|M.Sc Thesis||Department of Electrical Engineering|
|Supervisor:||Prof. Leviatan Yehuda|
|Full Thesis text|
Among the various frequency-domain solution methods for electromagnetic wave scattering problems, the source model technique (SMT) is distinguished by the relative simplicity of its software implementation. Moreover, under certain conditions it can be the most effective solution technique. Nevertheless, while the use of the SMT in frequency-domain solutions is quite common, there is a relative void in its implementation in time-domain solutions. The aim of this research was to determine the properties and the limitations of the SMT in time-domain solutions of three-dimensional transient problems and in turn to find ways of overcoming these limitations. As with any other boundary formulation, the difficulty of applying the SMT to scattering problems in the time domain stems from instabilities caused by
spurious modes with widely disparate time scales. In this work a study of the factors, both spatial and temporal, influencing the convergence of the SMT solution has been carried out. Concerning the spatial setup, it was found that the use of combined sources, placed and oriented in such a way that their maximal radiation is directed towards the field sampling points, is essential for solution stability. In addition, there are bounds on the extent to which the sources can be densely spaced. As to the temporal discretization, it was found that the use of high-order discontinuous expansion functions renders the time-marching scheme both stable and rapidly converging. Representative computational results that demonstrate these findings are
shown and discussed.