Otimização de estruturas reticuladas incluindo não-linearidade geométrica

Abstract

It is common in many types of framed structures, often used as roofs of gymnasiums, hangars, warehouses, etc, the structural analysis considering a geometrically nonlinear behavior. This analysis (more complex) allows for an evaluation closer to the real behavior of the structure and thus its design requires to an evaluation considering this behavior in order to avoid undesirable structural problems. Structural optimization problems are studied in this dissertation using a bioinspired metaheuristic, the Genetic Algorithm (GA), searching for optimal solutions (minimizing costs and satisfying safety constraints) for structures with geometrically nonlinear behavior. It seeks for optimal structural designs taking into account sizing design variables, such as areas or dimensions of the cross sections of the bars. The GA makes use of special encodings of the chromosomes, by using simple and multiple cardinality constraints, searching for the best grouping of the bars of the structures. Comparative studies are made by considering the optimization of structures through evaluations using linear and nonlinear analysis, and highlighting the major differences encountered when using either type of analysis in seeking for the optimal solutions.