tailieunhanh - A method for selecting optimal sections for members of skeletal structures

This paper presents a simple method for selecting optimal sections for members of skeletal structures from an initially given set of sections. This is an extension of evolutionary structural optimization (ESO) to the sizing optimization problem with discrete design variables. | Vietnam Journal of Mechanics, NCST of Vietnam Vol. 23, 2001, No4 (234 - 246) A METHOD FOR SELECTING OPTIMAL SECTIONS FOR MEMBERS OF SKELETAL STRUCTURES CHU Due NHA Ministry of Education and Training ABSTRACT. This paper presents a simple method for selecting optimal sections for members of skeletal structures from an initially given set of sections. This is an extension of evolutionary structural optimization (ESO) to the sizing optimization problem with discrete design variables. Member sensitivity index for section sizing is derived from the optimality criterion. Optimization process is an iterative process of analysis, sensitivity calculation and section selection until optimality criterion is satisfied or constraints are violated. The proposed optimization procedure has been implemented into a structural analysis and optimization program called FEMOPT written on MATLAB programming language. Illustrative examples demonstrate the effectiveness of the proposed method. 1. Introduction In design of skeletal structures (trusses, frames), member sections must be selected from an initially given set of sections. This is clear for steel structures where sections have to be chosen from available range produced by manufacturers. Selecting optimal sections for members to minimize the weight (or cost) of a structure is always the desire of any designer and this can be achieved by employing a structural optimization method. However, due to complexity of structural optimization when dealing with real problems and not popularity of optimization software packages, the problem is usually solved by trying several sets of member sections and choose the best among them [1] . The obvious limitation of this "trial" method is not able to identify the optimal solution if it is not included in trial sets. Based on idea of Evolutionary Structural Optimization (ESO) for shape and topology optimization of plates under stress consideration proposed by Y. M. Xie and G. P. Steven [2], the