Re: line integral and antiperiodic BC

["David Meeker" <dcm3c@xxxxxxx>]

Ack--this is one of those sort of annoying bugs, because I can't 
duplicate it on my installation--I get the same result in all 3 cases. 
I think what is happening is that, in your installation, the 
program is deciding that points that are on the boundary are 
actually just outside of the boundary. The B.n integral works by 
just taking the difference of the values of A at the endpoints of 
the contour, so if you get an A=0 from a point that it thinks is 
outside your geometry, that would explain the difference in reported 
values.

Anyhow, I'll look into this and see if there is some way that I can 
make this more robust...

Dave.


--- In femm@xxxx, <vecerkaj@xxxx> wrote:
> Hi everybody
> I've met the problem with simple line integral calculating the total 
flux and
> average Bn. I posted example, where the situation occured. It 
presents a PM
> synchronous machine (only half of machine, used antisymetric BC) 
> When I chose the contour for the calculation of average Bn over one 
rotor pole
> in the airgap (see example.gif) I realized, that the result depends 
on the
> way, how I selected a contour for the calculation. 
> 1. When the startpoint is selected inside of the region and endpoint 
is on the
> boundary (antisimetric BC), the result is Bn1=0 (nonsense).
> 2. When the startpoint is selected on the boundary, and endpoint is 
inside the
> airgap, the result of line integral is certain value Bn2=nonzero.
> 3. When I choose the contour arbitrary (by right button of mouse) in 
the
> similar location and I obtain the different value of the Bn3 
(Bn3=2*Bn2).
> 
> It seems, that correct value of Bn is the value Bn3. But I expected 
the same
> values obtained by different ways(except sign).
> 
> If anybody interested, I included also fem file. Maybe I make all 
the time
> same stupid mistake, can anybody show me the right way or a reason 
of
> differences?
> 
> Note the example uses linear characteristic of iron, just because of
> computation speed (in reality I use nonlinear BH characteristics).
> 
> Thanks in advance
> 
> George Vecerka
> Department of power electric and electronic
> Faculty of Electrical Engineering and Computer Science
> Brno University of Technology
> Czech Republic
> 
> 
> ____________________________________________________________________
> Get free email and a permanent address at 
http://www.netaddress.com/?N=1