Investigação de técnicas de condições de contorno para problemas de eletromagnetismo em alta-frequência

• Ano de defesa: 2007
• Autor(a) principal: Junia Taíze Santos Roberto Compart
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: Engenharia elétrica
• Link de acesso: http://hdl.handle.net/1843/BUDB-8D4LWJ

The main purpose is to develop a critical study on the use of absorbing boundary conditions (CCA) for high-frequency electromagnetic scattering problems in bidimensional closed domains. In particular, the investigation is focused in the precision of two sets of CCAs. Thefirst is constituted of analytical conditions, with emphasis in the techniques suggested by Mur, Trefthen and Higdon; the second is composed of absorbing conditions, with emphasis in the method proposed by Berenger, the perfect matching layer (PML). The finite difference time domain (FDTD) method is the numerical technique used to solve the electromagnetic scattering problem and to implement the boundary conditions. The model used in the numerical simulations is a bidimensional domain, filled with air, with 100 cells in the x direction and 50 cells in the y direction. The electromagnetic field source was represented by a hard source, located in the center of the domain.The numerical results obtained by the FDTD were validated using the definition of relative error, which compares the numerical result with the analytical solution. The precision of CCAs was investigated using the definition of local error, calculated in points of the domain were the highest reflections are expected; and the definition of global error, thatrepresents the accumulated error in the domain throughout time. The results indicate that the relative errors of the analytical CCAs are in the order of 10-3 whereas in the CCA PML it is in the order of 10-5. The CCA PML is, therefore, 100 times more accurate. The local error obtained for the analytical CCAs are in the order of 10-2 whereas the CCA PML is in the order of 10-6. For the global error, the analytical CCAs presented errors in the order of 10-4 and the CCA PML in the order of 10-5. The CCA PML presented a higher precision, but also presented a higher difficulty ofimplementation. On the other hand, the analytical CCAs of Mur, Trefethen and Higdon showed less accurate results, when compared to the CCA PML, but an easier implementation.