Otimização robusta aplicada em projetos de antenas impressas

• Ano de defesa: 2019
• Autor(a) principal: Fernando José de Souza Magalhães
• 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: HFSS; MATLAB; Otimização; Otimização robusta; Projeto de antenas impressas; Projeto robusto; Engenharia elétrica; MATLAB (Programa de computador); Otimização combinatória; Antenas (Eletrônica)
• Link de acesso: http://hdl.handle.net/1843/30973

This work proposes the design of patch antennas using robust optimization to obtain antennas that are less sensitive to uncertainties in the manufacturing methods and the electrical characteristics of the materials used. Low-cost methods and materials may have significant uncertainties that impact the final performance of a patch antenna and a robust optimization approach may be one way to address such uncertainties to reduce the cost of patch antenna designs. Two robust optimization approaches are proposed: one using the mono-objective methodology with genetic algorithms and the other using the multi-objective methodology with the NSGAII algorithm. Both methodologies are modified to reduce the sensitivity of antenna performance parameters to the uncertainties of design variables. The proposed methodology was implemented in two types of antennas, where the uncertainties of the manufacturing method and also the uncertainties of the FR-4 substrate were considered. The optimization routine was implemented using MATLAB, where the antenna performance parameters were obtained through the Ansys developer HFSS program. Two antenna models were made and validated through the use of the equipment and semi-anechoic chamber provided by the university, and it was observed that there was a reduction in sensitivity between the selected and less robust antennas of the last generation obtained by the algorithm, reducing the sensitivity. of the gain realized at 15,25% for the rectangular patch antenna and 18,07% for the circular polarization patch antenna. The axial ratio of the circular polarization patch antenna was reduced by 11,89%.