Dimensionamento ótimo de pilares mistos preenchidos de aço e concreto

• Ano de defesa: 2019
• Autor(a) principal: Lourenção, Jéssica Salomão
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal do Espírito Santo; BR; Mestrado em Engenharia Civil; Centro Tecnológico; UFES; Programa de Pós-Graduação em Engenharia Civil
• Palavras-chave em Português: Design; Concrete Filled Steel Tube; Optimization; Dimensionamento; Pilares Mistos Preenchidos; Otimização; Otimização estrutural; Colunas; Construção mista; MATLAB (Programa de computador); Engenharia Civil; 624
• Link de acesso: http://repositorio.ufes.br/handle/10/11230

The use of steel profiles of tubular sections for filled composite columns is an interesting solution for civil construction. This type of profile, due to its geometric shape, presents great resistance to the soliciting forces in general, having three types of crosssections, rectangular, square and circular, with or without longitudinal reinforcement. In Brazil, the design of concrete filled steel tube is covered by NBR 8800:2008, however, considering that the tubular profiles have certain peculiarities, the Brazilian Association of Technical Standards published the NBR 16239:2013 in order to present more accurate calculations for these types of profiles. In Europe, the design is covered only by EN 1994-1-1:2004. Thus, in order to reduce costs, in general, due to the existing load, it is desired to define an optimum geometry. In this way, this work aims to gather the criteria for the elaboration of a software with graphical interface that will realize the optimal dimensioning of concrete filled steel tube using Matlab software under the prescriptions of NBR 8800:2008, NBR 16239:2013 and EN-1994-1-1:2004. The optimization methods used were deterministic algorithms, using the Inner Method and Sequential Quadratic Programming Method, and probabilistic algorithm, using the Genetic Algorithm Method. The results found in the software developed demonstrated that the Inner Method and the Sequential Quadratic Programming Method obtained the lowest costs while the results found for the Genetic Algorithm Method were relatively higher due to the use of tubular profile catalogs.