多子域模式重复结构多目标拓扑优化

将多子域模式重复方法引入到基于变密度法的传热以及力热一体化拓扑优化中.基本思想是将全域划分为多个大小相等的单胞,将设计变量定义在其中多个单胞上,并通过全等映射复制到其他单胞.引入了连续变化的密度过滤方法,以解决棋盘格问题和边界模糊问题.分别以散热弱度和局部温度作为目标函数,对传热结构进行了优化设计;引入多目标设计公式,对力热一体化结构进行了优化设计,均得到了清晰的模式重复结构.比较发现,多子域模式重复优化效果虽然不如全域设计,但要优于单域模式重复设计,而且继承了模式重复模块化特征显著和综合性能更均衡的优点.此外,单域和多子域模式重复结构均有更强的抗缺陷能力.

Multi-objective topology optimization of multi-subdomain pattern repetition

A new topological optimization based on variable density method and pattern repetition is applied to design thermal conductive structures and thermal-mechanics integration structures. The whole domain is divided into periodically arranged cells, and the design variables are defined only on several cells and are then mapped to the others. To eliminate checkerboard and vague boundary, a continuation change method is adopted for density filtering. Different objectives are used to design thermal conductive structures while multi-objective optimization is applied to design thermal-mechanics integration structures. All of them have clear pattern repetition structures. The comparison of structures shows that multi-subdomain pattern repetition is not as good as global design but better than single domain pattern repetition in optimization effect. Besides, multi-subdomain pattern repetition structures are more comprehensive in performance and have obvious modular characteristics. Furthermore, both multi-subdomain and single domain pattern repetition structures have stronger defect resistance than global designs.