Space-mapping optimization of lossy coupled-resonator filters exploiting rational approximation models

Abstract

Space mapping is a useful method for optimization of microwave structures using full-wave electromagnetic simulations. This paper proposes an efficient and robust implementation of the space mapping algorithm for lossy coupled resonator filters. The method makes use of a circuital-parameter-based coarse model that presents numerous advantages over traditional coarse model such as 1) the ability to model cross-coupling and losses; 2) direct correction/calibration of the coarse model in order to include parasitic interactions extracted from the fine model; and 3) improved parameter extraction (PE) through estimation of rational polynomials. A numerical examples of the proposed methodology corresponding to a 4th order substrate integrated waveguide (SIW) filter with transmission zeros is presented in order to demonstrate the performance and versatility of the method.