摘要
针对螺栓连接的加筋板屈曲性能优化建模复杂和计算量较大的问题,提出了一种基于Kriging代理模型和多点约束(multi-point constraint,MPC)近似模型的混合优化策略.首先,运用MPC连接建立近似加筋板模型进行有限元屈曲分析,代替试验设计中大量的高精度加筋板模型分析.然后,通过Kriging代理模型建立MPC参数的预测函数,并在优化迭代中更新代理模型样本点,保证近似模型的计算精度.最后,基于建立的MPC近似模型,对螺栓连接加筋板进行了轻量化设计和性能最优化设计.算例结果表明:该文提出的混合优化方法与传统优化方法相比,其计算效率提高了10倍左右.在轻量化设计中,使加筋板结构在保证屈曲承载力不变的情况下减重了26.18%.在性能最优化设计中,在质量无明显变化的情况下使极限屈曲承载力提高了23.67%.
To address the modeling complexity and large computational load in optimizing the buckling performance of bolted stiffened panels,a hybrid optimization strategy based on the Kriging surrogate model and the multi-point constraint(MPC)approximation model was proposed.Firstly,the MPC connection was utilized to establish an approximate stiffened panel model for the finite element buckling analysis,to replace the analysis of many high-precision stiffened panel models in the experimental design.Then,a prediction function for MPC parameters was built with the Kriging surrogate model,and the sample points of the surrogate model were updated during optimization iterations to ensure the computational accuracy of the approximation model.Finally,based on the established MPC approximation model,the lightweight design and the performance optimization design of bolted stiffened panels were conducted.The numerical results demonstrate that,the proposed hybrid optimization method improves the computational efficiency by approximate 10 times compared with traditional optimization methods.In the lightweight design,the weight of the stiffened panel structure reduces by 26.18%while maintaining the same buckling capacity.In the performance optimization design,the ultimate buckling capacity increases by 23.67%without significant change in the structural mass.
作者
夏甘霖
孟增
吴枝根
XIA Ganlin;MENG Zeng;WU Zhigen(College of Civil Engineering,Hefei University of Technology,Hefei 230009,P.R.China)
出处
《应用数学和力学》
北大核心
2025年第9期1108-1118,共11页
Applied Mathematics and Mechanics
基金
国家自然科学基金(面上项目)(12372195,11572108)。
关键词
加筋板结构
屈曲分析
代理模型
结构优化设计
stiffened panel
buckling analysis
surrogate model
structural optimization design
