版权所有:内蒙古大学图书馆 技术提供:维普资讯• 智图
内蒙古自治区呼和浩特市赛罕区大学西街235号 邮编: 010021
作者机构:Beihang Univ Sch Astronaut XueYuan Rd 37 Beijing 100191 Peoples R China
出 版 物:《STRUCTURAL AND MULTIDISCIPLINARY OPTIMIZATION》 (结构和多学科最佳设计)
年 卷 期:2019年第59卷第1期
页 面:1-19页
核心收录:
学科分类:08[工学] 0812[工学-计算机科学与技术(可授工学、理学学位)] 0801[工学-力学(可授工学、理学学位)] 080102[工学-固体力学]
基 金:National Natural Science Foundation of China
主 题:Stacking sequence optimization Blending design Multipoint approximation Composite structures
摘 要:The stacking sequence optimization problem for multi-region composite structures is studied in this work by considering both blending and design constraints. Starting from an initial stacking sequence design, unnecessary plies can be removed from this initial design and layer thicknesses of necessary plies are optimally determined. The existence of each ply is represented with discrete 0/1 variables and ply thicknesses are treated as continuous variables. A first-level approximate problem is constructed with branched multipoint approximate functions to replace the primal problem. To solve this approximate problem, genetic algorithm is firstly used to optimize discrete variables, and meanwhile, a blending design scheme is proposed to generate a blended structure. Starting from the thinnest region, this scheme shares all layers of current thinnest region with its adjacent regions. For non-shared layers in the adjacent regions, local mutation is implemented to add or delete plies to make them efficient designs. The whole process is repeated until the blending rule is satisfied. After that, a second-level approximate problem is built to optimize the continuous variables of ply thicknesses for retained layers. Those procedures are repeated until the optimal solution is obtained. Numerical applications, including a two-patch panel and a corrugated central cylinder in a satellite, are conducted to demonstrate the efficacy of the optimization strategy.