The finite-difference time-domain (FDTD) method has been widely applied in solving electromagnetic problems. Its capability of handling electrically large or high-Q structure problems is, however, limited by the requirements of large computation memory and time. Such requirements are due to the numerical dispersion errors and the CFL stability condition. So far, most of the research efforts have been focused in developing schemes such as MRTD and PSTD that possess low numerical dispersion and therefore require low computation memory. In this paper, we will present another direction in improving the FDTD computation efficiency: removal of the CPL stability condition. In other words, we will present an unconditionally stable 3D finite-difference time-domain method where the FDTD time step, is no longer restricted by the CPL stability condition but by the modelling accuracy of the FDTD algorithm only. As a result, FDTD iteration number and CPU time are reduced. To further reduce numerical dispersion, anisotropy and memory of the method, a high-order scheme is also presented. Theoretical studies and numerical examples will be presented to validate the proposed schemes.
[1]
Richard Holland.
Implicit Three-Dimensional Finite Differencing of Maxwell's Equations
,
1984,
IEEE Transactions on Nuclear Science.
[2]
M. Krumpholz,et al.
MRTD: new time-domain schemes based on multiresolution analysis
,
1996
.
[3]
H. H. Rachford,et al.
The Numerical Solution of Parabolic and Elliptic Differential Equations
,
1955
.
[4]
K. Yee.
Numerical solution of initial boundary value problems involving maxwell's equations in isotropic media
,
1966
.
[5]
T. Namiki,et al.
A new FDTD algorithm free from the CFL condition restraint for a 2D-TE wave
,
1999,
IEEE Antennas and Propagation Society International Symposium. 1999 Digest. Held in conjunction with: USNC/URSI National Radio Science Meeting (Cat. No.99CH37010).