Electromagnetic modeling of radiation and coupling characteristics of phased arrays covered by frequency selective surface radomes

First, we present an efficient cascading procedure for analyzing FSS composite comprising of multiple FSS screens having unequal periodicity embedded in dielectric layers. In this approach, an iterative method is introduced to calculate the global period TG for the whole FSS composite, so that only one set of Floquet harmonics is needed of the analysis for the entire composite. Also, a systematical method is adopted to calculate and fill the scattering matrix of each sub-system with entries corresponding to the global period. In addition, the formula for the cascading technique is reformed so that only one matrix inversion is involved. This reduces the time and memory complexity associated and accelerates the acquisition of the scattering properties of the entire FSS composite.; Second, the characteristic basis function method (CBFM) is applied in conjunction with the finite difference time domain (FDTD) method to accurately predict the far field patterns of large finite phased arrays. This can be applied to arrays either standing alone or operating in close proximity of FSS radomes. The CBF's are established by dividing the aperture field distribution of small/moderate size system into sub-regions corresponding to the center, edges, and corners. They are then used to synthesize the aperture field of the large system for far field pattern calculation. This proposed method is especially designed to alleviate the heavy computing resource loads associated with solving electrically large systems, such as large finite arrays covered with FSS radomes. In contrast to the direct method, the increase in simulation time and the burden placed on the memory requirements are incrementally small in the present approach as the problem size is increased from moderate to large.; Finally, the serial parallel FDTD (SPFDTD) method is introduced to accurately evaluate the coupling coefficient between large arrays with a separation that is not large enough to allow the use of ray techniques. The original computation domain is divided into relatively small sub-domains, inside which parallel implementation of the FDTD updating using the MPI library is executed. This is done for the number of time steps designated before proceeding to the next adjacent sub-domain. Tangential field components on the interface are down-sampled in time domain and put away to external files. The fields are then retrieved and recovered as the equivalent sources with an interpolation process for those missed time instants in the following region. The procedures are repeated for every sub-domain until reaching the last one, which is where the coupling is measured. (Abstract shortened by UMI.)...