Multiband FSS Analysis and Synthesis Based on Parallel Non Coupled Metallic Strips Using WCIP Method

Abstract

A novel parallel non coupled metallic strips frequency selective surface (FSS) is presented. The design is based on a filter composed of horizontal and vertical metallic strips. The structure has the advantage of simplicity and gives multi resonant frequencies easily controlled by a simple variation in the strips length. The proposed FSS rejects frequencies at 7GHz, 9.4GHz with bandwidths of 0.646GHz, 0.793GHz respectively when the structure is excited with an x polarized plane wave, and two frequencies at 9.1GHz and 11.2 GHz with bandwidth of 280.3MHz, 63MHz respectively when the structure is excited with a y polarized plane wave. To suppress frequencies ideal diode on reverse bias are inserted. The simulated results obtained using WCIP (Wave Concept Iterative Method) are compared to the COMSOL Multiphysics 4.3b software results, and measurements, a good agreement is observed. Then, an FSS synthesis approach based on non coupled parallel metallic strips is presented. It provides synthesized FSS with resonant frequencies ranging from 4.5GHz and 12.25 GHz for metallic strips lengths inversely varying from 5mm to 19mm. To validate the synthesis approach, measured resonant frequencies are used as desired resonant frequencies to determine the metallic strips lengths. As a result the metallic strips FSS dimensions are extracted. At this stage the WCIP method is used to characterize the synthesized FSS. Three FSSs are synthesized based on the fabricated FSS and good agreement between measurements and synthesized FSSs results is recorded.

Synthesis of Multiband Frequency Selective Surfaces Using Machine Learning With the Decision Tree Algorithm

“…In [12], the FSS synthesis is performed using a hybrid multiobjective bioinspired method, combining a GRNN network and an optimization algorithm (MOCS -Multi-Objective Cuckoo Search), where the dimensions and periodicity of geometry are obtained from resonance frequency and bandwidth estimations. In [13] and [14], the physical dimensions are obtained from the desired resonance frequencies using GA with ANN (Artificial Neural Network), and WCIP (Wave Concept Iterative Process), respectively. In [15], the dimensions of geometry are obtained from the angle of incident wave and resonance frequency desired.…”