Re: boundary condition and torque calculation

["David Meeker" <dmeeker@xxxxxxxx>]

--- In femm@xxxxxxxxxxxxxxx, "Juan Tapia" <jtapia@xxxx> wrote:
> 
> 
> 
> Dave
> 
> I was looking for information about periodic and antiperiodic boundary
> condition for FE analysis and I found a file in the files section of the
> yahoo groups website. I downloaded (pbc-examples.zip) and proved the
> example. My question is about the torque value that I got. For the model
> which consider one-half of the machine (speedmotor.fem) using
> antiperiodic boundary condition I calculated -1.88e-3Nm, but when I
> calculate with the complete model (360 degree) I got 9.47e-3Nm which is
> one half of the previous value. Do those value make sense? I mean, If I
> have 180 degree model I obtain double value respect to the complete
> model?
> 
> Juan

This model uses "circuit" properties to define the currents in the
coils. Remember that the program only understands regions that are in
parallel, not series. If you just use mirroring to create the other
half, you'll add regions in parallel, halving the current in the
coils. The overall result would be as you'd described.

At any rate, I realize that the way that the currents are defined with
the "circuit" properties is confusing, and I'm head-scratching about
the right way to redo this functionality in the program so that the
construction and connectivity of coils is done better and in a more
intuitive way.

Dave.
--
David Meeker
http://femm.berlios.de/dmeeker