Conference Agenda

In this article, we evaluate the coercive field for electrical steel sheets from the interaction of magnetic domain walls with mechanical defects. The stress distribution is computed for the dislocation lines and the interstitial atoms, modeled as a point force. The orientation of the magnetization in the domain wall with mechanical defect is computed with a finite element method. The coercive field is determined from the variation of the domain wall energy with the relative position of the domain wall and the defect. Finally, the results are homogenized and the macroscopic coercive field is compared with experimental results. The prediction of the coercive field corresponds to the measured one for electrical steel sheets with various grain sizes.

The energy-based hysteresis model with dry friction type dissipation using magnetization as internal variable is extended considering a magnetoelastic Gibbs free energy. Magnetic field and mechanical stress loadings are considered and a multiscale model is used to represent the anhysteretic behavior. An explicit evaluation scheme inspired by the vector play approach is proposed. By comparison with uniaxial measurements, the model is shown to generate reasonable piezomagnetic loops and gives very satisfying results in a case with stress and field varying simultaneously.

The modelling and simulation of ferromagnetic materials requires the consideration of magnetic hysteresis. We here discuss the incorporation of the energy-based vector hysteresis model of Henrotte et al., based on previous work of Bergvist, into a vector potential formulation for the magnetic field equations. The field model describing a single step in a load cycle is phrased as a convex minimization problem which allows us to establish existence and uniqueness of solutions and to obtain systematic approximations by appropriate finite element discretization. Consistency of the model with the governing field equations is deduced from the first order

optimality conditions. In addition, two globally convergent iterative solution strategies are presented for the solution of the governing minimization problems. The efficiency of the approach is illustrated by numerical tests for a typical benchmark problem.

In this digest, a Fourier neural operator (FNO) is used to model the dynamic hysteresis behavior of ferromagnetic materials. First, a 1D finite element (FE) model is built, in which, both hysteresis and eddy current are considered. By applying a sinusoidal source at the boundary (magnetic field H) with different frequencies and peak values, a bunch of corresponding dynamic magnetic flux density (B) are generated. Then, all dynamic hysteresis loops (DHLs) are split into 2 groups: training data and test data. Attention is paid to feasibility and accuracy of proposed method in DHLs modeling.

This paper presents an adaptive time-stepping strategy designed for semiconductor transient simulations. The strategy aims to minimize overall computational cost by balancing nonlinear iterations with local truncation error (LTE) estimation. In addition to LTE-based time-step control, the method dynamically adjusts time steps based on the convergence behavior of Newton iterations. While the total number of time steps may slightly increase, the reduction in step size enhances Newton iteration convergence rates, resulting in lower overall computational overhead.

Numerical results of high-voltage semiconductor switching processes demonstrate that the strategy effectively captures sharp physical variations, significantly reduces failed Newton iterations, and enhances overall computational efficiency.