节点文献
直升机飞行力学模型辨识研究
Research on Helicopter Flight Dynamics Model Identification
【作者】 孙涛;
【导师】 张呈林;
【作者基本信息】 南京航空航天大学 , 飞行器设计, 2010, 博士
【摘要】 直升机飞行力学建模研究是直升机飞行控制系统、飞行仿真系统研制的基础。本文系统研究了直升机飞行力学模型的系统辨识建模理论及实现方法。论文围绕着直升机辨识中的先验知识获取和采用、辨识方法及辨识实现手段几个主要部分开展研究。以单旋翼带尾桨的小型无人直升机为主要研究对象,并致力于将研究方法和结论在向直升机系统辨识所存在的共性问题推广。研究中注重引入经验及理论分析所提供的先验知识,以使结果在能够正确描述直升机飞行力学特性的同时具有明确的物理意义及良好的适用性,能够更容易的为工程技术人员所接受和采用。文章首先采用机理建模的方法分析了直升机的飞行力学特性并得到了辨识模型集。分析过程中根据直升机主要气动部件所受气动力,建立了直升机的刚体6自由度运动方程。在此基础上,又研究了旋翼挥舞运动及小型直升机贝尔-希勒伺服小翼的挥舞运动方程并分析了其对直升机响应的影响。通过对直升机飞行力学模型特性的分析,获得了辨识的先验知识。针对直升机是一个非线性不稳定多输入-多输出系统及机体振动大的特点,研究了直升机飞行力学模型辨识方法。提出了首先采用子空间方法获得初始模型,然后采用频域加权的预报误差法获得最优模型的两步辨识方法。以本方法为主要工具,论文对算例直升机模型辨识进行了仿真,结果表明本方法是有效的。为了开展辨识试飞试验及研究辨识试验相关技术,本文开发了一套直升机飞行力学模型辨识试验系统。根据辨识试验开展的需要设计了各子系统的软硬件。同时,为了满足辨识的要求研究了系统的导航滤波算法,提出了适用于直升机飞行数据测量的多传感器数据融合方法,并对系统特性进行了分析。以本文所开发的辨识系统为基础,进一步研究了辨识试验中可能遇到的关键技术问题。最后,为了检验本文所提出的辨识方法以及所开发的辨识试验系统的有效性,以航模直升机为辨识对象开展了辨识试验,分别辨识了对象直升机的6自由度和8自由度线化飞行力学模型。采用未参与辨识计算的试飞数据验证了辨识结果的有效性,并将获得的两模型进行了对比。为了进一步验证辨识所得模型的准确性及实用性,利用辨识所得模型设计了多环PID控制器并通过试飞对其有效性进行了验证。
【Abstract】 Modeling of helicopter flight dynamics is the foundation of helicopter flight controller and flight simulator development. This thesis systemically studies theories and methods of helicopter flight dynamics model identification. And concentrates on three correlative contents: transcendental information acquisition and adoption, identification methods and identification flight experiment technology. In the thesis, the analysis and study are mainly based on small-scaled single main rotor helicopter, and some conclusions are extended to the normal helicopter to solve the common problems on helicopters. This thesis pays attention to the utilization of prior knowledge, such as engineering experience and theoretical conclusions, so as to make the results capable of describe the helicopter flight dynamic natures correctly as well as of clear physical meaning and easy to be accepted by engineers.Firstly, in order to get the identification model group, the thesis researches the characters of helicopter flight dynamics and modeles helicopter flight dynamics using mechanism modeling method. The force and moments of the main aerodynamic components are considered. With these forces and moments the thesis built up the 6-DOF movement equations of the helicopter. Based on the 6-DOF equations the flapping motions of main rotor and Bell-Hiller bar are also modeled to analyse the influence to helicopter response. Through the analysis on helicopter flight dynamic characters and mechanism model the transcendental information is acquired, which is essential for the identification research in the next step.The helicopter is a multi-input multi-output non-linear unstable system with very high vibration. The characters make the identification of helicopter very difficult, since there are many problems during identification modeling. Aiming at solving these problems, the system identification method is researched in this thesis. A two step identification method is proposed, which firstly identify the initial model using subspace identification method and then identify the optimized model using frequency weighted prediction error method. With this identification method this thesis makes the simulation to identify the example helicopter flight dynamics model. The result shows that the method is efficient and accurate.In this thesis a helicopter flight dynamic identification experiment system is develop in order to carry out the flight experiments and research the problems encountered. According to the requirements of the experiment the subsystems are designed. Meanwhile, the navigation algorithm is researched and a multi-sensor data fusion method is proposed. Based on the system designed in the thesis, the common problems in experiment are also researched.At the end of the thesis, in order to verify the identification method and the experiment system the flight experiment is carried out with model-scaled helicopter. The 6-DOF and 8-DOF linear flight dynamics model of the helicopter are identified. The identification results are verified with the flight data which is not used in the identification calculation. Aiming at verifying the identified model further, the thesis designed the flight controller using identified result, and carried out flight test to verify the validity of the controller.
【Key words】 helicopter; flight dynamics; system identification; subspace method; prediction error method; frequency domain weighting;