Improving material models in FEMM.

Check these threads;



[link ] [/url]

Thanks,

I see in one of the threads that I'm not the only one to notice odd behavior in the materials models. Here are all of the loops from my FEA (FEMM). I'm sure they were collected from various sources and I understand the batch to batch variations can be quite drastic but i still find it odd that they would chose models that show opposite of what really happens.



Here they are plotted on a log scale to see the differences a bit more clearly.



In any event, these are all the DC magnetization curves, and what I am actually looking for are the BH loops (showing losses and remenance). I have a good conceptual idea of how the various alloys and treatments will behave from Bozorth, but it would be great to get all of the info of the various alloys plotted on the same (X,Y) graph.

dave

FEMM provides what are called "normal permeability curves". It represents the tips of the minor hysteresis loops.

To adjust the hysteresis, you need to adjust the paramter Phihmax which controls the phase angle between B and H. Small angle between B and H corresponds to line between the two endpoints and no losses. As you increase the angle, it gives you a wider oval between the two endpoints (more losses). It is not exact, but I believe pretty close for most points below the knee of the curve with proper selection of Phihmax.

Then there is also a property Hc that can be specified. I'm not sure what it does.... seems like it would be redundant with the other parameters since if we know the shape of the oval we know Hc.



=====================================
Eng-tips forums: The best place on the web for engineering discussions.

My (limited)experience with FEMM is that it does not properly blend the supplied B/H data with the air line. Instead, it just extrapolates the slope of the last two points out as far as it needs to. Then if you run a steel part into heavy saturation FEMM will give you ridiculously high B values. You need to make sure the last points in the data have a relative permeability of 1. Other magnetic FEA programs that I've used do this automatically for the user.

dgallup

This is not "ridiculously high B values". As one of the old users for FEMM we had bad results with old versions of FEMM because on high H values the program didn't have values of B. What happens is that when the steel parts are in heavy saturation they start to behave as an air therefore, the permiability of the iron is equal to the permiability of aor.

Therefore, Dr. David Meeker added the air permiability values to the steel B-H curve so that the program can work with heavily saturated manetic circuits without crashing or bad results. Therefore, the straight line beyond B=~1.9 to 2.0 (for stainless steels it will be from lower B values) is basically the B-H curve of the air.

bit of advice with the software, it does not have an adequate uninstall package, so you may spend a bit of time clearing the various entries from your registry.

israelkk

"Therefore, Dr. David Meeker added the air permeability values to the steel B-H curve so that the program can work with heavily saturated magnetic circuits without crashing or bad results. Therefore, the straight line beyond B=~1.9 to 2.0 (for stainless steels it will be from lower B values) is basically the B-H curve of the air."

When I used FEMM (quite some time ago) it did not have this feature. That is what I called "blending with the air line". It's good to know FEMM has this now, it is absolutely necessary for accurate result in strong fields. Maybe I will try FEMM again.