亚轨道飞行器的固液火箭发动机的高维多目标设计优化  

High-Dimension Multi-Objective Design Optimization of Hybrid Rocket Motor for Sub-Orbit Flight Vehicle

作　　者：孙兴亮[1] 朱浩[1] 田辉[1] 俞南嘉[1] 蔡国飙[1] 

机构地区：[1]北京航空航天大学宇航学院,北京100191

出　　处：《宇航学报》2014年第11期1291-1298,共8页Journal of Astronautics

摘　　要：为完成应用在亚轨道飞行器上的固液火箭发动机系统设计优化,且考虑到系统设计优化问题的复杂性以及采用单目标优化难以获得全局最优解,建立了固液火箭发动机系统设计的4目标设计优化模型,采用非支配排序遗传算法获得超空间Pareto解集。利用基于变偏好区间的超径向可视化技术对超空间Pareto解集进行直观地分析,提高最优解选择的效率。最终获得了全局最优的固液火箭发动机系统设计方案,验证了建立的优化模型的准确性和相比于单目标优化的优势,论证了固液火箭发动机应用在亚轨道飞行器上切实可行以及变偏好区间超径向可视化技术处理超空间Pareto解集的高效性。Considering the complexity of system design optimization and the difficulty of applying the single objective optimization to find a global optimization solution,a four objective optimization model is established to complete the optimal design of the hybrid rocket motor for the sub-orbit flight vehicle. First,the hyperspace Pareto set is generated by using the non-dominated sorting genetic algorithm. Second,the hyper-radial visualization based on the variable preference range is applied to deal with the hyperspace Pareto set. Finally,the global optimization design of the hybrid rocket motor for the sub-orbit flight vehicle is found,the possibility of powering the sub-orbit flight vehicle with hybrid rocket motor is proved.The result indicates that the high-dimension multi-objective optimization model is accurate and the hyper-radial visualization based on variable preference rang is efficient.

关 键 词：亚轨道飞行器 固液火箭发动机 高维多目标优化 Pareto解集可视化 非支配排序遗传算法 

分 类 号：V436[航空宇航科学与技术—航空宇航推进理论与工程]