[Siegunn]

Dear Foamers,

i'm working with a simple steadystate-solver for calculating the current density within a conductor.
The solved equation is fvm::laplacian(sigma,V),where V is the electrostatic potential and sigma is the conductivity. The current density J is simply calculated by J=sigma*(-fvc::grad(V)).

In principle the solver works fine but in the case of an 1-dimensional resistor a problem appears at the cells next to a surface with a step in sigma, which can be seen in the attached picture below. The highlighted cellvalues correspond to the calculated current density, which should be homogenous over the whole volume.

I think this problem comes from a missing boundary condition at this surface, which should be something like this:

n * (J_a-J_b)=0,
where n is the surface normal vector and J_a, J_b the current densities of both sides to the surface.

Actually i have no idea how to set this boundary condition, because it is not a static value or gradient. Could someone give me a hint to this problem?

Cheers
Marcel

wrongJ.jpg