Re: some proposals for future developments

["atreding" <atreding@xxxxxxxxxx>]

--- In femm@xxxxxxxxxxxxxxx, Mark Smith <mark.smith@xxxx> wrote:
> Hi
> I might be off the ball here but many years ago I was trying to see 
if it
> was practical to measure the air gap in a solenoid valve by 
measuring the
> inductance, which my quickfield model suggested should be possible. 
I found
> that my measured inductance's were much lower than expected which 
turned out
> to be that the permeability at low field strengths was much lower 
than my
> material model due to the fact that the B-H curves are for ideal 
shaped
> fully annealed samples & my cores were "as machined". I went to the 
expense
> of having some parts hydrogen annealed and the results were much 
more in
> agreement with the FEA analysis. Saturation field strengths were 
not so
> different.
> Regards
> Mark

Thanks Mark. Yes, we figured it was partly due to changes in 
permiability at low levels so we actually had some ring samples of 
the iron tested and made BH curves from them, in both the annealed an 
unannealed condition. There were differences (mostly at low levels) 
but not enough to explain the kind of errors we were getting. 
(Although maybe for inductance it makes more difference - we were 
only concerned with force.) One thing that did help (although not 
completely) was to treat the iron as a weak permanent magnet oriented 
against the coil. That is, we took the Hc of the iron from the 
samples we tested and put that into the materials library as a new 
material. This was enough to partly fix things and I've done it on 
low saturation models since. But it still is somewhat unsatisfying as 
this "solution" is really just a fudge factor. From what you are 
saying, it seems that maybe the finished part itself needs to be 
annealed to bring it in line with the model, implying that the 
machining stresses are causing a larger Hc than our sample measured. 
That would be unfortunate since there would be no way to predict it, 
and annealing is too expensive an operation to do for some of these 
units.
But thanks again for your insight - it is nice to know that others 
have had similar problems and that something as simple as 
annealing/machining stresses might be enough to explain it - even if 
I can't calculate around it.
Andy