改进的RBF神经网络在翼梢小翼优化设计中的应用
收稿日期: 2013-09-17
修回日期: 2013-12-05
网络出版日期: 2014-01-08
基金资助
国家“863”计划（2012AA01A304）
Application of an Improved RBF Neural Network on Aircraft Winglet Optimization Design
Received date: 2013-09-17
Revised date: 2013-12-05
Online published: 2014-01-08
Supported by
National High-tech Research and Development Program of China (2012AA01A304)
为了提高径向基函数（RBF）神经网络模型的预测精度，在其基础上提出了一种自适应RBF神经网络模型。该预测模型在RBF神经网络模型表达式中引入自适应向量（向量维数与样本点自变量维数相同），采用优化搜索方式确定自适应向量值，从而提高模型预测的准确度和普适性。与其他RBF神经网络模型的改进相比，本文直接从改变基函数的形式入手，使用较少的参数优化达到对网络模型的自适应构造；该方法本质上改变了基函数网络中心与宽度对网络模型预测的作用以及样本点自变量向量的各个维对因变量的影响度，其对目标问题具有自适应性。将本文的自适应RBF神经网络模型应用在基于机身+机翼+翼梢小翼模型的翼梢小翼优化设计中，在约束弯矩的情况下进行巡航减阻优化设计，设计结果验证了该预测模型的可行性，表明其具有一定的工程实用价值。
白俊强 , 王丹 , 何小龙 , 李权 , 郭兆电 . 改进的RBF神经网络在翼梢小翼优化设计中的应用[J]. 航空学报, 2014 , 35(7) : 1865 -1873 . DOI: 10.7527/S1000-6893.2013.0487
A self-adaptive radial basis function (RBF) neural network is proposed in order to improve the prediction accuracy of the original RBF. A self-adaptive vector with the same dimension as the sample vector is introduced into the traditional RBF network. In contrast to other RBF neural network models, the current approach achieves the self-adaptive construction of the network by altering the form of the basis function directly, which reduces the number of variables to be optimized. This adaptive approach substantially changes the impact of the center and width of the RBF neural network on its prediction as well as the influence of each variable of the independent vector on the dependent vector. Thus the introduced vector enables the adaptability of the RBF neural network with respect to variant problems. Moreover, the accuracy and the universality of the prediction model are also improved due to the optimization of the self-adaptive vector. The proposed self-adaptive RBF neural network is applied to a winglet optimization design of a wing-body-winglet configuration. The optimization objective is to minimize the cruise drag with wing-root bending moment restriction. The optimization result confirms the effectiveness and the capability for engineering application of the self-adaptive RBF neural network.
[1] Qiu Y S, Bai J Q, Hua J. Flow field estimation method based on proper orthogonal decomposition and surrogate model[J]. Acta Aeronautica et Astronautica Sinica, 2013, 34(6): 1249-1260.(in Chinese) 邱亚松, 白俊强, 华俊.基于本征正交分解和代理模型的流场预测方法[J].航空学报, 2013, 34(6): 1249-1260.
[2] Mu X F, Yao W X, Yu X Q. A survey of surrogate models used in MDO[J]. Chinese Journal of Computational Mechanics, 2005, 22(5): 608-612.(in Chinese) 穆雪峰, 姚卫星, 余雄庆. 多学科设计优化中常用代理模型的研究[J]. 计算力学学报, 2005, 22(5): 608-612.
[3] Su W. Aerodynamic optimization design based on computational fluid dynamics and surrogate model. Xi'an: School of Aeronautics, Northwestern Polytechnical University, 2007.(in Chinese) 苏伟. 基于CFD技术和代理模型的气动外形优化设计方法研究. 西安: 西北工业大学航空学院, 2007.
[4] Shi Y, Han L Q, Lian X Q. Method design and examples analysis of neural networks[M]. Beijing: Beijing University of Posts and Telecommunications Press, 2009: 85-86.(in Chinese) 施彦, 韩力群, 廉小亲.神经网络设计方法与实例分析[M].北京: 北京邮电大学出版社, 2009: 85-86.
[5] Moody J E, Darken C J. Fast learning in networks of locally-tuned processing units[J].Neural Computation, 1989, 1(2): 281-294.
[6] Zhang B. Overview on design of RBF network based on fuzzy clustering[J]. Microcomputer & Its Applications, 2012, 31(12): 1-3.(in Chinese) 张彬.基于聚类算法的RBF神经网络设计综述[J]. 微型机与应用, 2012, 31(12): 1-3.
[7] Zhu C J. Zhang Y. Research of improved fuzzy C-means clustering algorithm[J]. Journal of Henan Univercity, 2012, 42(1): 92-95.(in Chinese) 朱长江, 张缨. 模糊-均值聚类算法的改进研究[J]. 河南大学学报, 2012, 42(1): 92-95.
[8] Liu W J, Guo J. Study on the OLS algorithm of RBF neural network[J]. Journal of Tianjin Polytechnic University, 2001, 21(2): 71-73.(in Chinese) 刘文菊, 郭景. RBF 神经网络中心选取OLS算法的研究[J].天津工业大学学报, 2001, 21(2): 71-73.
[9] Liu M Q, Liao X X. A robust learning algorithm for RBF neural networks[J]. Journal of Huazhong University, 2000, 28(2): 8-10.(in Chinese) 刘妹琴, 廖晓昕.RBF神经网络的一种鲁棒学习算法[J].华中理工大学学报, 2000, 28(2): 8-10.
[10] Yuan C R. Artificial intelligence principles and applications[M]. Beijing: Tsinghua University Press, 2000: 1-4.(in Chinese) 袁曾任. 人工神经网络及其应用[M]. 北京: 清华大学出版社, 2000: 1-4.
[11] Haykin S. Neuarl networks: a comprehensive foundation[M]. 2nd ed. Patparganj: Person Education(Singapore) Pte.Ltd., 2005: 28-37.
[12] Cai Y H, Qi R Y, Cai J, et al. Online modeling for switched reluctance motor using radial basis function neural network and its experimental validation[J]. Acta Aeronautica et Astronautica Sinica, 2012, 33(4): 705-714.(in Chinese) 蔡永红, 齐瑞云, 蔡骏, 等. 基于RBF神经网络的开关磁阻电机在线建模及其实验验证[J]. 航空学报, 2012, 33(4): 705-714.
[13] Pang Z, Xu W H. Learning algorithm for RBF neural networks based on improved k-means algorithm[J]. Computer Engineering and Applications, 2012, 48(11): 161-163.(in Chinese) 庞振, 徐蔚鸿. 一种基于改进k-means 的 RBF 神经网络学习方法[J]. 计算机工程与应用, 2012, 48(11): 161-163.
[14] Jiang Y Q. Aircraft winglet design[M]. Beijing: Aviation Industry Press, 2009: 1-2.(in Chinese) 江永泉. 飞机翼梢小翼设计[M]. 北京: 航空工业出版社, 2009: 1-2.
[15] Jiang W, Jin H B, Sun W P. Aerodynamic optimization for winglets based on multi-level response surface method[J]. Acta Aeronautica et Astronautica Sinica, 2010, 31 (9): 1746-1751.(in Chinese) 姜琬, 金海波, 孙卫平. 基于多级响应面法的翼梢小翼气动优化设计[J]. 航空学报, 2010, 31(9): 1746-1751.
[16] Sederberg T W, Parry S R. Free-form deformation of solid geometric models[J]. Computer Graphics, 1986, 20 (4): 151-160.
[17] Zhu X X. Free curves and surfaces modeling technology [M]. Beijing: Science Press, 2000: 236-243.(in Chinese) 朱心雄. 自由曲线曲面造型技术[M]. 北京: 科学出版社, 2000: 236-243.
[18] Huang J T, Gao Z H, Bai J Q, et al. Laminar airfoil aerodynamic optimization design based on Delaunay graph mapping and FFD technique[J]. Acta Aeronautica et Astronautica Sinica, 2012, 33(10): 1817-1826.(in Chinese) 黄江涛, 高正红, 白俊强, 等.应用Delaunay图映射与FFD技术的层流翼型气动优化设计[J]. 航空学报, 2012, 33(10): 1817-1826.
/
| 〈 |
|
〉 |
CJA