Decomposition and Space Mapping for Reduced-Cost Modeling of Waveguide Filters

Abstract

In this work, we present a technique for low-cost surrogate modeling of waveguide filters. The proposed methodology is based on the decomposition of the filter structure. Some of the decomposed parts are modeled using response surface approximations (RSAs). The RSA models are subsequently combined with analytical models of the waveguide sections to form an initial filter surrogate. As a result of electromagnetic couplings between the decomposed parts, which are not accounted for by the initial surrogate, its accuracy is limited. This misalignment is reduced by applying space mapping at the level of the complete filter structure. Decomposition approach allows us to greatly reduce the computational cost of creating the surrogate because the time required to simulate the structure in parts is much lower than the time for simulating the entire filter. Moreover, the number of parameters describing each part is lower than for the entire filter. The presented technique is demonstrated using two test cases. Application examples are also given.