NEURO-GENETIC OPTIMIZATION OF MAGNETIC HYSTERESIS INTEGRATES IN ELECTROMAGNETIC SYSTEMS
Résumé
In this work we have presented an approach for calculating the hysteresis loop of Jiles-Atherton model using the magnetic inductance as the independent variable is proposed to be used directly in the calculation time step finite volume applied to the numerical analysis of nonlinear magnetic fields. This model is characterized by five parameters that must be identified and optimized for better representation of the measured characteristics. The parameters set of the Jiles–Atherton hysteresis model identified by using a real coded genetic algorithm. The parameters identification performed by minimizing the mean squared error between experimental and simulated magnetic field curves. The method verified by applying it to an axi-symmetrical ferromagnetic system. The calculated results validated by experiences performed in a Single Sheet Tester's frame (SST). In this work, we are interested to develop a model based on feed-forward neural networks of which can describe magnetic hysteresis by taking account the influence of some external sizes.
Origine : Fichiers produits par l'(les) auteur(s)
Loading...