Neural-Network-Based Model for Dynamic Hysteresis in the Magnetostriction of Electrical Steel Under Sinusoidal Induction

In this paper, we present a model for the dynamic hysteresis behavior of magnetostriction in electrical steel under sinusoidal induction. The model can be used for the numerical calculation of vibrations in magnetic cores. In order to keep the calculation time of the method to an acceptable level, we developed a neural-network-based model, which predicts magnetostriction loop shapes of the material under a limited set of circumstances but offers fast evaluation time. As an example, we apply the model to a grain-oriented electrical steel and present an error analysis. The model can be extended for use with nonsinusoidal induction.