Aplicação da transformada rápida de Fourier em equações integrais no domínio do tempo para a caracterização de propagação radioelétrica

• Ano de defesa: 2021
• Autor(a) principal: Renata Alves Antunes Teles
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; Brasil; ENG - DEPARTAMENTO DE ENGENHARIA ELÉTRICA; Programa de Pós-Graduação em Engenharia Elétrica; UFMG
• Palavras-chave em Português: Transformada rápida de Fourier; Equações integrais no domínio do tempo; Método dos momentos; Marching-on-in-time; Terreno suavemente irregular; Engenharia elétrica; Equações integrais; Fourier, Transformações de; Método dos momentos (Estatistica)
• Link de acesso: http://hdl.handle.net/1843/35411

This work proposes accelerating prediction of propagation of an electromagnetic wave based on time-domain integral equations by using Fast Fourier Transform (FFT). In this problem, a vertically - or parallel - polarized wave with near graze over a smoothly irregular terrain is considered, so that it can be trated as a perfect magnetic conductor and, by applying the equivalence principle, replaced by equivalent superficial magnetic currents which can be calculated using the Method of Moments (MoM). For the formulation in time domain, marching-on-in-time is used on time discretization for computing time convolution. FFT is applied to accelerate those convolutions. Results from Time-Domain Magnetic Field Integral Equation (TD-MFIE) and Time-Domain Electric Field Integral Equation (TD-EFIE) were compared to frequency domain integral equations - MFIE and EFIE - and to Time-Domain Uniform Theory of Diffraction (TD-UTD), and their simulation times were compared to TD-MFIE with FFT. There were highly concordant results between the methods, despite some differences shown on TD-UTD given the fact that it does not consider double diffractions and diffractions in edges with an internal angle greater than 180°. Using FFT to calculate time convolution integrals reduces significantly calculation time in 28 times, which shows its potential application in this problem, but a greater study about signals envolved in operations is required.