Hello Christian,
yes, that should be the case. But up to now one can put into the
hp::FECollection whatever one likes to, e.g. one FESystem (FE_Q, 2,
FE_Q, 1) and one FE_Q, and activate them on different cells. Then there
cannot come out anything useful from DoFTools::count_dofs_per_block.
This should be prevented by some checks, if one wants to add this
functionality.
Best Regards,
Markus
Am 16.03.11 09:00, schrieb Christian Goll:
Hi Markus,
Quoting Markus Bürg <[email protected]>:
Hello Christian,
you are right, unfortunately this concept is not easy to implement for
the hp-context, because for example it might be the case that there are
different-sized FESystems in the hp::FECollection. Anyway this does not
mean that it is impossible. One just has to take much more care than in
the non-hp-context. If you want to work on this, I would be glad to
help you with the implementation.
Best Regards,
Markus
as I intend to 'abuse' the hp-concept to allow for a discontinuity in
my solution on an interface in the interior of my domain (see problem
below), so I will give it some thought and come back to your offer.
Just one question: What do you mean by "different-sized FESystems"?
The number of components must be the same for every FESystem in a
FECollection, or have I misunderstood you?
best regards,
christian
Am 15.03.11 17:53, schrieb Christian Goll:
Hi Markus,
I am particularly interested in DoFTools::count_dofs_per_block,
which does not exist for a hp::DoFHandler. Further investigation in
this direction brought me to the assumption, that the whole
concept of grouping components into a block is missing in the
hp-context, i.e. in the non-hp-case, the number of blocks is stored
in FiniteElementData, which is a base class of FiniteElement, but
I did not find something similiar for the FECollection.
best regards
christian
_______________________________________________
dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
Quoting Thomas Carraro <[email protected]>
I have two domains (Omega_1 and Omega_2), in both of them I have only
one >variable (u_1 resp. u_2), that has a discontinuity on the
interface between >the two domains, i.e.
in "Omega_1" we have
"u_1 != 0" and "u_2 = 0"
in "Omega_2" we have
"u_1 = 0" and "u_2 != 0"
so I actually need double degrees of freedom at the interface.
The idea was to use a hp_collector like
fe_1 ( FE_Q<dim>(1), 1, FE_Nothing<dim>(), 1 ),
fe_2 ( FE_Nothing<dim>(), 1, FE_Q<dim>(1), 1 ),
fe_collection.push_back (fe_1);
fe_collection.push_back (fe_2);
in this way we would have double degrees of freedom only on the
interface
_______________________________________________
dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
