Computational Methods for Large Electromagnetic Problems
With a Focus on Efficient, Optimal, and Predictable Methods for Array Antennas
Time: Mon 2026-12-14 15.00
Location: F3 (Flodis), Lindstedtsvägen 26 & 28
Video link: https://kth-se.zoom.us/j/64271463634
Language: English
Subject area: Electrical Engineering
Doctoral student: Harald Hultin , Elektromagnetism och plasmafysik
Opponent: Professor Cyrille Menudier, University of Limoges, Limoges, France
Supervisor: Professor B. Lars G. Jonsson, Elektromagnetism och plasmafysik; Professor Oscar Quevedo-Teruel, Kommunikationssystem
QC 20261113
Abstract
Array antennas are versatile and used in several disciplines. Accurate modeling of finite arrays requires full-wave analysis to account for edge effects and the fact that each Embedded Element Pattern (EEP) is unique. Full-wave analysis of large arrays is challenging, as the simulation problem grows fast with array size. To address these challenges, a MultiLevel Fast Fourier Transform (MLFFT)-based method is presented to enable full-wave analysis of large arrays by exploiting the periodic array geometries. This avoids common approximations such as unit-cell models. The method combines the MLFFT with new data formats and preconditioners in an iterative solver, that enable efficient analysis of arrays containing more than 1000 elements. Compared with an iterative solver without MLFFT, solution time is reduced by over an order of magnitude. The method can be used to compute the EEP and impedance of each element in the array. The accuracy of rules of thumb and infinite-array models in predicting finite-array behavior is evaluated, and new predictors for active-impedance error as a function of array size are introduced. An optimization method for low Side-Lobe Level (SLL) using EEPs is also presented. Here, the optimization is formulated directly in terms of EEPs, thus installation effects are inherently included. These low SLL methods are extended to provide a tradeoff between Total Active Reflection Coefficient (TARC) and SLL. For more general installed-antenna problems, the thesis investigates the inclusion of near-field terms in an in-house Shooting and Bouncing Rays (SBR)-solver. The study shows that near-field terms improve the accuracy of the solver for a platform with minimal diffraction. Together, these contributions improve the analysis, prediction, and optimization of large antenna arrays while providing insight into the accuracy of solvers for installed-antenna problems.