Improving the Jiles-Atherton model by introducing a full dynamic dependence of parameters

An improvement of the Jiles-Atherton model by introducing a dynamic dependence of all five parameters is presented. The values of parameters dynamically change as a function of the magnetic field excitation. The proposed model aims to improve the poor performance of the classic model Jiles-Atherton in generalizing both saturated and minor loops of a given material. Non-linear optimization techniques such as Genetic Algorithm (GA), Trust-Region-Reflective (TRR) and Levenberg-Marquardt (LM) were used by a hybrid configuration for the parameter identification, using hysteresis loops generated with the Preisach model as target. Different cases of excitation trends are investigated and final validation and comparison results are presented