Re: [femm] Determining inductance per meter with 2D analyses

["Robert Macy" <macy@xxxxxxxxxxxxxx>]

That explains it! I did indeed have unbounded boundary conditions. 

- Robert -


----- Original Message ----- 
From: David Meeker 
To: femm@xxxxxxxxxxxxxxx 
Sent: Thursday, September 12, 2002 7:56 AM
Subject: Re: [femm] Determining inductance per meter with 2D analyses


gopherlid wrote:

> Anybody have an idea how to calculate inductance per length?
>
> I tried using the formula 1/2 L I^2 = Energy
>
> but the Energy value changed too much based upon number of nodes. 
> We're talking 100% It's as though, more nodes, more energy with no
> asymptotic limit in sight!
>
> It's a very simple structure, a wire with a slab of metal around it,
> so the results should be straight forward. However, the "magnetic
> field energy" seems to relate to the number of nodes *and* the number
> of nodes in the metal, which makes no sense, since there is less than
> 0.01% energy in the metal. But changing the nodes in the metal
> actually made the inductance change by a factor of 2:1 
>
> Any ideas? Plus ideas on how to know when enough is enough?
>
> - Robert -

It sounds like problem that you are trying to solve is ill-posed. The 
inductance of a single wire in an unbounded 2-D space is infinite. The 
result that you get when you try to analyze this problem then strongly 
depends on how you truncate the solution domain. To get a configuration 
with a bounded inductance, all of the current in your domain must add up 
to zero, or you must model a domain where the existence of the return 
wire is implied your the choice of boundary conditions.

There are some examples that could be relevant to you at:
http://femm.berlios.de/induct1/induct1.htm
http://groups.yahoo.com/group/femm/message/347

Dave.
--
David Meeker <http://femm.berlios.de/dmeeker>