Determination of thickness and permeability tensor using the combination (models-neural networks)

Abstract : The purpose of this paper is to describe an improved microwave method for predicting the material’s thickness and the saturation magnetization and the damping factor through the neural networks. These characteristics provide the permeability tensor components using the combination between theoretical models and neural network. Neural networks learn the relationship between the scattering parameters and the outputs. The networks’ performances result from both simulation and measurement thin ferrite samples.