期刊
OPTIMIZATION AND ENGINEERING
卷 20, 期 1, 页码 251-275出版社
SPRINGER
DOI: 10.1007/s11081-018-9408-3
关键词
Topology optimization; Linear systems; Linear solvers; Approximate reanalysis; Nonlinear Programming; 90C30; 65K05; 49M37; 65F05; 15A23
类别
资金
- Fundacao de Amparo a Pesquisa do Estado de Sao Paulo [2013/05475-7, 2013/07375-0]
- Conselho Nacional de Desenvolvimento Cientifico e Tecnologico [302915/2016-8, 430678/2016-9]
A classical problem in topology optimization concerns the minimization of the compliance of a static structure, subject to a volume constraint upon the available material. Assuming that the structure is under small displacements and it is composed of a linear elastic material, the evaluation of the objective function demands the solution of a linear system. Hence, within the computational optimization process of addressing topology optimization problems, the cost of evaluating the objective function may be an issue, especially as the discretized mesh is refined. This work pursues the approximate reanalysis technique in combination with the Sequential Piecewise Linear Programming method for obtaining optimized structures. Numerical evidences are presented to corroborate the usage of this blend in a study composed by three distinct strategies in three benchmark test problems. A further analysis has been performed concerning the impact of the computation of the gradient vector of the objective function, pointing out room for additional savings.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据