Learning Forum

[tawhidtarek]

I have run an eddy current simulation for an actuator in Ansys Maxwell. The main target was to observe the coil inductance with respect to the frequency. We got our result in the simulation. However, when we verify the result with the experimental one, they do not match at all. I have checked the material properties and dimensions of the .aedt file. But still, I have not been able to find out the reason behind this huge discrepancy.

[DELI]

Hi, this is a pretty big difference, could you tell us more detail about this design?
is the coil design the same as real test? dimension, geometry, the number of strand.
is the core design matching with real test?
Is the excitation matching the real test? Is there any saturation on the core?
How is mesh? Mesh quality would affect calculation accuracy.
is there any movement in real test? If yes, does it include spring force and/or friction? Eddy Current solver is a steady state solver which will not consider motion.

[tawhidtarek]

The coil design is the same in terms of dimension, geometry, and the number of strands.
The core is also the same.
The excitation is also matching the real test. You can find the equivalent circuit of the test in the attachment.
The mesh quality is also good. I have checked the mesh statistics under solution. The standard deviation is in the range of e-11. The picture of the mesh has been attached.
There is no movement in the real test. The terminals of the coil were connected to the inductance measurement device.

[tawhidtarek]

Hy,
Just to add, there was no saturation in the core in the simulation. And we have used HP4194A Impedance Analyzer to measure the inductance in different frequencies.
Regards Tarek

[DELI]

Mesh looks pretty good, I can see some sort of eddy current from the mesh plot.
Do you have eddy effect on the core in this simulation?
Looking at the plot of L vs Frequency of real test and simulation, it seems the calculated inductance from simulation reacts to frequency change, which looks like eddy effect. However, the experimental L curve doesn't react a lot from frequency change, is the real core laminated?

[tawhidtarek]

Hy Thanks a lot for your comment. The real core is laminated. In the material properties, I also set the core as laminated in the simulation.
However, the eddy effect was "on" for all the parts in the simulation. I set the eddy effect off the materials except for the magnet . I reran the simulation and now the results look like this:

still, there is a significant difference between the values. Can you please think of any other issues that can cause the difference? Thanks in advance.

Regards Tarek