飞行器总体不确定性建模与优化设计方法  

Uncertainty Modeling and Optimization Method for Overall Design of Flight Vehicle

作　　者：张海瑞[1] 王浩[1] 王尧 洪东跑[1] ZHANG Hairui;WANG Hao;WANG Yao;HONG Dongpao(China Academy of Launch Vehicle Technology,Beijing 100076,China)

机构地区：[1]中国运载火箭技术研究院,北京100076

出　　处：《宇航学报》2023年第4期486-495,共10页Journal of Astronautics

基　　金：国家自然科学基金(U20B2028);技术基础课题(JSZL2020203B001)。

摘　　要：针对考虑不确定性的飞行器总体设计迭代周期长、优化困难等问题,提出一种基于裕度量化解耦策略的飞行器总体不确定性建模与优化设计方法。首先采用主动学习策略开展基于优化加点Kriging的阈值不确定性分析,实现了不确定性裕度的高效求解。进而提出一种适用于飞行器总体设计的裕度量化解耦策略,将双层嵌套的不确定性优化问题解耦为确定性优化与不确定性分析过程顺序执行,高效给出满足概率约束的优化设计方案。该方法解决了传统解耦方法采用嵌入式算法而导致的工程应用困难的问题。通过非线性多约束数值案例以及滑翔飞行器工程案例,验证了该方法能够在保证精度的前提下,提高不确定性优化设计效率。Aiming at the problems such as long iteration period and difficult optimization in the overall design of flight vehicle considering uncertainty,an uncertainty modeling and optimization method for flight vehicle overall design based on margin quantification and decoupling strategy is proposed.Firstly,a Kriging-based active learning strategy is used to carry out the threshold uncertainty analysis efficiently.Then,a margin quantification and decoupling strategy for the overall design of flight vehicle is proposed.The two-layer nested uncertainty optimization problem is decoupled into the sequence execution of deterministic optimization and uncertainty analysis processes,and an efficient optimization design scheme satisfying probability constraints is given.This method solves the problem of engineering application difficulty caused by traditional decoupling method using embedded algorithm.The results of numerical and engineering cases show that the new method can improve the efficiency of uncertain optimal design on the premise of ensuring the accuracy.

关 键 词：飞行器总体设计 不确定性优化设计 KRIGING模型 主动学习策略 

分 类 号：V421.1[航空宇航科学与技术—飞行器设计]