基于代理模型的高超声速飞行器协同优化设计  

Collaborative optimization of hypersonic vehicle based on surrogate model

作　　者：马树微[1] 杜斌[1] 陈万春[1] 

机构地区：[1]北京航空航天大学宇航学院,北京100191

出　　处：《北京航空航天大学学报》2013年第8期1042-1047,共6页Journal of Beijing University of Aeronautics and Astronautics

基　　金：航天支撑技术基金资助项目(2011115155)

摘　　要：建立了类X-51A高超声速飞行器的多学科模型,包括外形、气动、推进、质量、热流、弹道等学科.采用多学科设计优化(MDO,Multidisciplinary Design Optimization)方法中的协同优化(CO,Collaborative Optimization)进行了计算,计算流程中构建了两个子系统优化器与弹道学科的代理模型,使用代理模型替代了计算耗时的真实模型.同时为了对比,采用了多学科可行法(MDF,Multidisciplinary Feasible Method)进行了优化计算.结果表明,二者的优化结果基本一致,优化之后的飞行器质量比初始构型飞行器降低了15.2%,航程提升了63.5%,并且使用代理模型后计算效率也得到了很大的提升.验证了基于代理模型的协同优化方法应用于高超声速飞行器设计中的有效性.The muhidisciplinary models of hypersonic vehicle which similar to the X-51A were built, including configuration, aerodynamics, propulsion, mass, heat and trajectory. The collaborative optimization （CO） method which is one of the muhidisciplinary design optimization （MDO） method was used, the surro- gate models of two subsystem optimizers and trajectory model were built to replace the real models which are time-consuming in the process. Meanwhile, multidisciplinary feasible （MDF） method was used to calculate for comparison. The results show that both the optimization results are basically the same, the mass of the ve- hicle is 15.2% lower and the range is 63.5% higher than the initial vehicle after optimization,and the compu- tational efficiency has also been greatly improved by using the surrogate models. The validity of the CO based on surrogate model applying to hypersonic vehicle design is verified.

关 键 词：高超声速飞行器 多学科设计优化 协同优化 代理模型 

分 类 号：V211.3[航空宇航科学与技术—航空宇航推进理论与工程]