Add setup with circuits + impedance bc

[arvedes]

@ettaka I setup a more complex test case, currently I cannot solve it with https://github.com/ElmerCSC/elmerfem/commit/bf4e04227020cf64c911118b82479679374e7834. I tried the following:

• case_circuit_impedance-bc.sif gives a trivial result (maybe there is still an issue in the setup?)
• case_circuit.sif doesn't work, the solver doesn't converge. It used to work with https://github.com/ElmerCSC/elmerfem/commit/0ff29c7213bfe81ab9f79de7650118d69bec1ab4 (but required much more memory). I also tried to solve with mumps but got a segmentation fault in WhitneyAVHarmonicSolver (with both https://github.com/ElmerCSC/elmerfem/commit/bf4e04227020cf64c911118b82479679374e7834 / https://github.com/ElmerCSC/elmerfem/commit/0ff29c7213bfe81ab9f79de7650118d69bec1ab4).
• case_layer-conductivity.sif works (even though there was this issue on the develop branch https://github.com/ElmerCSC/elmer-elmag/issues/10).

My setup is based on this Dockerfile.

[ettaka]

Hi @arvedes ! Good job. I see at least one typo: https://github.com/nemocrys/elmer-elmag/blob/circuits_impedance-bc/InductionHeating/harmonic_open3Dmassive_circuit.definition#L69

Needs to point to the Boundary Condition, not Body.

[arvedes]

Thanks, it's working now with a relaxed linear system convergence tolerance of 1e-6. Do you have an idea how to improve that?