摘要
将多载荷作用下的连续体结构拓扑优化设计看作为一种对单元样本的模式识别,然后利用模式识别的K邻近(KNN)方法实现连续体结构的拓扑优化设计。具体过程是以设计区域全体单元为样本空间,利用单元在各载荷作用下的应力分量作为单元样本的特征向量,并以欧氏距离作为识别标准将单元分为两类,最终通过逐步删除低应力贡献度的结构单元来达到结构优化。最后还给出了基于KNN方法的多载荷作用下连续体结构拓扑优化设计示例,结果表明此设计方法是完全正确且可行的。
In this paper the topology optimization design of multi-loaded continuum structure is taken as a kind of structural unit pattern reeognition, and K Nearest Neighbor （KNN） method of pattern recognition is used to conduct the topology optimization design. The entire design region is taken as the initial sample space, and the element stresses under multi-load are regarded as the feature vectors, and the feature vectors＇ Euclidean distance is regarded as the recognition standard to classify all the unit samples into two classes so as to obtain the topology optimization result. Several examples of the topology optimization design are given and made based on KNN method. The results indicate that this design method is feasible.
出处
《机械科学与技术》
CSCD
北大核心
2012年第6期873-876,共4页
Mechanical Science and Technology for Aerospace Engineering
基金
长江学者和创新团队发展计划(IRT0816)
"高档数控机床与基础制造装备"科技重大专项课题(2010ZX04014-014)
国家自然科学基金项目(50875039)
"十一五"国家科技支撑计划项目(2009BAG12A02-A07-2)资助
关键词
结构优化
拓扑优化
K邻近
多载荷
连续体
structure optimization
topology optimization
K-nearest neighbor
multi-load
continuum structure
