多体分离气动特性智能预测技术研究  

Investigation on intelligent prediction technology of aerodynamic characteristics of multi-body separation

作　　者：姜金俊[1] 陈连忠[1] Jiang Jinjun;Chen Lianzhong(China Academy of Aerospace Aerodynamics,Beijing 100074,China)

机构地区：[1]中国航天空气动力技术研究院,北京100074

出　　处：《空天技术》2024年第3期24-32,共9页Aerospace Technology

摘　　要：下一代超声速或高超声速战斗机内埋武器系统可能会采用后射分离(Aft Ejection Separation,AES)概念,后射分离属于飞行器多体分离一种类型,不仅存在多体飞行器间的流动与气动干扰,而且存在多体飞行器间的相对运动,严重影响载机的安全性。开发了一种内置式风洞试验系统并在FD-12风洞中集成,获取了典型工况下的后射分离试验结果,并基于综合随机决策树的非线性模型回归技术的人工智能算法对其余攻角α_(A)、侧滑角β_(A)下后射分离气动特性进行初步智能建模预测研究,给出将近100次随机状态下试验值与预测值对比,并用来预测给定其余典型状态下后射分离气动数据。结果表明,基于综合随机决策树的非线性数学模型的回归技术优化算法对后射分离气动干扰试验,在马赫数Ma=2.0,攻角-5°≤α_(A)≤5°,侧滑角-3°≤β_(A)≤8°的样本空间内具有一定的预测性。The next-generation supersonic or hypersonic fighter aircraft may use the concept of aft ejection separation(AES)for the internal weapon system.The AES belongs to a type or form of multi-body separation,not only has the flow and aerodynamic interference between the multi-body aircraft,but also has the relative motion between the multi-body aircraft,which seriously affects the separation safety of the aircraft.In this paper,an internal wind tunnel captive trajectory system is developed and integrated in the wind tunnel to obtain the aerodynamic and motion variation laws under typical conditions,and the artificial intelligence algorithm based on comprehensive random decision tree is used to preliminarily predict the aerodynamics of the aft ejection separation of the internal weapon.The results show that the optimization algorithm of nonlinear model regression technology based on comprehensive random decision tree has a certain predictability in the sample space of Mach numberMa=2.0,angle-of-attack(-5°≤α_(A)≤5°)and sideslip angle(-3°≤β≤8°)for tail separation test.

关 键 词：内埋武器 后射分离 气动干扰 人工智能 综合随机决策树 

分 类 号：V211.7[航空宇航科学与技术—航空宇航推进理论与工程] V212.1