基于AEKF的直升机在线系统辨识及仿真验证  

Online system identification and simulation validation of the helicopter based on the AEKF

作　　者：聂博文[1] 李国强[1,2] 靳秋硕[1] 张卫国 章贵川[1] NIE Bowen;LI Guoqiang;JIN Qiushuo;ZHANG Weiguo;ZHANG Guichuan(China Aerodynamics Research and Development Center,Mianyang 621000,China;National University of Defense Technology,Changsha 410073,China)

机构地区：[1]中国空气动力研究与发展中心,四川绵阳621000 [2]国防科技大学,湖南长沙410073

出　　处：《飞行力学》2024年第3期19-25,共7页Flight Dynamics

摘　　要：针对直升机飞行动力学建模面临桨叶动态挥舞、旋翼动态入流和旋翼/机身干扰等复杂难题而导致置信度不高的问题开展了研究。首先,针对常规布局直升机,建立了非线性动力学模型,获得了配平点处的线性化模型。然后,采用增广扩展卡尔曼滤波(AEKF)方法,通过多正弦正交激励,在线辨识获得了状态空间方程模型,初步验证了在线辨识方法的有效性。最后,在同一配平点处,向非线性模型和系统辨识模型施加相同的对偶方波激励,对比两种模型的响应历程。结果表明,基于AEKF的直升机在线辨识模型具有良好的置信度。It is difficult to model the flight dynamics of helicopter due to the challenges, such as blade dynamic flapping, dynamic inflow, and rotor/fuselage interference. As a result, the nonlinear model is always of low fidelity. Firstly, the nonlinear flight dynamics of a conventional helicopter is modeled, and a linearized model is obtained at the trim point. Then, a set of orthogonal multi sinusoidal excitations are applied to the nonlinear model and the method of augmented extended Kalman filter(AEKF) is implemented to identify the state space model, and the effectiveness of the online identification method is preliminarily verified. Finally, at a trimming point, a set of doublet excitations are applied to the nonlinear model and the identified model simultaneously, and the resulting time history of the flight states are compared with each other. The results show that the online identified model using the approach of AEKF is reliable.

关 键 词：飞行动力学 在线辨识 增广扩展卡尔曼滤波 飞行仿真 

分 类 号：V275.1[航空宇航科学与技术—飞行器设计]