Projeto ótimo de colunas semi-enterradas

Abstract

Semi-embedded columns are structural elements with many day-a-day practical applications, mainly in foundation/structures of platforms, bridges, civil constructions, towers, etc. These elements have unstable behavior, requiring the development of reliable theoretical models for its structural project. The objective of this work is the optimal project of semi-embedded columns, whose decision variables are the dimensional parameters of the column. For this purpose is writing a multi-objective optimization problem that seeks to maximize the distance among the first three natural frequencies, at the same time is considering other conflicting objective function related to minimize the volume of the column. This optimization problem is subject to the constraints that limit the height of the foundation and verify if the critical load of Euler was obeyed. The dynamic model developed for the column uses the formulation of Lagrange to describe the behavior in a nonlinear soil. To solve the optimization problem two numerical methods are applied. These methods work with populations and are based on the natural process of evolution: the Differential Evolution method and Genetic Algorithm. The obtained results are compared and the methods performance analyzed by the following criteria: computational effort, algorithmic effort and execution time. Several numerical simulations checking the effectiveness of the developed methodology are presented. The optimal project of semi-embedded columns is calculated, resulting in a safe solution in terms of prevention of the dynamic oscillations and economic in terms of material cost, according to the conditions set by the designer.