Hello Alexander,
then you are right and inserting the dot product of orientation vector
and shape function should work. Where do you get the orientation vector
from? Should this be another argument of the function?
Best Regards,
Markus
On 25.04.2012 15:21, Alexander Grayver wrote:
Markus,
My previous email wasn't clear.
For EM, we can have some vector that defines orientation of the unit
dipole, then source is defined as following
/f(p) = n * \delta (p - p_s);/ n = {n_x n_y n_z} -- orientation vector
Elements of this vector should then be assigned to corresponding
cell's components (i.e. E{x,y,z}).
For instance, x-directed unit dipole would give us 3 component vector
with only 1st element == 1.
On 25.04.2012 15:38, Markus Bürg wrote:
Hello Alexander,
I am not sure about that. How is a vector-valued delta-function defined?
Best Regards,
Markus
On 25.04.2012 14:24, Alexander Grayver wrote:
Hello,
I realized that this function works only for scalar FE. Looking at
the implementation it seems the only thing I need to change when
considering vector FE is this:
for(unsigned int i=0; i<dofs_per_cell; i++)
rhs_vector(local_dof_indices[i]) = fe_values.shape_value(i,0);
Here, instead of /shape_value/ I should use /shape_value_component/
for each component and then put dot product to the rhs_vector?
Thanks.
--
Regards,
Alexander
_______________________________________________
dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
_______________________________________________
dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
