Re: [femm] Re: Circuits in FEMM

[dcm3c@xxxxxxx]

In a message dated 9/5/01 12:52:32 PM Eastern Daylight Time, 
frank.e.lenning@xxxxxxxxxx writes:


> Thanks Dave.  What is the equation that FEMM solves for the time 
> harmonic case and where does it do it in the FEMM source code?

I've put a brief note in the "files" section on the extra equation that the 
program solves associated with each current constraint:
http://groups.yahoo.com/group/femm/files/currentconstraint.pdf

The source code associated with it lives (for the 2d planar case) in the 
prob2big.cpp and (for the axisymmetric case) prob4big.cpp files of fkn.  The 
code that enforces it is sort of strewn around, because the extra equation 
really gets built implictly.  Only the upper diagonal of the matrix is 
stored, because of the symmetry of the problem, so the extra equation gets 
built by adding extra forcing terms associated with the unknown gradient to 
the elements that live in the regions associated with the circuit.  Look for 
the comments that show where the circuit parts live (e.g. lines 
78-89,294-302, 335-340 in prob2big, similar implementation in prob4big) .  In 
the source code, the voltage gradients are included in the vector of unknowns 
right after the nodal values of potential.  The voltage gradients also get 
written to disk in the .ans file.

Dave.