Machined core not as strong as simulations

[abishek20]

Hello,
I had simulated a core (low carbon steel 1008) wound with Cu wire in Ansys magnetostatics. An armature (permanent magnet in this case) placed in a gap of the core was simulated to calculate forces.

However the force I am getting from an actual machined 1010 steel core is significantly less (more than factor of 2).

Why is this happening? Is this due to stress formed by machining that could be affecting the grain structure? Should I look into annealing/plating with high permeability mat? please help !

 

[IRstuff]

Are you having a typo error? You simulate one composition but use a different composition and then ask why there might be a difference?

TTFN

FAQ731-376

 

[abishek20]

No. I didnt have 1008 material. However 1008 and 1010 are very close, I dont think the material change can warrant a factor of 2 change in performance.

 

[MagnetDave]

That seems like a lot of change. On the merits of composition, I would expect something more like <10% change. That said, the stress state of the material is a potential source of the problem - generally, annealed materials have the best magnetic properties. Was your 1010 cold-rolled? That could be a source of your headache.

However - You should also check your model. What is the material curve for the 1008 like? Perhaps you have something unrealistic there. Secondarily, what is the cooling on your coil like? Heating of the wires could also be part of the problem. Lastly, what sort of magnet did you model and what did you use?

I realize these are all measurement-model agreement questions that you would ask for every problem, but given the short first post, I gotta ask.

 

[IRstuff]

A simulation is an abstraction and simplification of reality. That being the case, any difference between the simulation and the as-built object should be subject to scrutiny.

Is everything in the as-built structure IDENTICAL to the simulation?

TTFN

FAQ731-376

 

[dgallup]

The most important factor in a simulation is the material B/H curves you input. Where did they come from? Some solvers come with pretty poor material properties data built in. I have no idea what, if anything, Ansys provides. If you are getting into high flux densities, look at how the curve is blended into the "air line" at saturation. I know FEMM used to do a very poor job in this area when I tried it years ago which led to wildly optimistic force predictions. Others have posted that this has been corrected.

I use MagNet in conjunction with a material properties database I have built up over the years. It is usually within about 10% of measurements but always high. Everything is always perfect in a simulation, real parts have imperfections.

 

[MagMike]

I suggest re-running your model without the armature, at DC current. Run the same current through the copper coil, measure the magnetic field on the face of the steel and see how it compares to Ansys.

Adjust the material properties of the 1008 in Ansys until it matches your readings. This will help partially validate your model. After this, re-run the model with the armature.

BTW: You mentioned your armature is a permanent magnet. If it is Alnico or Hard Ferrite, there is a chance it became slightly demagnetized.