Re: [femm] Few questions ...

[dcm3c@xxxxxxx]

In a message dated 11/16/00 1:43:35 PM Eastern Standard Time, 
jrcamacho@xxxxxx writes:


> Dear finite elements experts,
> We are currently working with linear induction motors on femm, the 
> machine is double face with an aluminum linor 1/2 cm thick in the 
> midle. About this configuration we have few questions:
> 1. In a double face three-phase linear induction motor with a 
> circular linor, how one can calculate the propelling force applied to 
> the linor?


Specify three-phase currents in the coils by defining either current 
densities or "circuit" properties with the appropriate complex components so 
that the phasing is right.  Then, analyze the problem at the slip frequency 
of interest.  To get forces, you can either integrate Lorentz (that is, J X B 
forces) over the volume of the secondary, or you can integrate Stress Tensor 
over the surface of the stator.  The Lorentz force tends to converge faster, 
and I'd therefore suggest it. Unfortunately, motion isn't implemented in the 
current version of femm (and might not be for quite some time -- convection 
is a pain to deal with...).  You can only get force results for the "locked 
shuttle" case.  When there is motion, you can get significant drag from 
higher harmonics that doesn't show up in the locked shuttle case, and the end 
effects caused by motion can have a very significant effect (especially in 
short-!
stator LIMs, but not so much in long-stator LIMs). 

> 2. By marking the whole area of the core in one of the faces, the 
> program is giving us a core loss equal to zero. How the area should 
> be marked to give a different value?

Hmmm--You should be able to evaluate the core losses by selecting the core 
area and then performing the "total losses" integral.  To get reasonable core 
losses, you must define a conductivity, lamination thickness, packing factor, 
and hysteresis angle for your material--if you don't have at least some of 
these defined, you will always come up with zero core losses. 

3
> . How one can introduce the third dimension in the program? The 
> results given are a fuction of the metrical unity for the third 
> dimension?

Well, femm is fundamentally a 2-D program.  I've tried to kludge in 3-D 
effects in linear induction motor problems in the past by modifying 
(lowering) the secondary's conductivity to take account of the resistance of 
the part of the current path in which the currents have to "turn around," but 
this is at best approximate.  Each harmonic has its own apparent conductivity 
(because the fraction of the current loop in the "edge" sections of the plate 
is different), and there are additional leakage fluxes associated with theses 
secondary currents and the coil end turns that are neglected.

Anyhow, I've been messing around with LIMs a lot lately, but becuase of the 
somewhat inherently 3-D nature of these machines, my models have been largely 
analytical, lumped parameter ones.  I have then used FEMM from time to time 
to "giggle test" these analytical models by considering degenerate cases of 
the analytical models that coincide with a 2-D analysis.

Dave.
--
http://members.aol.com/dcm3c