Waveguide Microwave Imaging: Solids Volume Fraction of Particulate Materials

An original modeling-based microwave imaging technique for determining the volume fraction of solid material in dielectric powders is described. The desired characteristic is determined by analyzing S-parameter measurements in a waveguide containing the sample with the help of an artificial neural network trained by data from 3D FDTD simulation. The powder sample is represented by a mixture of air and millimeter-scale particles reproduced in the FDTD model. Computational tests with 20 to 40 mm cubic samples of SiC and ZrO2 powders in WR340 show that the solids volume fraction is determined with less than 5% error.