dolfin team mailing list archive

[Anders Logg <question152565@xxxxxxxxxxxxxxxxxxxxx>]

Question #152565 on DOLFIN changed:
https://answers.launchpad.net/dolfin/+question/152565

    Status: Open => Answered

Anders Logg proposed the following answer:
On Tue, Apr 12, 2011 at 05:41:41PM -0000, Neilen Marais wrote:
> New question #152565 on DOLFIN:
> https://answers.launchpad.net/dolfin/+question/152565
>
> I'm trying to solve a standard electromagnetic problem to test my code. The problem is an infintesimal electric dipole. The linear form involving current would look something like:
>
> V = FunctionSpace(mesh, "Nedelec 1st kind H(curl)", order)
> u = TrialFunction(V)
> v = TestFunction(V)
>
> L =  dot(j*k0*Z0*J_imp, v)*dx
>
> where (ignoring the difficulties of the imaginary part) j*k0*Z0 is essentially a constant, J_imp is a function that represents the impressed current and v is the testing function. This is fine if J_imp is a volumetric current distribution that can easily be represented as a function in V.
>
> An infintesimal electric dipole is modelled as a current point source. In other words J_imp is a spatial dirac delta function. This kind of source is also used to excite some antenna models, which is what I'm working towards. How could one get dolfin to integrate such a dirac-delta function?

There's currently no support for that in DOLFIN. It would require some
work in UFL (adding something like a "Dirac" element) and in FFC
(knowing that it should correspond to evaluating the proper basis
function at a point).

I think this would be useful to add but don't have the time at the
moment.

--
Anders

-- 
You received this question notification because you are a member of
DOLFIN Team, which is an answer contact for DOLFIN.