A ten-parameter model for the static hysteresis simulation of ferromagnetic materials

Abstract

This work proposes a model with dynamic parameters based on the classic Jiles-Atherton model for magnetic hysteresis. The goal of this study is to investigate whether the improved model is able to generalize the material behavior correctly when minor loops are involved. Two non-linear optimization techniques are used for parameters identification: a hybrid algorithm based on Genetic Algorithm (GA), Trust-Region-Reflective (TRR) and Levemberg-Marquardt (LM), and a novel continuous technique called Continuous Flock of Starlings Optimization (CFSO). Hysteresis loops used as reference were generated with the Preisach model, and the analysis is performed on a wide set of virtual materials and excitation waveforms.