Discussion Forum

Hi, I'm new to COMSOL and I was trying to do a stationary magnetic fields (mf) simulation in 2D axissymmetric in COMSOL. The model was really simple, which consisted of:

1. a cylindrical permanent magnet (with its axis aligned with the overall symmetric axis(z axis))
2. a cylindrical iron core (whose axis also aligned with z axis)
3. copper coil
4. air.

And for different domain, I have different size settings under my mesh node, with the one in area 1 being the finest since it's smallest in terms of its geometry, and following that is area 2 and area 3, and finally, area 4 has the coarsest mesh.

Since this project is mainly about the force exerted on the permanent magnet and from what I've learned, the value of electromagnetic force can be subjected heavily to the size of your mesh, I decided to use the "adaptation and error estimates" function to see if my results converge. I left all the settings in the "adaptation and error estimates" dropdown menu as default except for "maximum number of adaptations" which I set as 5, and ran the study.

However, after the study was done, I checked the mesh, and I encountered a very strange phenomenon: the meshes in area 1 and area 2, aka the areas with their edges overlapped with the z axis, remained almost the same as adaptation level increased, while the meshes in area 3 and area 4 got significantly smaller.

So I want to ask: How did this "resistance to refinement" in certain areas happen and more importantly, how to avoid it so meshes in every domain can be refined consistently across the board? I would greatly appreciate any help!!!

Also I've attached some results of the meshes as well as the google drive link to my mph. file down below.

Mph file