LatticeOPT: a heuristic topology optimization framework for thin-walled, 2D extruded lattices
LatticeOPT: a heuristic topology optimization framework for thin-walled, 2D extruded lattices
复制标题
DOI：
10.1007/s00158-022-03397-5
复制
发表时间：
2022-05
影响因子：
3.9
通讯作者：
Junyan He;Shashank Kushwaha;D. Abueidda;I. Jasiuk
中科院分区：
文献类型：
--
作者：
Junyan He;Shashank Kushwaha;D. Abueidda;I. Jasiuk
本文介绍了一个纯粹的启发式拓扑优化框架，以提高比能量吸收薄壁，挤压网格结构。该框架通过迭代更新网格壁厚来优化网格截面设计。工作的主要新奇是我们提出的两个新的厚度更新计划。第一个更新方案是壁向比能量吸收的均匀化的直接陈述，而另一个方案是基于双向渐进结构优化方法启发的壁向灵敏度参数的均匀化。这两种方案的基础上的中心思想，均匀化的某些领域的变量，这已被广泛采用在以前的薄壁结构的优化框架。所提出的框架具有很高的潜力，因为它可以直接与商业有限元软件包，只需要每个元素的能量吸收的信息。不需要有限元刚度矩阵，该框架可以用于显式动力学模拟来处理高度非线性问题。给出了三个数值算例：（1）轴压柱的优化，（2）动态三点弯曲下的网格加强梁的优化，和（3）爆炸载荷下的网格填充夹层板的优化。结果表明，该框架可以有效地提高比能量吸收，只需25次非线性有限元模拟。图形摘要
This paper introduces a purely heuristic topology optimization framework to improve specific energy absorption for thin-walled, extruded lattice structures. The framework optimizes the lattice cross section design by iteratively updating the lattice wall thicknesses. The main novelty of the work is the two novel thickness update schemes we proposed. The first update scheme is a direct statement of homogenization of wall-wise specific energy absorption, while the other scheme is based on the homogenization of a wall-wise sensitivity parameter inspired by the bi-directional evolutionary structural optimization method. Both schemes are based on the central idea of homogenization of certain field variables, which has been widely employed in previous optimization frameworks for thin-walled structures. The proposed framework has high potential because it can work directly with commercial finite element packages, and only requires information on the energy absorption of each element. Without the need for the finite element stiffness matrix, this framework can be used with explicit dynamics simulations to treat highly nonlinear problems. Three numerical examples are presented: (1) optimization of a column under axial compression, (2) optimization of a lattice-reinforced beam under dynamic three-point bending, and (3) optimization of a lattice-filled sandwich panel under blast loading. The results show that the framework can effectively increase specific energy absorption with as few as 25 nonlinear finite element simulations.Graphical abstract