Dear David:Your request was for a recommendation for a commercial program that would be good for examining the distant leakage field. Although this is certainly possible to calculate this sort of thing with finite elements, boundary element techniques are well-suited to this particular problem since no outer boundary of the solution region need be defined. One example of such a commercial code would be IES Amperes, which is a good and fairly intuitive program. As far as freeware, there is Radia (http://www.esrf.fr/machine/groups/insertion_devices/Codes/Radia/Radia.html), which is actually a volume integral code (rather than a boundary integral code) but has the the same nice properties with respect to evaluating the field at distant locations. However, you have to have Mathematica installed to be able to run Radia.
I need your help again. Please see attachment in figures about the question:
1. for ferrite, we want to know the leakage field at
one side region about 3~5 times the ferrite dimension.
basically it is a 3D problem. You suggested to use BEM
methods to solve it. But I found BEM is difficult to
pick up.
My original suggestion still stands--at a big enough distance, everything looks like a dipole (or perhaps a collection of a small number of dipoles). You could use a 2D finite element solution to back out a dipole moment for your device as seen from a distance, and then shove that dipole moment into the equation for the field of a dipole in 3D space.2. Is this a good idea? To change the vector problem into a potential problem as in "Answer.bmp"? Due to mu_r is much larger than 1, we can claim each side have different potentials. The rest will be a finite-difference problem to get all the potential distribution of the region. B will be the gradient of potential.
Please comment and suggest.
Best Wishes,
Bruce