Multilayer FSS Optimizer based on PSO and CG-FFT

Traditionally, Frequency Selective Surfaces (FSS) are designed to produce a certain spectral response, in one or more frequency bands, using a methodology that manually tunes and simulates the behavior of the structure. In some cases, due to the high filtering requirements of an application, it could be necessary make FSS with several periodic surfaces, in order to improve the frequency response of the final prototype. This process is difficult to be carried out if the filtering response to synthesize cannot be directly obtained with a combination of classical FSS unit cell geometries. To overcome this problem, in this paper a methodology of optimization that automates the hard and costly search is introduced, combining the PSO and an analysis methodology of structures with several periodic surfaces that act as the cost function: the Conjugate Gradient Fast Fourier Transform (CG-FFT). This approximation uses a multilayer Green function calculated applying the transmission line model that considers the mutual coupling effects by translating the field from the source FSS to the other periodic surfaces. Representative results including the optimization of a three layer FSS with two planar periodic surfaces are reported and discussed in order to show the usefulness of the approach proposed.