基于BP神经网络的冰形特征参数预测

柴聪聪; 易贤; 郭磊; 王俊

doi:10.11729/syltlx20200016

基于BP神经网络的冰形特征参数预测

doi: 10.11729/syltlx20200016

柴聪聪^{1, 2,},
易贤^2, ,,
郭磊^1, ,,
王俊¹

1.
电子科技大学计算机科学与工程学院, 成都 611731
2.
中国空气动力研究与发展中心结冰与防/除冰重点实验室, 四川绵阳 621000

基金项目:

国家自然科学基金 11472296

详细信息

作者简介:
柴聪聪(1997-), 女, 河南新乡人, 硕士。研究方向: 结冰预测、深度学习。通信地址: 四川省成都市高新区西源大道2006号(611731)。E-mail: 18224004883@163.com

通讯作者:
易贤, E-mail: yixian_2000@163.com

郭磊, E-mail: leiguo@uestc.edu.cn

中图分类号: V211.24
计量
- 文章访问数: 498
- HTML全文浏览量: 201
- PDF下载量: 38
- 被引次数: 0
出版历程
- 收稿日期: 2020-02-12
- 修回日期: 2020-04-18
- 刊出日期: 2021-06-25

Prediction of ice shape characteristic parameters based on BP nerual network

CHAI Congcong^{1, 2
,},
YI Xian^{2
, ,},
GUO Lei^{1
, ,},
WANG Jun¹

1.
School of Computer Science and Engineering, University of Electronic Science and Technology of China, Chengdu 611731, China
2.
Key Laboratory of Icing and Anti/De-icing, China Aerodynamics Research and Development Center, Mianyang Sichuan 621000, China

摘要

摘要: 机翼结冰影响了飞机飞行的气动特性，严重时将会引起事故，对冰形特征参数进行预测对翼型气动特性研究以及后续防除冰措施具有重要的意义。本文利用BP神经网络，建立翼型冰形特征参数预测模型，并采用k折交叉验证进行网络结构选择，以气象与飞行条件作为输入，结冰极限、冰角高度和角度等冰形特征参数作为输出。结果表明：预测的冰形特征参数（除下冰角高度外）与数值结果相对误差低于5%，证明该方法具有较好的预测效果。
- 机翼结冰 /
- 冰形特征参数 /
- 神经网络 /
- k折交叉验证
Abstract: Airfoil icing affects the aerodynamic characteristics of aircraft flight, which can lead to accidents when it is serious. The prediction of ice shape parameters can effectively prevent accidents. In this paper, BP neural network is used to establish the prediction model of airfoil ice shape characteristic parameters, and k-fold cross validation is used to select the network structure, in which the meteorological and flight conditions are the inputs, and ice shape characteristic parameters such as the ice limit, the ice angle height and angle are the outputs. The experimental results show that the relative error between the predicted ice shape parameters (except for the height of the lower ice angle) and the numerical results is less than 5%, which proves that the method has a good prediction ability.
- airfoil icing /
- ice shape characteristic parameters /
- neural network /
- k-fold cross validation

HTML全文

图 1 结冰几何特征描述^[16]

Figure 1. Description of icing geometric properties^[16]

下载: 全尺寸图片幻灯片

图 2 SAE规定的冰形特征参数示意图

Figure 2. Icing characteristic parameters provided by SAE

下载: 全尺寸图片幻灯片

图 3 霜冰的冰形特征参数

Figure 3. Icing characteristic parameters of glaze ice

下载: 全尺寸图片幻灯片

图 4 数值计算与实验冰形对比

Figure 4. Comparison between the numerical calculation method and the experiment

下载: 全尺寸图片幻灯片

图 5 单层BP神经网络模型

Figure 5. Single-layer BP network

下载: 全尺寸图片幻灯片

图 6 k折交叉验证

Figure 6. k-fold cross validation

下载: 全尺寸图片幻灯片

图 7 不同神经元模型的10折交叉验证平均测试误差

Figure 7. Average test error of 10 fold cross validation for different neuron models

下载: 全尺寸图片幻灯片

图 8 各算例的翼型结冰结果

Figure 8. Airfoil icing results of cases

下载: 全尺寸图片幻灯片

表 1 气象和飞行条件参数设置

Table 1. Meteorological and flight condition parameters setting

参数	设置范围
飞行速度v/(m·s^-1)	70~150
环境温度T/K	243~267
液态水含量LWC/(g·m^-3)	0.2~0.8
平均水滴直径MVD/μm	20~60
结冰时间t/s	360~1350

下载: 导出CSV

表 2 算例的气象和飞行条件参数

Table 2. Meteorological and flight parameters of cases

算例	v/(m·s^-1)	T/K	LWC/(g·m^-3)	MVD/μm	t/s
1	85	243	0.75	40	420
2	90	243	0.75	20	300
3	67	245	0.6	15	540
4	77	258	0.8	30	600
5	110	263	0.85	40	540
6	77	260	0.6	30	540
7	88	245	0.9	33	480
8	92	255	0.6	25	480

下载: 导出CSV

表 3 冰形特征参数计算与预测结果

Table 3. Calculation and prediction results of ice shape characteristic parameters

算例		S_u/(m·c^-1)	S_l/(m·c^-1)	h_u/(m·c^-1)	h_l/(m·c^-1)	θ_u/(°)	θ_l/(°)
1	C	0.0490	－0.1444	0.0213	0.0207	183.5	225.0
1	P	0.0489	－0.1472	0.0211	0.0220	186.8	225.5
2	C	0.0278	－0.0623	0.0147	0.0143	209.3	231.6
2	P	0.0269	－0.0631	0.0137	0.0120	212.8	229.3
3	C	0.0212	－0.0366	0.0116	0.0111	212.9	237.0
3	P	0.0228	－0.0349	0.0111	0.0118	215.0	230.4
4	C	0.0381	－0.1022	0.0363	0.0245	177.0	267.7
4	P	0.0357	－0.1021	0.0354	0.0317	176.5	265.4
5	C	0.0522	－0.1624	0.0365	0.0160	127.3	288.5
5	P	0.0481	－0.1760	0.0366	0.0178	129.0	284.1
6	C	0.0358	－0.1022	0.0242	0.0179	183.3	270.4
6	P	0.0368	－0.1002	0.0228	0.0251	185.6	270.9
7	C	0.0432	－0.1184	0.0298	0.0280	202.1	236.6
7	P	0.0401	－0.1251	0.0286	0.0353	194.8	226.8
8	C	0.0261	－0.0572	0.0163	0.0151	208.0	230.6
8	P	0.0255	－0.0557	0.0160	0.0167	206.5	232.0

下载: 导出CSV

表 4 冰形特征参数预测误差

Table 4. Prediction error of ice shape characteristic parameters

算例		S_u/(m·c^-1)	S_l/(m·c^-1)	h_u/(m·c^-1)	h_l/(m·c^-1)	θ_u/(°)	θ_l/(°)
APE/%	1	0.20	1.94	0.94	6.28	1.79	0.22
	2	3.23	1.28	6.80	16.08	1.67	0.99
	3	7.55	4.64	4.31	6.31	0.99	2.78
	4	6.30	0.10	2.48	29.39	0.29	0.87
	5	7.85	8.37	0.27	11.25	1.33	1.54
	6	2.79	1.96	5.79	40.22	1.23	0.19
	7	7.18	5.66	4.03	26.07	3.64	4.14
	8	2.30	2.62	1.84	10.60	0.72	0.61
MAPE/%		4.68	3.32	3.31	18.28	1.46	1.42

下载: 导出CSV

参考文献(19)

[1]	CEBECI T, KAFYEKE F. Aircrafticing[J]. Annual Review of Fluid Mechanics, 2003, 35(1): 11-21. doi: 10.1146/annurev.fluid.35.101101.161217
[2]	王建涛, 易贤, 肖中云, 等. ARJ21-700飞机冰脱落数值模拟[J]. 空气动力学学报, 2013, 31(4): 430-436. https://www.cnki.com.cn/Article/CJFDTOTAL-KQDX201304004.htm WANG J T, YI X, XIAO Z Y, et al. Numerical simulation of ice shedding from ARJ21-700[J]. Acta Aerodynamica Sinica, 2013, 31(4): 430-436. https://www.cnki.com.cn/Article/CJFDTOTAL-KQDX201304004.htm
[3]	RATVASKY T P, BARNHART B P, LEE S. Current methods modeling and simulating icing effects on aircraft performance, stability, control[J]. Journal of Aircraft, 2010, 47(1): 201-211. doi: 10.2514/1.44650
[4]	闫鹏庆, 牛亚宏. 人工模拟结冰飞行试验技术研究[J]. 民用飞机设计与研究, 2018(1): 71-74. doi: 10.19416/j.cnki.1674-9804.2018.01.013 YAN P Q, NIU Y H. Research on artificial icing flight test technology[J]. Civil Aircraft Design & Research, 2018(1): 71-74. doi: 10.19416/j.cnki.1674-9804.2018.01.013
[5]	高郭池, 丁丽, 李保良, 等. 气动除冰类飞机结冰风洞试验适航审定技术[J]. 实验流体力学, 2019, 33(2): 85-94. doi: 10.11729/syltlx20180067 GAO G C, DING L, LI B L, et al. Airworthiness certification technology about icing wind tunnel test for pneumatic de-icing aircraft[J]. Journal of Experiments in Fluid Mechanics, 2019, 33(2): 85-94. doi: 10.11729/syltlx20180067
[6]	李鑫, 白俊强, 王昆. 机翼积冰数值模拟[J]. 航空动力学报, 2013, 28(12): 2663-2670. doi: 10.13224/j.cnki.jasp.2013.12.005 LI X, BAI J Q, WANG K. Numerical simulation of ice accretions on aircraft wing[J]. Journal of Aerospace Power, 2013, 28(12): 2663-2670. doi: 10.13224/j.cnki.jasp.2013.12.005
[7]	张强, 陈迎春, 周涛, 等. 民用飞机机翼结冰试验与数值预测[J]. 航空动力学报, 2017, 32(1): 22-26. doi: 10.13224/j.cnki.jasp.2017.01.004 ZHANG Q, CHEN Y C, ZHOU T, et al. Test and numerical prediction of ice accretions on civil aircraft wing[J]. Journal of Aerospace Power, 2017, 32(1): 22-26. doi: 10.13224/j.cnki.jasp.2017.01.004
[8]	SAE. ARP5904, Airborne Icing Tankers[S]. Washington: Society of Automotive Engineers, 2007.
[9]	FOSSATI M, HABASHI W G. Multiparameter analysis of aero-icing problems using proper orthogonal decomposition and multidimensional interpolation[J]. AIAA Journal, 2013, 51(4): 946-960. doi: 10.2514/1.J051877
[10]	OGRETIM E, HUEBSCH W, SHINN A. Aircraft ice accretion prediction based on neural networks[J]. Journal of Aircraft, 2006, 43(1): 233-240. doi: 10.2514/1.16241
[11]	潘环, 艾剑良. 飞机结冰冰形预测的建模与仿真[J]. 系统仿真学报, 2014, 26(1): 221-224, 229. doi: 10.16182/j.cnki.joss.2014.01.009 PAN H, AI J L. Modeling and simulation of aircraft ice shape prediction[J]. Journal of System Simulation, 2014, 26(1): 221-224, 229. doi: 10.16182/j.cnki.joss.2014.01.009
[12]	CHANG S N, LENG M Y, WU H W, et al. Aircraft ice accretion prediction using neural network and wavelet packet transform[J]. Aircraft Engineering and Aerospace Technology, 2016, 88(1): 128-136. doi: 10.1108/aeat-05-2014-0057
[13]	张强, 高正红. 基于神经网络的翼型积冰预测[J]. 飞行力学, 2011, 29(2): 6-9. doi: 10.13645/j.cnki.f.d.2011.02.002 ZHANG Q, GAO Z H. Prediction of ice accretions based on the neural net[J]. Flight Dynamics, 2011, 29(2): 6-9. doi: 10.13645/j.cnki.f.d.2011.02.002
[14]	KIM H, BRAGG M. Effects of leading-edge ice accretion geometry on airfoil performance[C]//Proc of the 17th Applied Aerodynamics Conference. 1999. doi: 10.2514/6.1999-3150
[15]	BRAGG M, BROEREN A, ADDY H, et al. Airfoil ice-accretion aerodynamic simulation[R]. AIAA-2007-0085, 2007. doi: 10.2514/6.2007-85
[16]	RUFF G, ANDERSON D. Quantification of ice accretions for icing scaling evaluations[R]. AIAA-98-0915, 1998. doi: 10.2514/6.1998-195
[17]	SAE. AIR5666, Icing wind tunnel interfacility comparison tests[S]. Washington: SAE Aero-space Information Report, 2012.
[18]	易贤. 飞机积冰的数值计算与积冰试验相似准则研究[D]. 绵阳: 中国空气动力研究与发展中心, 2007. YI X. Numerical computation of aircraft icing and study on icing test scaling law[D]. Mianyang: China Aerodynamics Research and Development Center, 2007.
[19]	王小川, 史峰, 郁磊. MATLAB神经网络43个案例分析[M]. 北京: 北京航空航天大学出版社, 2013. WANG X C, SHI F, YU L. Analysis of 43 cases of MATLAB neural network[M]. Beijing: Beijing University of Aeronautics & Astronautics Press, 2013.