飞行过载与飞行姿态半实物仿真技术研究

发布时间：2018-10-16 16:53

【摘要】：本文围绕飞行过载与飞行姿态仿真的技术问题,将三轴飞行模拟转台与离心机结合,提出一种新型飞行过载模拟器。并且对飞行过载与飞行姿态耦合后进行了飞行过载模拟器的动力学分析,然后建立了飞行过载模拟器的动力学模型,并且分析了飞行过载模拟器的转矩耦合问题及验证解耦的可行性和有效性,最后将飞行过载与飞行姿态耦合后的模型进行了仿真,并且对仿真结果进行了分析。论文主要工作如下:首先,介绍了本论文的课题来源和研究背景,提出了本文所研究问题的必要性和意义。具体介绍了国内外三轴飞行转台和离心机的研究成果,给出了飞行过载模拟器的研究综述。其次,介绍了飞行过载模拟器的系统组成及安装设计。指出三轴飞行模拟转台的功能是根据计算机的指令,实时跟随指令信号,通过控制转台滚动、俯仰、航向轴系运动,来模拟姿态角变化,达到对飞行器的姿态仿真的作用。并且进一步对其驱动装置的选择和结构设计进行了分析。介绍了离心机的工作原理。介绍了两种飞行过载模拟器,给出了结构图,并做出了简单的介绍。搭建出飞行过载模拟器的结构以后,就可以对其进行动力学分析。再次,主要进行飞行过载和飞行姿态的动力学分析,关键问题在于搭建在过载情况下飞行姿态仿真的模型。核心部分是搭建飞行过载模拟器的动力学方程。对三轴飞行模拟转台的动力学进行了分析,其中包括坐标变换,转矩方程,环轴间传递的耦合转矩方程的建立。对三轴飞行模拟转台的内框架、中框架及外框架控制模型进行建模,得到三个框架的动力学方程。对飞行过载模拟器的摩擦转矩进行分析,找到由重力及离心力产生的摩擦转矩的方程,最终得到飞行过载模拟器的动力学方程。还有,应用NARX神经网络解决了飞行过载模拟器轴间的转矩耦合。给出了飞行过载模拟器的一般性动力学模型,并且验证了模型的可逆性。这是使用NARX神经网络解耦控制的先决条件。IMC策略介绍如何实现解耦控制,在NARX神经网络中用于识别飞行过载模拟器的逆模型。提出了一个合适的实时学习校正算法训练的神经网络,保证了网络的收敛性。提出的解耦控制理论是对飞行过载模拟器的应用程序进行测试。应用程序演示的结果证明了NARX神经网络解耦控制在飞行过载模拟器三个轴解耦控制上的有效性。最后,完成了对不同过载情况下飞行姿态进行仿真实验的研究,并且对仿真图形进行了分析和对比,得出了过载量越大导致姿态角越小的结论。
[Abstract]:Based on the technical problems of flight overload and attitude simulation, a new flight overload simulator is proposed by combining the three-axis flight simulation turntable with the centrifuge. After the coupling of flight overload and attitude, the dynamic analysis of flight overload simulator is carried out, and then the dynamic model of flight overload simulator is established. The torque coupling problem of flight overload simulator and the feasibility and effectiveness of decoupling are analyzed. Finally, the model of flight overload and attitude coupling is simulated, and the simulation results are analyzed. The main work of this paper is as follows: firstly, the source and background of this thesis are introduced, and the necessity and significance of the research are put forward. This paper introduces the research results of three-axis flight turntable and centrifuge at home and abroad, and gives a summary of the research of flight overload simulator. Secondly, the system composition and installation design of flight overload simulator are introduced. It is pointed out that the function of the three-axis flight simulation turntable is to follow the command signal in real time according to the instruction of the computer, and to simulate the attitude angle change by controlling the rolling, pitching and heading shafting motion of the turntable, so as to achieve the function of the attitude simulation of the aircraft. Furthermore, the selection and structure design of the drive device are analyzed. The working principle of centrifuge is introduced. This paper introduces two kinds of flight overload simulators, gives the structure diagram, and makes a brief introduction. After the structure of the flight overload simulator is built, the dynamic analysis can be carried out. Thirdly, the dynamic analysis of flight overload and attitude is mainly carried out, and the key problem is to build the model of flight attitude simulation under overload. The core part is to build the dynamic equation of flight overload simulator. The dynamics of three-axis flight simulation turntable is analyzed, including coordinate transformation, torque equation and coupling torque equation between ring axes. The internal frame, middle frame and outer frame control model of the three-axis flight simulation turntable are modeled, and the dynamic equations of the three frames are obtained. The friction torque of flight overload simulator is analyzed and the equation of friction torque generated by gravity and centrifugal force is found. Finally the dynamic equation of flight overload simulator is obtained. In addition, NARX neural network is used to solve the torque coupling between the axes of flight overload simulator. The general dynamic model of flight overload simulator is given, and the reversibility of the model is verified. This is a prerequisite for decoupling control using NARX neural network. IMC strategy introduces how to realize decoupling control and use it to identify the inverse model of flight overload simulator in NARX neural network. A neural network trained by an appropriate real-time learning correction algorithm is proposed to ensure the convergence of the network. The decoupling control theory is to test the application program of flight overload simulator. The results of the application program demonstrate the effectiveness of the decoupling control of NARX neural network in the three-axis decoupling control of the flight overload simulator. Finally, the simulation experiment of flight attitude under different overload conditions is completed, and the simulation figure is analyzed and compared, and the conclusion is drawn that the larger the overload is, the smaller the attitude angle is.
【学位授予单位】：北京理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：V211

【相似文献】