基于神经网络的高原机场飞机离地速度计算方法  

Aircraft Lift off Speed Calculation Method at Plateau Airports Basedon Neural Network

作　　者：陈肯[1] 李梁昆 CHEN Ken;LI Liang-kun(Civil Aviation Flight University of China,Guanghan 618000,China)

机构地区：[1]中国民用航空飞行学院,四川广汉618000

出　　处：《航空计算技术》2024年第1期1-5,共5页Aeronautical Computing Technique

基　　金：国家自然科学基金资助(2021YFF0603904)。

摘　　要：针对现有飞行理论中对民航飞机在高原机场起飞离地速度计算研究不足,且计算方法单一、复杂的问题,利用MATLAB仿真平台构建在高原机场飞机起飞滑跑阶段模型,通过计算不同飞行条件下的飞机离地速度,得到大量训练样本,建立BP神经网络模型,将机场海拔、飞机质量、安全起飞速度、温度和可用起飞距离作为输入量,飞机离地速度作为输出量,对神经网络进行训练和测试,直至均方差达到要求。将神经网络所求估算值与仿真数据进行对比并进行分析。均方误差为4.1289,均方根误差为5.6121。使用神经网络计算方法效率高,误差小,可应用于不同高原机型在不同影响因素下的离地速度计算,对高原机场运行和机场跑道长度设计有一定参考价值。To address the shortage of existing flight theories on the calculation of lift off speed of civil aircraft at plateau airports,and the problem of single and complex calculation methods.By using MATLAB simulation platform to built a model of aircraft take off run stage at plateau airport,a large number of training samples were obtained by calculating the lift off speed under different flight conditions,established a BP neural network model,take airport altitude,aircraft weight,safe take off speed,temperature and available take off distance as input quantities and aircraft lift off speed as output quantity.The neural network was Trained and tested until the mean square deviation reaches the requirement.Compared and analyzed the estimated values obtained from the neural network with the simulation data.The mean square error is 4.1289 and the root mean square error is 5.6121.The neural network calculation method is efficient,with small error,and can be applied to the calculation of lift off speed for different plateau aircraft types under different influencing factors,which has some reference value for plateau airport operation and calculation of runway length determination.

关 键 词：高原机场 飞机性能 离地速度 神经网络 

分 类 号：V212[航空宇航科学与技术—航空宇航推进理论与工程]