A SPARSE-MATRIX CANONICAL GRID METHOD FOR ANALYZING MICROSTRIP STRUCTURES

In this paper, we illustrate the analysis of microstrip structures wit h a large number of unknowns using the sparse-matrix/canonical grid me thod. This fast Fourier transform (FFT) based iterative method reduces both CPU time and computer storage memory requirements. We employ the Mixed-Potential Integral Equation (MPIE) formulation in conjunction w ith the RWG triangular discretization. The required spatial-domain Gre en's functions are obtained efficiently and accurately using the Compl ex Image Method (CIM). The impedance matrix is decomposed into a spars e matrix which corresponds to near interactions and its complementary matrix which corresponds to far interactions among the subsectional cu rrent elements on the microstrip structures. During the iterative proc ess, the near-interaction portion of the matrix-vector multiplication is computed directly as the conventional MPIE formulation. The far-int eraction portion of the matrix-vector multiplication is computed indir ectly using fast Fourier transforms (FFTs). This is achieved by a Tayl or series expansion of the Green's function about the grid points of a uniformly-spaced canonical grid overlaying the triangular discretizat ion.