Scattering Analysis by a Stable Hybridization of the Finite Element Method and the Finite-Difference Time-Domain Scheme with a Brick-Tetrahedron Interface
Journal article, 2008

We present scattering computations performed with a newly developed stable hybridization of the finite element method (FEM) and the finite-difference time-domain (FDTD) scheme, which is based on Nitsche's method. This hybrid has not been tested on scattering problems previously, and here we compute the radar cross-section (RCS) for three different targets: (i) the perfect electric conducting (PEC) sphere, (ii) the NASA almond, and (iii) a generic aircraft called RUND. In order to assess the discretization errors associated with the hybrid, we provide comparisons with established results and techniques: (i) the Mie-series for the PEC sphere, (ii) the method of moments (MoM) implemented in the commercial code FEKO, and (iii) a stable FEM-FDTD hybrid that exploits pyramids and a curl-conforming representation of the electric field. The bistatic RCS for the PEC sphere shows second order convergence towards the analytical solution and a relative error of 2% is achieved for about 20 points per wavelength. The NASA almond has a low monostatic RCS in its horizontal plane, which makes it a good benchmark problem for scattering computations. It features a sharp tip and, as a consequence, the order of convergence for the RCS is lowered. Given a careful convergence study for the NASA almond, we achieve a highly accurate monostatic RCS by means of extrapolation and this result is state-of-the-art in the open literature.

finite element method

electromagnetics

FDTD

hybrid techniques

scattering analysis

Author

DAVID DEGERFELDT

Chalmers, Signals and Systems, Signal Processing and Biomedical Engineering

Thomas Rylander

Chalmers, Signals and Systems, Signal Processing and Biomedical Engineering

Electromagnetics

0272-6343 (ISSN) 1532-527X (eISSN)

Vol. 28 1-2 3-17

Subject Categories

Other Electrical Engineering, Electronic Engineering, Information Engineering

DOI

10.1080/02726340701818113

More information

Created

10/8/2017