[femm] Re: No Subject / Materials....

[Dcm3c@xxxxxxx]

In a message dated 3/14/00 1:09:48 PM Eastern Standard Time, 
jvilla@xxxxxxxxxxxxxxx writes:

> Hi,everybody
> First let me introduce my self. I'm Juan Luis Villa From University of
> Zaragoza (Spain);I'm developing my PHd THesis on the modelization of Lineal
> Induction Motors and your program is very usefull for me. But I have a
> problem with the modelization of induction current. I'm trying to
> modelizate induction current on an aluminium's sheet.
> 
> Does anybody Know how to do it?; is it posible to modelizate with femm?

femm could be somewhat useful in analyzing LIMs. One way to go about it is 
to model your machine and then run analyses at various frequencies. Since 
the model is static in the sense that the plate isn't moving, the frequency 
that you are specifying is the slip frequency of the machine. You can get 
some possibly useful results about the relationship between current, slip, 
and force at zero speed in this way.

However, you have to be careful in interpreting 2-D LIM models. Edge effects 
and higher harmonics aren't adequately represented in a 2-D model without 
motion, and these phenomena can have important implications on machine 
performance. I'd suggest using femm as a check on the magnetostatic fields 
predicted by analytical models (current sheet models, for example) and as a 
method of obtaining parameters for use in analytical models that include 
these extra 3-D effects.

If you just want to look at the effects of some current distribution in the 
plate, you can break the plate into a number of sections and apply volume 
currents so that a sinusoidal current distribution is approximated.


In a message dated 3/14/00 6:11:17 PM Eastern Standard Time, 
scott_norton@xxxxxxxxxxxx writes:

> Like others, let me first claim my ignorance about magnetics...
> (Why do the equations have to seem so much more complicated than
> electrostatics!?!?)

Well, for your sort of problem, you can at least gain some intuition from 
electrostatics. It is possible to formulate a problem with just magnets as 
scalar potential problems, just like electrostatics. The ends of your 
magnets are like surfaces with a distributed charge, the surface of any iron 
object is (ideally) at a constant potential, and so on.

> My mission: To increase the B field gradient in a very small
> localized volume. I have a couple nice fairly large magnets and I 
> want to introduce some strips of metal material around my small volume
> to cause large field gradients. I'd like to explore my options 
> using FEMM. 

If you can idealize your configuration as a 2-D or axisymmetric problem, femm 
ought to be able to do the trick.

> First of all, what type of material am I looking for that enhances
> magnetic field gradients? What's the material specification that tells
> one this? For instance, I've seen nickel strips introduced into a uniform
> magnetic field that was used to create gradients. Why was
> it chosen? What's better? If I was to model nickel in FEMM, do I need
> the BH curves or is magnetic permeability enough?

Rather than just pure nickel, the strips that you have seen are probably a 
high nickel alloy like mumetal, permalloy, or supermalloy. Mumetal and 
supermalloy are in the femm materials library. These materials have an 
extremely high permeability, but a fairly low saturation flux density.

Depending on what you are doing, these materials might be overkill--plain 
iron or some low carbon alloy might be adequate. The permeability of these 
materials isn't anywhere near as high as the nickel alloys, but it's still 3 
or 4 orders of magnitude more permeable than air (i.e. looks almost 
infinitely permeable with respect to air).

At any rate, without a knowing more about the application, it's hard to say 
too much more. For example, it's possible to think up configurations of just 
permanent magnets that have really steep field gradients--what is an 
acceptable material / geometry depends upon what you are trying to do.

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