欢迎光临《计算力学学报》官方网站！

ESO method based on a kind of dynamic deletion rate

DOI：10.7511/jslx201502022

 作者 单位 E-mail 罗静 重庆大学 机械工程学院, 重庆 400044 luojjune@163.com 张大可 重庆大学 机械工程学院, 重庆 400044 李海军 重庆大学 机械工程学院, 重庆 400044 龚姣 重庆大学 机械工程学院, 重庆 400044

渐进结构优化方法ESO(Evolutionary Structural Optimization)的基本思想是基于单元灵敏度,通过把无效或低效的单元逐步从结构中删除,从而得到优化的拓扑结构。经过20年的发展,其算法结构、理论研究以及实际工程应用领域已经取得了大量成果。在原始ESO算法中删除单元的数目是由固定删除率RR的取值决定的,设计者无法预期每一个迭代步删除单元的数量。对于一个初始满设计区域,合理的删除策略为随着迭代的进行,随着结构应力分布越均布,应该逐渐减少单元删除数量。基于此,本文构造了一种动态删除率,使得随着迭代的进行单元的删除数量逐渐减少,相对于前人构造的动态删除率更为简单明了,人为控制参数更少,并且通过算例证明该方法相比原始删除策略具有更好的优化效果。

The main idea of the evolutionary structural optimization ESO method is that the invalid element gradually be removed from the structure based on the element sensitivity, and so the topology optimization result is formed.The element deletion number in the original ESO algorithm is decided by the value of rejection ration(RR), and the designer cannot expect the element deletion number every iteration.For an initial full design area, a reasonable strategy for element deletion should gradually reduce the number of element deletion as the iteration proceeds, while the structural stress distribution is more uniform.Based on this, this paper constructs a dynamic deletion rate, the element deletion number gradually reduce as the iteration proceeds.It is more simple, and has less controlled parameters relative to the previous dynamic deletion rates.Through the examples it shows that the method can get slightly a better optimization effect compared with the original deletion strategies.