Selected Developments in Computational Electromagnetics for Radio Engineering

Dissertation for the degree of Doctor of Science in Technology to be presented with due permission for public examination and debate in Auditorium S4 at Helsinki University of Technology (Espoo, Finland) on the 26th of May 2001 at 12 o'clock noon. which are all gratefully acknowledged. Finally, I use the opportunity to express my gratitude to my wife Mari, who took the risk to marry a numerician. Abstract This thesis deals with the development and application of two simulation methods commonly used in radio engineering, namely the Finite-Difference Time-Domain method (FDTD) and the Finite Element Method (FEM). The main emphasis of this thesis is in FDTD. FDTD has become probably the most popular computational technique in radio engineering. It is a well established, fairly accurate and easy-to-implement method. Being a time-domain method, it can provide wide-band information in a single simulation. It simulates physical wave propagation in the computational volume, and is thus especially useful for educational purposes and for gaining engineering insight into complicated wave interaction and coupling phenomena. In this thesis, numerical dispersion taking place in the FDTD algorithm is analyzed, and a novel dispersion reduction procedure is described, based on artificial anisotropy. As a result, larger cells can be used to obtain the same accuracy in terms of dispersion error. Simulation experiments suggest that typically the dispersion reduction allows roughly doubling the cell size in each coordinate direction, without sacrificing the accuracy. The obtainable advantage is, however, dependent on the problem. In the open literature, a few other procedures are also presented to reduce the dispersion error. However, the rather dominating effect of unequal grid resolution along different coordinate directions has been neglected in previous studies. The so-called Perfectly Matched Layer (PML) has proven to be a very useful absorbing boundary condition (ABC) in FDTD simulations. It is reliable, works well in wide frequency band and is easy to implement. The most notable deficiency of PML is that it enlarges the computational volume-in open 3-D structures easily by a factor of two. However, due to its advantages, PML has become a standard ABC. In this thesis, the operation of PML in FDTD has been studied theoretically, and some interesting properties of it not known before are uncovered. For example, it is shown that, surprisingly, PML can absorb perfectly (i.e. with zero reflection) plane waves propagating towards almost arbitrary given direction at given frequency. Optimizing the conductivity profile …