利用聚类分析的宏细观多尺度并行拓扑优化设计方法研究

Investigation of Macro and Micro Multiscale Concurrent Topology Optimization Design with Cluster Analysis

多孔材料因为具有质量轻、比刚度和比强度大、隔振和隔热效果好等优点,越来越多的应用在航空航天和制造装备等领域。为充分发挥多孔材料的性能,本文提出同时考虑结构的宏观性能和细观微结构性能的多尺度并行拓扑优化设计方法,获得性能优良的多孔结构。论文采用聚类方法有效降低了计算成本,针对并行优化难收敛的问题提出改进模型,使迭代平稳收敛。最后以经典的悬臂梁、MBB梁和Michell结构为例进行优化设计,通过对宏细观多尺度并行优化结果的分析,验证了所提方法的有效性与正确性。

Porous structure is widely used in aerospace and manufacturing equipment field because porous structure usually has light weight,high specific stiffness and strength as well as excellent vibration isolation and thermal insulation performance.In order to get an excellent performance of porous materials,this paper proposes a multi-scale parallel topology optimization design method which considers simultaneously both the macroscopic and microstructure properties of the structure to obtain a porous structure with excellent performance.The clustering method is performed in this paper to effectively reduce the computational cost,and an improved model is proposed to solve the problem that is difficult to converge in parallel optimization,so that the iteration is smoothly closed.Finally,the classic cantilever beam,MBB beam and Michell structure are used as examples to assess the proposed method.The effectiveness and correctness of the proposed method are verified via comparing the optimization results.