面向乘波体气动外形设计的代理模型特性分析及优化  

Surrogate models characteristic analysis and optimization for waverider aerodynamic design

作　　者：季廷炜[1,2] 王崯瞩 张继发[1,2] 张帅[1,2] 

机构地区：[1]浙江大学航空航天学院,杭州310027 [2]浙江大学工程与科学计算中心,杭州310027

出　　处：《固体火箭技术》2015年第4期451-457,共7页Journal of Solid Rocket Technology

基　　金：国家自然科学基金(11102184)

摘　　要：代理模型在气动布局优化设计得到了广泛应用,研究探讨分析各代理模型的特性分析是重要的基础问题。文中基于3种Kriging和二阶多项式响应面模型等4种代理模型,以升阻比和容积率为样本,研究分析了其在乘波体优化设计中的的拟合精度和结构特性。研究表明,在高超乘波体优化设计中,基于自然指数型评估函数的Kriging模型具有相对突出的拟合精度。最后,在乘波体升阻力系数和容积率代理模型的多目标优化中,对比了基于自然指数型评估函数的Kriging模型和常用的基于高斯函数的Kriging模型的优化结果,发现基于自然指数型评估函数的Kriging模型对于峰值的拟合效果更好。The surrogate modes have been wildly implemented in aerodynamic shape optimization. It is a fundamental problem to analyze the characteristic of surrogate modes for aerodynamic shape optimization. Four different surrogate models including a quadratic response surface and three Kriging surrogates were constructed based on the values of lift-to-drag ratio and volumetric efficiency responses. The structure of the various surrogate modes was investigated and the main differences were addressed. The Kriging surrogate based on exponential correlations produced a relatively better overall performance. The multi-object evolutionary genetic algorithm was applied to obtain the optimum design based on the surrogates for these two objectives. The two Kriging surrogate based on Exponential correlations and Gaussian correlations were selected and used in the optimization process. The results shows that the Kriging surrogate based on exponential correlations produces a better result for peak value.

关 键 词：代理模型 气动布局 乘波体构型 飞行器概念设计 多目标优化 

分 类 号：V211.5[航空宇航科学与技术—航空宇航推进理论与工程]