Magnetic Property Calibration of Low-Permeability Materials for Vibrating Sample Magnetometers Using Neural Networks

Abstract

This article proposes a new method to calibrate the B–H properties measured by the vibrating sample magnetometer (VSM) for bar-shaped samples. It compensates for finite size and demagnetization effects using the neural networks (NNs) trained based on finite element analysis. Once the NNs are trained, they provide fast prediction of calibration results. The results from the proposed method are compared with those from the capacitive cancellation method, regarded as a reference due to its immunity to the effects of the demagnetizing field.