均质化(气候)
计算
参数化复杂度
多边形网格
数学
应用数学
数学优化
刚度
柯西分布
有限元法
计算机科学
算法
数学分析
材料科学
几何学
物理
生物多样性
热力学
复合材料
生物
生态学
作者
Haidong Lin,Shujuan Hou
标识
DOI:10.1016/j.cma.2023.116010
摘要
Recently developed cross-scale optimization methods are mainly the macro equivalent calculation of performances based on the traditional homogenization method. However, the traditional homogenization is limited to the classic continuum of Cauchy–Boltzmann. Therefore, it is inadequate to interpret the size dependence of the optimal result. Hence, a new cross-scale optimization is proposed based on Wei–Hutchinson strain gradient theory by employing the non-local homogenization model, which could describe and explain the size dependence during optimization process when considering micro structures. The topological optimization procedure simultaneously has the ability of coupled computing by using subdomain parameterized coarse meshes. The numerical computations involved in the entire model can be solved in one iteration, which helps to eliminate mesh dependencies and greatly reduce the computation time. It is shown that the final stiffness of the optimized periodic structure can be significantly increased by considering the strain gradient theory compared with the classic homogenization scheme in the process of cross-scale optimization.
科研通智能强力驱动
Strongly Powered by AbleSci AI