| 神经网络在直升机概念设计中的应用 |
| |
| 引用本文： | 孙伟,张呈林. 神经网络在直升机概念设计中的应用[J]. 南京航空航天大学学报, 2009, 41(1) |
| |
| 作者姓名： | 孙伟 张呈林 |
| |
| 作者单位： | 南京航空航天大学直升机旋翼动力学国家级重点实验室,南京,210016 |
| |
| 摘 要： | 运用人工神经网络算法得到了单旋翼布局直升机的总体参数和部件重量估算方法,为直升机方案论证初期提供了一个快速有效的决策手段.对神经网络应用了多种训练算法,并与传统回归公式进行比较,可以发现采用BR训练算法的神经网络具有比传统回归方法更好的非线性表达能力,同时又表现出比LM,SCG等常规训练算法更好的范化能力,对训练样本和检验样本都有较高的估算精度,适用于型号设计初期的参数估算.
|
| 关 键 词： | 直升机 概念设计 神经网络 范化能力 |
Application of Neural Networks in Helicopter Conceptual Design |
| |
| Sun Wei,Zhang Chenglin. Application of Neural Networks in Helicopter Conceptual Design[J]. Journal of Nanjing University of Aeronautics & Astronautics, 2009, 41(1) |
| |
| Authors: | Sun Wei Zhang Chenglin |
| |
| Affiliation: | National Key Laboratory of Rotorcraft Aeromechanics;Nanjing University of Aeronautics & Astronautics;Nanjing;210016;China |
| |
| Abstract: | A method for estimating single-rotor helicopter preliminary parameters and group weights with artificial neural networks is proposed.It is fast and effective in helicopter sizing stage.The neural networks are trained with several different training functions respectively.Compared with traditional multiple regression method,the neural network with BR training function was more accurate in estimating nonliner functions and is better in generalization than some normal network training functions,such as LM,SCG.... |
| |
| Keywords: | helicopter conceptual design neural network generalization |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
