Toward Adaptive Multiscale Computational Methods and Uncertainty Quantification for Electromagnetic Analysis of Microwave Circuits, Systems, and Antennas

One of the most severe challenges in electromagnetic analysis of microwave circuits, systems, and antennas is constituted by a need for resolving non-smooth and multiscale behavior in modeling and simulations. Precise modeling of such inconvenient solution behaviors is extremely burdensome and even for relatively simple structures, this can render conventional approaches infeasible. Material interfaces, sharp or semi-sharp geometric features, etc. drive poor solution regularity, resulting in extremely poor rates of convergence without new techniques. Similarly, when formalizing uncertainties in the analysis (e.g., for material parameters, surface roughness, etc.), current computational methods are incapable of handling this problem adequately in the presence of multiscale characteristics and non-smooth solution behavior. This is especially critical for enhancing the automation of design of microwave circuits, systems, and antennas and even more so in safety-critical and mission-critical applications.