Reply To: AlNiCo hysteresis simulation

[pavel.kirienko]

I think I found at least part of the problem: I had job parallelization enabled for the transient solver in the HPC options. This caused the time decomposition method (TDM) to be chosen for the problem. The TDM is unable to simulate hysteresis effects directly, as mentioned in the docs at Maxwell Help > Specifying Solution Settings > Setting Analysis Parameters for Transient Solutions > Time Decomposition Method for Maxwell Transient Designs:

Strictly speaking, TDM cannot support hysteresis modeling. But for soft hysteresis materials, especially for lamination core loss computation, the hysteresis loop, or the coercivity Hc, is very small. [NB: it is not small for AlNiCo]

This is quite dangerous as the simulation completes successfully without any kind of warning that the results are invalid. I disabled TDM and got much better results. They don’t quite match what I got with the bench prototype yet, though.

I ran two simulations:

1. With the BH curve provided with the LNG37 alloy from the standard material library replaced with a core loss model as explained in Maxwell Help > Assigning Boundaries and Excitations for 3D Designs > Setting Demagnetization/Magnetization Computations for Source Designs > Using the Hysteresis Model-Based Magnetization Approach.
2. With the original LNG37 material definition.

The first one resulted in a much better result that approximates what I get with the bench prototype. The second one resulted in the sample being demagnetized — the B vector field is oriented randomly inside the cylinder after the current pulse is removed, and the magnitude is about zero.

Could you please confirm that this:

A normal BH curve with its knee point is sufficient.

is sufficient to compute both demagnetization and magnetization of the body? That does not seem to be the case or I'm doing something wrong.