Helix (nonlinear magnetic) under Homogeneous Magnetic Field (Convergence problem)

Dear Sir,

I saw that you've been answering a lot of these questions in the COMSOL forums, and was wondering if you could solve me the following problem, since I've been fighting with it for days with no success:

I have a Helix (in the microscale) made out of a nonlinear magnetic material (Cobalt Steel Vacoflux 50) which I'm subjecting to a magnetic flux density of 1700 T exerted in the boundaries of a much bigger air sphere sorrounding it. When I simulate the problem, it converges if I chose the material to have a linear BH curve (B = mu*H) but doesn't converge if I use it's actual BH curve.

I have tried everything that involves changing the mesh and changing the damping of the newton algorithm with no success. Now my only idea is to simulate the problem with a linear BH curve first in the first step of the study and then use the results as initial values for the nonlinear problem (second step) but i'm not really sure how to do it, since there are so many options and so many places where one can define initial conditions.

I would very much appreciate your help! Thank you!

Best,

Michel Heusser
ETH Zürich