Re: [femm] Problem in force calculations

[Dave Squires <djsquires@xxxxxxxx>]

Eckart Nipp wrote:

> Hi,
>
> integration contours on surfaces are likely to cause errors in force
> computations. Furthermore take care that the contour is closed. Besides
> moving the outer boundaries far enough away, I would suggest to set
> zero A on them.

I have even been careful to make sure that the contour is equidistant
all the way around by using the snap grid feature and making sure I
go clockwise. The snap grid feature insures closure of the contour.

>
> Another question is the approximation of the actual 3D situation by a 2D
> problem. Is the geometry rotational symmetric? Or is the approximation
> by an infinitesimal long geometry applicable?

This might be part of it. In reality you have a 3D situation. This one is
not axi-symmetrical. It is just two magnets and an iron plate between them.

> In addition I think in this case you must take great care of dimensions,
> permeability, saturation, magnetization, etc.. Just think about what happens
> if the steel sheet is very thin or very thick. The behaviour would be
> completely differnt.
>

Of course. That is what I am trying to simulate. A thicker piece will show
a switchover from repulsion to attraction in real life.

Dave Squires