On the variational characterization of a laminated core model in ferromagnetic materials

Abstract

Accurate prediction of iron losses in transformers, electric motors, and generators at elevated frequencies requires accounting for eddy currents in laminated magnetic cores during field simulation. We study local material models for these lamination losses that resolve the inhomogeneous magnetic-field distribution across the thickness of the respective steel sheets via one-dimensional diffusion equations. This allows to incorporate skin-effect and eddy-current losses in the laminations at each material point of a two- or three-dimensional magnetic-field simulation. The macroscopic material response is described implicitly through a minimization problem involving a set of internal variables which encode the load history and current state of the material. The variational structure of these problems allows us to prove well-posedness, to include anisotropy and hysteresis, and to develop efficient globally convergent solvers for the evaluation of the macroscopic material response. The proposed framework thus allows to efficiently and reliably predict magnetic behaviour of laminated core material over a wide frequency range and is well-suited for integration into nonlinear magnetic field solvers.