Maxwell 2D: simulation with saturated cored

- April 23, 2025 at 12:04 pm
  
  slawomir.szot
  Subscriber
  
  I want to perform 2D Eddy simulation of magnetic circuit where core starts to saturate, in other words where B is on the knee of nonlinear B-H curve. Is this normal that it takes very long time to solve? I made few attemps to solve it but without succes, as you can see in convergence below solver was not able to achieve required error despite quite large number of mesh elements.
  I'm sure model is OK - when I lower excitations to be on linear part of B-H curve it solves in few seconds with roughly 3000 mesh elements.
- April 23, 2025 at 6:14 pm
  
  MirandaH
  Ansys Employee
  
  Hi, the energy error is going up after pass 11, there might be something going on. Maxwell treats the BH curve as a piecewise linear curve between the user-defined data points, do you have enough points at the knee area?
- April 25, 2025 at 1:05 pm
  
  slawomir.szot
  Subscriber
  
  I added more points in the knee area (now there is about 20 points there) but it didn't help:
  After plotting B field I noticed there are very small spots where B is extremely high, example screen after pass 10:
  What shall I do to improve this? Would changing parameters: nonlinear residual and max number of nonlinear iterations help?