Dynamic Ferromagnetic Hysteresis Modelling Using a Preisach-Recurrent Neural Network Model

Christian Grech; Marco Buzio; Mariano Pentella; Nicholas Sammut

doi:10.3390/ma13112561

Dynamic Ferromagnetic Hysteresis Modelling Using a Preisach-Recurrent Neural Network Model

Materials (Basel). 2020 Jun 4;13(11):2561. doi: 10.3390/ma13112561.

Authors

Christian Grech^{1

2}, Marco Buzio², Mariano Pentella^{2

3}, Nicholas Sammut^{1

2}

Affiliations

¹ Faculty of Information and Communications Technology, University of Malta, MSD2080 Msida, Malta.
² CERN, European Organization for Nuclear Research, 1211 Geneva, Switzerland.
³ Department of Applied Science and Technology, Politecnico di Torino, 10129 Turin, Italy.

Abstract

In this work, a Preisach-recurrent neural network model is proposed to predict the dynamic hysteresis in ARMCO pure iron, an important soft magnetic material in particle accelerator magnets. A recurrent neural network coupled with Preisach play operators is proposed, along with a novel validation method for the identification of the model's parameters. The proposed model is found to predict the magnetic flux density of ARMCO pure iron with a Normalised Root Mean Square Error (NRMSE) better than 0.7%, when trained with just six different hysteresis loops. The model is evaluated using ramp-rates not used in the training procedure, which shows the ability of the model to predict data which has not been measured. The results demonstrate that the Preisach model based on a recurrent neural network can accurately describe ferromagnetic dynamic hysteresis when trained with a limited amount of data, showing the model's potential in the field of materials science.

Keywords: ARMCO pure iron; Preisach; dynamic hysteresis loop; machine learning; magnetic properties; particle accelerators; recurrent neural networks.