Re: [femm] New FEMM user

[David Meeker <dmeeker@xxxxxxxx>]

eavogels wrote:
> Dear Mr. Meeker.
> My name is Eric Vogels and I will try to learn FEMM. I followed this 
> group almost half a year and I must admit that the level where 
> everyone is discussing models is too high for me to understand. 
> Unfortunately I have not the right education but since many years 
> I'm experimenting with magnets. Now I want to learn FEMM in order to 
> be able to build my models in the program first, before building in 
> real live. Do you think there is a change that I can use FEMM by 
> only reading the manual and looking at examples or is fysics 
> education neccesary. Please advice me. I only use permanent magnets, 
> mounted on wheels. I wonder is examples with permanent magnets are 
> available. I'm good in cutting and pasting other peoples work.
> In case you want to see where I work with in my spare time look at: 
> <http://fdp.hemsida.net>
> Thanks (and thank you for the nice program).
> Eric Vogels.
Configurations of permanent magnets can be modeled by femm, but they must
be configurations that can well-approximated by a 2D planar geometry or configurations
that are rotationally symmetric (e.g. like a typical speaker design).  

I know of many users who have been able to set up and analyze programs and
get OK results without much of a formal background in magnetism.  For example,
I've seen some pretty good simulations that have been created by high school
students who where interested in modeling coil guns. However, as has been
pointed out in several messages in the "FEMM Vs. commercial
code" thread, without a fairly good understanding of the physics behind
the program, it is possible to make models that aren't reasonable.  Since
the finite element is only an approximate method of solving the differential
equations that are usually employed to model magnetic phenomena, it is possible
to create geometries or meshes that give bad results if you don't have some
understanding the numerical methods behind the program (and their shortcomings).
 Most importantly, it can be especially difficult to understand the results
of the program without a good theoretical understanding of what the program
is supposed to be doing.  In some sense, the program is really just like
any calculator--it crunches the numbers, but doesn't have any intrinsic understanding
of the problem that it's solving--that's the user's job.

Anyhow, I guess that I have been sort of remiss in putting up permanent magnet
examples on the website.  As an example, a fun PM configuration to look at
is a Halbach array, which is an array of permanent magnets that tends to
concentrate the PM flux on one side of the array.  Since it's only about
1K, I've attached a sample of a wavelength's worth of such an array.  It's
a good example in the sense that it shows how permanent magnets can be been
defined and oriented in the program. There have also been some pretty good
threads that had to do with specific PM machine modeling problems on the
mailing list. (e.g. the collection of messages that was spawned by http://groups.yahoo.com/group/femm/message/1229
having to do with small brushless DC motors).

Dave.
-- 
David Meeker
dmeeker@xxxxxxxx
http://femm.berlios.de/dmeeker

Attachment: zip00037.zip
Description: Zip compressed data