Font Size: a A A

Development of PSTD methods for wireless communication applications

Posted on:2005-05-26Degree:Ph.DType:Dissertation
University:University of South CarolinaCandidate:Liu, XiaopingFull Text:PDF
GTID:1458390008986094Subject:Engineering
Abstract/Summary:
In recent decades, a variety of computational electromagnetic methods have been developed and applied in the analysis and design of high-speed microwave and millimeter-wave components and systems, as well as in the prediction and evaluation of large-scale scattering phenomena. Among these methods, finite difference time domain (FDTD) is a popular, straightforward, and robust scheme for solving Maxwell's equations. However, numerous examples show that a sampling density of 20 cells per minimum wavelength is necessary in order to obtain reasonably accurate results. This requires a large amount of computer memory and time, hence strictly limiting the applications of FDTD to electrically small objects.; In order to improve the computational efficiency and accuracy of FDTD, pseudo-spectral time domain (PSTD) methods have been proposed and are forming an important and rapidly developing field. In this dissertation, we research the two dominant types of PSTD: Fourier PSTD and Chebyshev PSTD. Fourier PSTD only requires 2 cells per minimum wavelength while Chebyshev PSTD needs pi cells per minimum wavelength; both have an infinite order of accuracy. We have developed systematic transformed-space non-uniform Fourier PSTD (TSNU-PSTD) schemes and field solvers for one-, two- and three-dimensional problems. We have also developed systematic Chebyshev PSTD algorithms and field solvers for 1D and 3D applications. These algorithms and field solvers have been successfully applied to various electromagnetic wave propagation and scattering problems and proven that PSTD schemes are much more efficient than the FDTD method in terms of the requirement of computer memory and CPU time. PSTD methods have great potential in the simulations of large-scale problems.; We have also explored the algorithms to evaluate the space-transformation factor (STF) in the Fourier TSNU-PSTD method. In addition to the development of PSTD schemes, we have developed the flexible and easy-to-use software tool CurveReader which can read the coordinates of an arbitrary black curve under a white background in polar and Cartesian (both linear and logarithmic) coordinate systems. Finally, we have developed the highly efficient non-uniform body-of-revolution FDTD (BOR-FDTD) algorithm, field solver, and the user-friendly mesh generator, for the analysis and design of microwave components and devices with rotationally symmetric geometries.
Keywords/Search Tags:PSTD, Methods, Cells per minimum wavelength, Developed, FDTD
Related items