Skip to Main Content
An unsymmetrical printed circuit board with a ground plane has been modeled using Genetic Algorithm by its equivalent dipole set for the Near-Field matching and Far-Field prediction. Far-Field measurements have been predicted using Near-Field measurements in which the actual radiation source has been replaced by an equivalent set of dipoles. In previous work, the Genetic Algorithm has been employed to find out the equivalent dipole set only within the surface boundary of the radiation source. Since most of the practical radiation sources are unsymmetrical in nature, the magnetic fields above and below the ground plane are significantly different. In this paper, the dipole locations are allowed to lie outside the surface boundary of the original source to match the unsymmetrical Near-Field distribution and to predict its Far-Field strength. Simulation results show that the improved Genetic Algorithm not only demonstrates its viability as a Far-Field predictor but also that Genetic Algorithm can be optimized to search for an equivalent dipole set outside the boundary of source. By taking four different cases of observation planes, it is also shown that increasing the number of observation planes as compared to increasing the number of matching points on a single observation plane gives improved results for the Near-Field matching and Far-Field approximation of unsymmetrical printed circuit board. Other important aspects like the selection of observation points, the number of equivalent dipoles required have been addressed.
Date of Conference: 12-13 March 2010