MAGNETIC MATERIAL PROPERTY IDENTIFICATION USING NEURAL NETWORKS

Abstract

In the numerical solution of field problems for many devices, the modeling of the hysteresis properties of a material is extremely important. While considerable work has been published on models for handling this behavior, it is still difficult to obtain the parameters for these models from measured data. This paper proposes the use a neural network for this operation. [Vol. 12, No. 2 (1997), Special Issue on Numerical Field Calculations in Electrical Engineering, pp 44-49]