Re: [femm] Force and torque from stress tensor

[David Meeker <dcm3c@xxxxxxxxxxxxx>]

Alberto Perez wrote:

> As you know I'm simulating an induction motor. Now I'm calculating
torque and
> force from stress tensor of two poles motor. I have calculated as
Meeker says in
> his manual and I have seen example of horsecore in femm web. To
calculate Fmax,
> I'm simulating in magnetostatic and I take 180º of the one pole
airgap. I get the
> following values: [...]

> The forces calculate from femm and from me are similars but torques
aren´t them.
> I don?t understand torque from stress tensor, is this torque the
torque max can I have
> obtain? Can I calculate the max electromagnetic torque at
magnetostatic? Or I have to
> calculate with nominal frequency = 50Hz and starting current?
>
> What relations is between force and torque? I have high force but the
torque is too low.
>
> Thank you in advance

You can't really calculate the torque on an induction motor at DC. At
DC, there's no torque (except, perhaps, for some cogging torque). What
you can do instead is perform runs at a range of low frequencies. The
frequencies correspond to the slip frequency, and the torque that you
get out of the machine corresponds to what you'd get at synchronous
speed (e.g. 50 Hz) minus that slip frequency. You'd have to run a
range of frequencies to find out the max torque for a given current.
The for a fixed current, the torque rises with increasing slip
frequency, up to a frequency equal to 1/(rotor time constant). For
higher slip frequencies, the torque decreases.

There's not necessarily any correlation between the force result and the
torque result. The force integral yields the net translational force on
the rotor--this force doesn't tend to spin the rotor. In contrast, if F
is the force dictated by the stress tensor at some point on your
contour, the contribution of that point on the contour to the torque is
computed by r X F, where r is a vector going from (0,0) to the point of
interest, and the X represents the cross-product. When this quantity is
integrated on a contour that runs all the way around the rotor, you get
the total torque on the rotor. If you think about it, you can imagine
cases where you can get a net translational force but no torque (e.g. in
a magnetic bearing), or a torque without any net translational force
(which is actually what one desires in a rotating machine).

Dave Meeker