Accelerated PMM solutions via model-based parameter estimation and fast Fourier transform techniques

Since their introduction, frequency selective surfaces (FSS) [1] have found a wide range of applications as electromagnetic filters, radomes, and more recently even as components of metamaterials. Spatial-domain and spectral-domain formulations of the periodic moment method (PMM) [1, 2], as well as periodic versions of the hybrid finite element boundary integral (PFEBI) method [3] have been widely used for the analysis of FSS. PMM formulations only requires the discretization of one unit cell in a two-dimensional periodic FSS screen and employ a layered-media Green’s function to account for the presence of the dielectric substrate and superstrate. This leads to one important advantage of the PMM; namely, it is very fast compared to the PFEBI-based method when it comes to the analysis of FSS structures composed of a homogenous layered-media. The most computer-intensive operations in the PMM are filling the impedance matrix and solving the linear matrix equation to obtain the current distribution on the FSS screen. Under certain circumstances of practical importance, these operations can become computationally burdensome, especially when performing frequency and angle sweeps. Impedance matrix interpolation techniques were first proposed in [4] to significantly reduce the large matrix fill times associated with certain antenna analysis problems. A modified three-sampling-point interpolation technique was later applied in [5] to more efficiently predict the response of FSS screens.