I want to determine the resistance
of a rotor bar in a squirrel-cage induction motor, including the skin
effect. I tried using FEMM but the results are not looking right.
I drawed the rotor bar in FEMM,
created a boundary (Prescribed A with all the parameters zero) so that
no flux should cross to the outside of the bar, as was done in the
example on skin effect on the website. (Is this the right method for
this case, or must I draw the whole induction motor?)
I think that you'd want to draw the bar in the slot so that you get
losses due to the slot leakage flux passing through the conductor.
I took a bar length of 1 metre for
simplicity sake (Problem -- depth = 1000), and used a circuit property
to have a current of 50 Ampere flow through the bar. (Bar material =
copper)
To calculate the DC resistance, I
used the Block integral in FEMMview to find the area, and then
calculated the resistance to be 359 microOHM
When using the block integral for
resistive losses (problem frequency = 0Hz) I get 0,898 Watt which
corresponds to my calculation using P = I^2*R
However, when I change the frequency
to 50 Hz, the resistive losses are 0,449 Watt, which is less than the
loss at 0 Hz! (I wanted to use this loss to determine the effective
rotor-bar resistance at 50 Hz)
Am I doing anything wrong? or
understanding anything wrong?
The convention that FEMM uses is that all current, voltages, current
densities, flux densities, potentials, etc. are represented as
amplitude rather than RMS quantities. If currrent "I" is in terms of
amplitude the time average power loss is