基于约束边界抽样的起重机金属结构可靠性优化  

Reliability-based Design Optimization of Crane Metal Structure Based on Constraint Boundary Sampling

作　　者：范小宁[1] 王凯[1] 余畅 FAN Xiaoning;WANG Kai;YU Chang(Department of Mechanical Engineering,Taiyuan University of Science and Technology,Taiyuan 030024)

机构地区：[1]太原科技大学机械工程学院,太原030024

出　　处：《机械工程学报》2023年第8期288-298,共11页Journal of Mechanical Engineering

基　　金：山西省基础研究计划资助项目(20210302123212)。

摘　　要：针对全区域Kriging代理模型的起重机金属结构可靠性优化设计,为保证优化结果的正确性仍需较多高保真样本从而造成计算成本偏高、计算效率偏低的问题,提出了基于约束边界抽样的Kriging代理模型的起重机金属结构可靠性优化设计,根据抽样规则只在约束边界区域内抽取较多高保真样本,提高约束边界区域Kriging代理模型的拟合精度,减少构建非约束边界区域高精度代理模型所造成的时间消耗。通过实例验证:基于边界取样的Kriging代理模型的起重机金属结构可靠性优化与全约束域Kriging代理模型的可靠性优化设计相比,在取得同样优化结果的情况下,计算成本大约可节省78%,可见边界取样法在基于代理模型的可靠性优化设计中可以在保持良好准确性的同时大幅提高设计效率。The reliability-based design optimization(RBDO)for crane metal structures(CMSs)based on the Kriging surrogate models(KSMs)in the whole region,to ensure the correctness of the optimization results,a large number of high-fidelity samples are still required,which results in a high computational cost.A CMSs’RBDO based on KSMs of the constraint boundary sampling(CBS)is proposed.According to this sampling rule,only more high-fidelity samples are drawn in the constraint boundary region,which improves the fitting accuracy of the constraint boundary region and reduces the time consumption caused by constructing high-precision surrogate models in unconstrained boundary regions.Comparison studies with the RBDO for CMSs based on KSMs of sampling in the whole region indicated that the proposed approach could save the calculation cost by about 78%under the same optimization results.The CBS method could enhance the design efficiency greatly,meanwhile maintaining good accuracy in the RBDO based on KSMs.

关 键 词：起重机金属结构 Kriging代理模型 约束边界抽样法 可靠性优化设计 

分 类 号：TH215[机械工程—机械制造及自动化]