Dear Jean-Paul,

Thank you again for a detailed explanation.

However, please excuse my ignorance... 

>
> 1. Create a vector-valued finite element system (maybe an 
> FESystem<dim>(FE_Q<dim>(1),dim) ) and use it to distribute DoFs onto a new 
> DoFHandler.
> 2. Create a mass matrix "M" with the new DoFHandler.
> 3. Create a RHS vector "f" of the form f_i = \int_\Omega f(x) \phi_i(x) dx 
> . Here f(x) = \grad \psi (x) is the gradient of the solution evaluated at 
> point x and \phi_i is a vector-valued shape function. So you need to 
> obviously compute the gradient of the solution at all quadrature points 
> when doing integration - you would do this with the function I mentioned 
> previously.
>
> What makes me confused here is that, according to your explanation, two 
different DoFHandler could be used simultaneously.

I mean let's say dof_handler1 is for solution and dof_handler2 is for M

To get \grad\psi, I have to use dof_handler1, however, to assemble matrix M 
and vector f I have to use dof_handler2...

So, in the assemble_system() I don't know whether I have to use 

cell= dof_handler1.begin_active(), endc = dof_handler1.end();
or 
cell= dof_handler2.begin_active(), endc = dof_handler2.end()

Also, It seems that I have to use two different FEValues<dim> one for 
solution the other for the vector...

In the assemble_system() for M and f, can I use these two things 
simultaneously...?

Thank you

Kyusik.

 

-- 
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 dealii+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to