Kriging Methodology for Predicting Material Uncertainty Impact on FEM Scattering Computations

Abstract

We present a Kriging interpolation surrogate function for use in reconstruction of probability density function in scattering uncertainty quantification problems, focusing on predicting material uncertainty impact on finite-element scattering computations. The results show extremely high reconstruction accuracy while reducing computation time by approximately two orders of magnitude relative to the conventional Monte Carlo approach. This tool offers a robust option for future uncertainty quantification problems in the field of computational electro magnetics.