Discussion Forum

I'm trying to model the force-field topology of various permanent magnet geometries as the magnets get closer and further away from each other, using 3D modeling mfnc. Unfortunately, every time I attempt to set up a nested Parametric Scan to cover a range of magnet separations and off-axis alignments, the scan invariably halts after 20 or so evaluations with a "Degenerate Polygon -- could not insert point" type of message.

It does not reproducibly occur during the same iteration, and the magnets are not coming into close enough contact that a mesh node would touch another domain (in fact, during the parametric scan, they are getting further and further away from each other). I do notice that after a parametric sweep, the Mesh node has altered itself from "Physics Controlled" to having a entry for "User Defined, Remaining Domains"

I did set up the geometry for the model in COMSOL, using a Work-Plane and a Revolve transform. But this seems to be due to something about the parametric scan itself: If I put in the exact same parameters by hand as the failed location in the parametric search, the mesh and evaluation occur without problem.

How can I either avoid encountering the "Can't insert point" error, or at least manage to restart/continue the parametric scan with the results being placed into the correct Probe Table that had been being generated -- Continue does NOT do it?

Thanks!