Using neural networks in the identification of Preisach-type magnetostriction and field-temperature models

Abstract

It is well known that the identification process of Preisach models is, basically, reduced to the determination of weight functions of the operators upon which the model is built. Recently, two-dimensional Preisach models have been proposed as a possible approach for modeling one-dimensional magnetostriction as well as field-temperature effects on magnetized bodies. In this article, the use of neural network machinery is suggested to solve identification problems of such models. The suggested technique has been implemented and experimentally tested. Testing results suggests that this technique can lead to reasonably accurate simulation results.