基于混合自适应变异鸽群优化的飞机等效系统拟配  

Equivalent fitting technique for aircraft via mixed adaptive mutation pigeon-inspired optimization

作　　者：张兆宇 段海滨[1,2] 罗德林 ZHANG Zhaoyu;DUAN Haibin;LUO Delin(School of Automation Science and Electrical Engineering,Beihang University,Beijing 100083,China;Peng Cheng Laboratory,Shenzhen 518000,China;School of Aerospace Engineering,Xiamen University,Xiamen 361102,China)

机构地区：[1]北京航空航天大学自动化科学与电气工程学院,北京100083 [2]鹏城实验室,广东深圳518000 [3]厦门大学航空航天学院,福建厦门361102

出　　处：《厦门大学学报（自然科学版）》2022年第6期933-942,共10页Journal of Xiamen University：Natural Science

基　　金：科技创新2030-“新一代人工智能”重大项目(2018AAA0102403);国家自然科学基金(91948203,U20B2071,T2121003)。

摘　　要：现代高阶增稳飞机因模型阶次升高致使飞行品质难以分析,一般参考美军标提出的低阶等效系统(low-order equivalent system,LOES)概念以及一系列标准,在频域上辨识得到LOES后开展飞行品质分析,从而避免因模态耦合严重造成按传统方法无法分析计算的弊端.故本文提出了一种基于混合自适应变异机制的改进鸽群优化算法(mixed adaptive mutation pigeon-inspired optimization,MAMPIO),用于在频域上计算LOES.试验部分在飞机本体开环模型和增稳控制模型的纵向通道进行双拟配仿真计算,并同经典鸽群优化算法及用于高增稳飞机双拟配的混沌差分进化算法对比,通过绘制失配度收敛曲线及拟配频段的频率响应和频率失配曲线,证明在拟配频率段本文算法具有更高的拟配精度.在横航向通道上,由于LOES模型参数较多,使用MAMPIO在拟配精度方面会超出失配包络线.故针对横航向通道的等效拟配,提出一种先采用MAMPIO粗略搜索一组可靠的参数,再采用最小二乘法以MAMPIO的搜索结果作为迭代初值进一步获取LOES参数的拟配方法.包线内五组典型状态点处的仿真结果表明,此横航向拟配方法能够克服最小二乘法的初值敏感性,同时显著提升滚转、侧滑通道的拟配精度.Difficulties arise when the flight quality of modern high-ordered aircrafts with control augmentation systems(CAS)is analyzed due to the increase of the model order.The military standard of the United States proposes the concept of low-order equivalent system(LOES)and series of criterion,which requires engineers to perform LOES identification and evaluate the flight qualities.The drawback caused by strongly coupled flight models is avoided.Therefore,in this paper,a pigeon-inspired optimization method based on mixed adaptive mutation mechanism(MAMPIO)is proposed.This approach can handle equivalent fitting of LOES on frequency domains.Double-fitting simulations are carried out in the longitudinal channel of aircraft s open-loop model and closed-loop CAS model.The basic PIO and chaotic differential evolution algorithms are compared.The accuracy of the proposed method is demonstrated by analyzing evolutionary curves of mismatch function,frequency response curves and frequency response mismatching curves.In the lateral directional channel,fitting accuracy of MAMPIO exceeds mismatch envelope due to more unknown parameters.Therefore,a method is proposed,in which a reliable group of parameters is roughly searched by MAMPIO,then serves as the initial value of the least square method to obtain final parameters of LOES.Simulation results on five state points indicate that,the approach discussed above can overcome the sensitivity of initial value of the least square method.Moreover,the equivalent-fitting accuracy is elevated notably.

关 键 词：等效拟配 鸽群优化 自适应 变异 飞行品质 

分 类 号：V212.12[航空宇航科学与技术—航空宇航推进理论与工程]