Mohammad Reza Ghasemi; Akbar Azadi
Abstract
One of the major purposes of optimization in civil engineering is to perform a suitable design for the structure. This goal has to fulfill technical criteria and contain the minimum economical costs. Building frames are of the most customary civil engineering structures. Therefore, optimization of these ...
Read More
One of the major purposes of optimization in civil engineering is to perform a suitable design for the structure. This goal has to fulfill technical criteria and contain the minimum economical costs. Building frames are of the most customary civil engineering structures. Therefore, optimization of these types of structures could be of a great concern from the economical viewpoints. One of the current obstacles in such optimization problems is the local convergence debility. Thus, using means of tackling this problem seems necessary. Genetic Algorithm which is one of the optimization methods inspired by nature, has overcome this problem. In order to solve such problems, genetic algorithm needs a multiple analyses of structures. Therefore, in this study attempts were made to introduce and embed new formulae into a newly developed program to handle new techniques for selection and mutation as genetic operations. As for the aspects of application, the introduced techniques were inspected and investigated in the optimization of some planar and special braced steel frames. The outcome through comparisons proclaimed a considerable decrease in numbers of analyses as well as significant increases in the speed of convergence.
Mohammad Reza Ghasemi; Mohammad Reza Mostakhdemin Hosseini
Abstract
Due to the probabilistic nature and uncertainties of structural parameters, reliability-based optimization will enable engineers to account for the safety of the structures and allow for its decision making applicability. Thus, reliability-based design will substitute deterministic rules of codes of ...
Read More
Due to the probabilistic nature and uncertainties of structural parameters, reliability-based optimization will enable engineers to account for the safety of the structures and allow for its decision making applicability. Thus, reliability-based design will substitute deterministic rules of codes of practice. Space structures are of those types that have an exceedingly high range of applicability aspects in civil engineering. Therefore optimization of such structures with great and considerable number of members will be economically wise. For this purpose, the optimization process could be carried out using various mathematical models. One such model is to minimize weight while considering elements failure probability as constraints. Another form is to minimize weight and then regarding the whole structure system reliability as constraint. The third type could be to minimize failure probability as well as its weight, while taking into account the structural system reliability as the constraint. In this research each of the above forms were studied and the results were compared. Also, apart from reliability considerations for the members, the reliability of nodes was also taken into account. Node failure means that node displacement in at least one direction exceeds that of the allowable value. As well the effect of various stochastic parameters such as load, yield stress, modulus of elasticity and cross section were studied. The stochastic parameters discussed in this study are statistically independent and possess standardized normal distribution. To avoid local convergence during the process of optimization, Genetic Algorithms is used as means of optimization. This study show that with increasing the members or system admissible failure probability, optimum weight of structure increases, but with increasing the coefficient of variation of load or yield stress, optimum weight increases.