摘要
适航噪声分析是民用直升机研发的一个环节,根据直升机的设计参数,建立三种适航状态的分析模型,获得直升机的噪声水平。文章在提炼直升机噪声的主要影响参数的基础上,利用神经网络适用于非线性预估的特点,采用美国联邦航空局(FAA)的资讯通报发布的直升机适航噪声飞行试验数据,进行神经网络的设计和训练,建立了基于神经网络的直升机适航噪声分析方法,在保证分析误差的基础上,实现了直升机适航噪声的预估。
The prediction of helicopter certificated noise is one step of a civil helicopter development ,which needs to construct models of three flight conditions required by helicopter noise certification ,to predict helicopter noise levels, using the helicopter parameters as input. Based on the nonlinear character of neural network, the neural network for helicopter certificated noise prediction was modeled, and trained using the helicopter certificated noise levels data published by FAA. For the significant effects of neural network inputs to performance, the general parameters affecting heli? copter noise levels were analyzed in the paper. The results indicated that the desirable helicopter noise levels could be simulated from the neural network.
作者
仲唯贵
ZHONG Weigui(China Helicopter Research and Development Institute,Jingdezhen 333001,China)
出处
《直升机技术》
2019年第1期6-9,19,共5页
Helicopter Technique
关键词
直升机
适航噪声
神经网络
helicopter
certificated noise
neural network
