Discussion Forum

Hi All,

I am trying to implement a nonlinear Mu_r in my MEF simulation of a simple toroid. In my current simulation I am using the 'Nonlinear_mu-metal" from the ACDC Module's example: one-sided magnet under industrial applications (this is 4.2 by the way). I can run and solve the example one-sided magnet just fine and it is in 3D. I have set up a model of one turn of a toroid with a constant terminal current of 1 [A]. This model works with any constant value of mu_r.

However, whenever I try to use the one-sided magnet mu_r (the example model uses mfnc) it tells me there is a circular variable dependency detected and fails miserably. I have looked at many of the other circular dependency threads (most are unsurprisingly uninformative as it can completely depend on the model). I was wondering if anyone has tried implementing this mu_r from the example:

1-1/(1/(1-mu_rmax)-(mu0_const*(H1^2+H2^2+H3^2)^(1/2)/B_sat)^(1/N))
mu_rmax = 40000; B_sat = 1.2 [T]; N = 1;

PS: I have tried using the BH loop defining to solve this model as well and it just never solves the model (the convergence plot either levels out or "does not converge" finally).

Could this mean that my method of constant current shouldn't be used, or is there something I'm overlooking?

Thanks in advance,
Brock