Development of an open optimization framework for aeronautical applications

• Ano de defesa: 2014
• Autor(a) principal: Alexandre Pequeno Antunes
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Instituto Tecnológico de Aeronáutica
• Palavras-chave em Português: Projeto de aeronaves; Algoritmos genéticos; Estrutura de aeronaves e Helicópteros; Redes neurais; Configurações aerodinâmicas; Fabricação de aeronaves; Engenharia aeronáutica
• Link de acesso: http://www.bd.bibl.ita.br/tde_busca/arquivo.php?codArquivo=3082

The aeronautical industry, more precisely the aircraft manufacturing sector, is a demanding business area in which the design development cycles are continuously shrinking whilst the technical requirements are becoming more stringent due to the fierce competition. The present work considers the adoption of the multidisciplinary design optimization concept, which is also known by the MDO acronym, as a way of adapting to this new reality. In the MDO concept, the design is performed in a concurrent fashion through the integration of the engineering processes in environments know as ``frameworks';';. The work presents the development of a set of tools that can be adopted as numerical procedures inside existing frameworks or they can be coupled to create the basic structure of an open MDO framework, focused in aeronautical engineering and with special attention to aerodynamic design problems. These tools are embedded in different modules and they are employed in a series of study cases focused in aeronautical applications. These studies have shown how aspects associated with the choice of the geometrical parameterization and the upper and lower range limits of the parametric variables can yield different geometries during the optimization process. Moreover, the present work shows that only those geometric parameterizations that consider high order polynomials can guarantee that the same final geometry is achieved at the end of the optimization process. The increase in the polynomial order leads to optimized solutions with lower drag coefficients. The aerodynamic optimizations performed with a neural network have shown the benefits that the approximation methods can provide in terms of computational cost. The complete set of tools developed during this work can contribute to improve the capability of the Computational Fluid Dynamics group at Instituto de Aeronáutica e Espaço (IAE) and at Instituto Tecnológico de Aeronáutica (ITA) by the incorporation of this open environment for analysis and multidisciplinary optimizations. These tools can become an initial structure focused in a collaborative research activity between academia and industry.