基于BP神经网络的飞机性能计算

Aircraft Performance Calculation Based on BP Neural Network

针对人工查询性能图表进行飞机性能计算误差大、速度慢等问题,研究了使用BP神经网络进行飞机性能计算的可行性。首先介绍了飞机起飞爬升性能计算的理论模型;其次根据传统性能计算方法的局限性提出了用BP神经网络进行插值计算的思路,并以B737-800W起飞爬升限重为例,使用Matlab进行编程计算;最后将得到的计算结果与波音性能软件计算结果进行对比,误差在可以接受的范围内,证明将BP神经网络这一计算方法应用于飞机性能计算是可行的。

Compared with the traditional manually method of aircraft performance calculation,a new algorithm based on the BP neutral network is raised,which has better calculation efficiency and less error.First the theoretical model of takeoff climb limit calculation is introduced,then according to the limitations of traditional method of calculation raise that using BP neutral network to solve the problem. After that with the example of B737- 800 W takeoff climb limit calculation,using Matlab as a platform to prove it.Through the comparison of the results between BP neutral network and Boeing calculation software,we find that the error is acceptable. The results show that the BP neural network used in calculating the performance of the aircraft is feasible.