Jacobi-Davidson iteration based reduced order finite element models for radar cross-section

Abstract

Finite element modeling of large-scale scattering problems like radar cross-section (RCS) lead to large matrices. Model order reduction (MOR) using eigenspace projection of these models involving plane wave excitation requires solution of nonlinear eigenvalue problems. This paper proposes a methodology to expediently compute wideband RCS. JacobiDavidson iteration is used to generate the projection spaces. Reduction by many orders in the system size and the computational cost is exemplified.