This paper presents a framework composed by Domain Decomposition Method (DDM) that integrates a local mesh refinement scheme based on Huygens’ equivalence principle and the integration of multibranch basis functions for the efficient analysis of multiscale problems. The proposed method allows the improvement of the precision of the solution while the solving time is reduced via the decomposition of the problem.

Slotting the impedance matrix to perform the matrix-vector product through the use of FFT, procedures named slotFFT techniques, is an efficient way to accelerate the solution of electromagnetic analysis of finite periodic structures through the SIE-MoM framework. This work presents the combination of a slotFFT technique with two different matrix compression methods, one based on full domain power-decoupled macrobasis, and other based on skeletons to increase the achieved acceleration, showing the versatility of the proposed slotFFT algorithm to be integrated with different frameworks.

The efficient and precise analysis of large finite periodic arrays in the scope of the nanotechnology is a current challenge in computational electromagnetics. SlotFFT techniques [1]-[2] have proven to efficiently perform the analysis of this kind of problems. It is a transparent method which maintains the precision of the surface-integral equation-method of moments (SIE-MoM) framework [3]-[4] but reducing the cost of the matrix-vector product (MVP) from O(N 2 ) to O(nMlog(M)), where n is the number of unknowns of the element, M the number of elements, and N=nM.

Approaching large finite periodic structures with the surface integral equation-method of moments (SIE-MoM) framework is a challenging kind of problems in terms of memory and computing time. Slotting the MoM matrix with slotFFT techniques has proven to be an efficient approach to reduce both metrics through the application of FFT. In this work we propose the integration of a matrix compression method with a controlled impact on the precision based on full-domain distributed basis to improve the acceleration given by the slotFFT method.