Efficient Identification of Jiles–Atherton Model Parameters Using Space-Filling Designs and Genetic Algorithms

Abstract

The Jiles–Atherton model is widespread in the hysteresis description of ferromagnetic, ferroelectric, magneto strictive, and piezoelectric materials. However, the determination of model parameters is not straightforward because the model involves numerical integration and the solving of ordinary differential equations, both of which are error prone. As a result, stochastic optimization techniques have been used to explore the vast ranges of these parameters in an effort to identify the parameter values that minimize the error differential between experimental and modelled hysteresis curves. Because of the time-consuming nature of these optimization techniques, this paper explores the design space of the parameters using a space-filling design. This design provides a narrower range of parameters to look at with optimization algorithms, thereby reducing the time required to identify the optimal Jiles–Atherton model parameters. This procedure can also be carried out without using expensive hysteresis measurement devices, provided the desired transformer’s secondary voltage is known.