Microwave Characterization of Dielectrics Materials Using the Combination of a Finite Element Technique and Neural Networks

摘要

In this paper, a microwave characterizationtechnique based on the use of a coaxial discontinuity is presented. The inversion is carried out by neural networks (NN) to determine the complex permittivity of dielectric materials from the measurement of the admittance at the plane of discontinuity between coaxial waveguide and a measurement cell. In order to train the NN, several data sets are created by finite element modeling.