This article proposes an advanced methodology to deal with the complexity of composite materials modeling up to 60 GHz. For radiofrequency (RF) requirements, it has been demonstrated that the distribution of conductive inclusions plays a major role. Since their locations are intrinsically subject to uncertain assumptions, the Monte Carlo (MC) technique is considered as a golden standard. Unfortunately, the computational costs involved by coupling full-wave electromagnetic (EM) simulations and

more »

... remains prohibitive. The aim of this proposal is to demonstrate the interest of stochastic reduced order method (SROM) to tackle computational constraints, jointly with the statistical precision needed for a realistic description of RF composites.