face_dof_indices[i]);
}
}
}
Second approach is to utilize system_to_block_index() to target the first
block directly, which circumvents checking component indices:
// ... [Interface detection loops from step-46]
...if(face_is_on_solid_boundary)
{
// v = 0 on boundary condition
cell->face(face_no)->get_dof_indices(local_face_dof_indices,0);
for(unsigned int i = 0; i < local_face_dof_indices.size();++i)
{
if(fluid_fe.system_to_block_index(fluid_fe.face_to_cell_index(i,
face_no) ).first == 0)
constraints.constrain_dof_to_zero(local_face_dof_indices[i]);
}
}
My questions are
1. Which of these two approaches is considered the correct or safest way in
deal.II when handling a mix of primitive and non-primitive elements in an
hp context?
2. I am also aware of face_system_to_base_index(i). How does this function
differ from the 2 functions used above in this context and would it be a
better alternative?
Any advice or feedback would be greatly appreciated!
Regards,
Devansh Sonigra.
--
The deal.II project is located at http://www.dealii.org/
For mailing list/forum options, see
https://groups.google.com/d/forum/dealii?hl=en
---
You received this message because you are subscribed to the Google Groups
"deal.II User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion visit
https://groups.google.com/d/msgid/dealii/CAGHTkKvEG%3Dk-DJ970RKELBTWn3U4AsP-Oj9EE-W3JVGpP4iQUg%40mail.gmail.com.
