基于深度神经网络的高速翼型流场降阶模型

陈泽伟; 李立; 孔轶男; 陈刚

doi:10.7638/kqdlxxb-2024.0050

欢迎来到能源电力期刊网！

English Version 期刊登录

您当前的位置：

首页 >

文章列表页 >

基于深度神经网络的高速翼型流场降阶模型

研究论文 | 更新时间：2026-05-28

- 基于深度神经网络的高速翼型流场降阶模型
- Deep neural network-based reduced-order modeling of high-speed airfoil flow field
- 空气动力学学报 2025年43卷第10期页码：33-43
- 作者机构：
  
  1. 西安交通大学，复杂服役环境重大装备结构强度与寿命全国重点实验室,西安,710049
  2. 西安交通大学航天航空学院，先进飞行器服役环境与控制陕西省重点实验室,西安,710049
  3. 中国航空工业集团公司西安航空计算技术研究所,西安,710000
  4. 西北工业大学计算机学院,西安,710129
  5. 中国空气动力研究与发展中心,绵阳,621000
- 作者简介：
  
  陈泽伟（2000—），男，福建漳州人，硕士研究生，研究方向：复杂流动人工智能建模技术. E-mail：zwchen2024@163.com
  陈刚*，教授，研究方向：智能飞行器气动总体设计. E-mail：aachengang@xjtu.edu.cn
- 基金信息：
  
  国家自然科学基金（92371201, 52192633）;航空科学基金（ ASFC-20220019070002）;陕西省杰出青年基金（2022JC-03）
- DOI：10.7638/kqdlxxb-2024.0050
  中图分类号： V211.3
- 纸质出版：2025
- 稿件说明：
移动端阅览
陈泽伟, 李立, 孔轶男, 等. 基于深度神经网络的高速翼型流场降阶模型[J]. 空气动力学学报, 2025, 43(10): 33−43.

CHEN Zewei, LI Li, KONG Yinan, et al. Deep neural network-based reduced-order modeling of high-speed airfoil flow field[J]. 2025, 43(10): 33-43.
陈泽伟, 李立, 孔轶男, 等. 基于深度神经网络的高速翼型流场降阶模型[J]. 空气动力学学报, 2025, 43(10): 33−43. DOI： 10.7638/kqdlxxb-2024.0050

CHEN Zewei, LI Li, KONG Yinan, et al. Deep neural network-based reduced-order modeling of high-speed airfoil flow field[J]. 2025, 43(10): 33-43. DOI： 10.7638/kqdlxxb-2024.0050.

摘要

基于深度神经网络的流体建模近年来备受关注。然而，之前的研究主要集中在低速和亚声速流动上，对于高速流场的重构研究相对较少。为了解决高速翼型绕流的快速预测问题，本文提出了一种基于深度神经网络技术的流场降阶模型。该模型利用全连接神经网络和反卷积神经网络，能够建立翼型绕流工况与超声速流场之间的映射关系。首先，通过数值模拟方法构建翼型超声速流场数据集，其中攻角和来流马赫数作为输入数据，流场状态信息作为输出数据。其次，根据输入输出特征构建深度神经网络模型，并进行训练，损失函数均方根误差收敛至

0.0019

。最后，对模型的预测精度和泛化性能进行了分析。神经网络模型在测试集上的流场预测均方根误差低于0.004，最大相对误差值集中在0.03附近，且3个流场状态变量的相关系数均高于0.99。利用该模型对给定工况下的高速定常流场进行预测，结果与数值模拟结果形态高度一致，3个流场变量的相关系数均超过0.998，不同截面处的流场预测值与CFD预测值吻合良好，表明该模型具备良好的预测精度和内插泛化能力。此外，神经网络模型同样具备对数据集以外的马赫数工况进行流场预测的能力，在马赫数小于13的范围内具备一定的外推状态泛化能力。与CFD计算相比，深度神经网络降阶模型的预测速度提升了至少两个数量级，而且所需流场预测数据越多，深度神经网络模型效率优势越显著。

Abstract

The flow field prediction based on deep neural networks has attracted considerable attention in recent years. However

previous studies mainly focused on low-speed and subsonic conditions

whereas much less attention has been paid to reconstructing supersonic and hypersonic flow fields. To rapidly and accurately predict supersonic and hypersonic airfoil flows

this paper proposes a flow field reduc

ed-order model based on deep neural networks

utilizing fully-connected neural and deconvolutional neural networks to establish a mapping relationship between the flow conditions and flow fields. Firstly

a dataset of supersonic airfoil flow fields is constructed by numerical simulations in a wide range of the angles of attack and incoming Mach numbers. Secondly

a deep neural network model is constructed and trained

with the root mean square error of the loss function converged to 0.0019. Finally

the prediction accuracy and generalization performance of the model are analyzed. The root mean square error of the neural network model for the test set is less than 4×10

–3

the maximal relative error is about 0.03

and the correlation coefficients between true and predicted flow fields are higher than 0.99

indicating that the model has good prediction accuracy and interpolation generalization ability. In addition

the neural network model is also able to predict the flow fields for Mach numbers outside the dataset

exhibiting good generalization ability for extrapolated conditions in the range of the Mach number less than 13. Compared to numerical simulations

the prediction speed of the deep-neural-network-based reduced-order model is faster by at least two orders of magnitude

and the efficiency is proportional to the amount of predicted flow fields.

关键词

Keywords

references

TERRY N B, CONE P P. Hypersonic technology: An evolution in nuclear weapons?[J]. Strategic Studies Quarterly, 2020, 14(2): 74−99.

ZHANG T T, WANG Z G, HUANG W, et al. A review of parametric approaches specific to aerodynamic design process[J]. Acta Astronautica, 2018, 145: 319−331. doi: 10.1016/j.actaastro.2018.02.011

WANG Z, XIAO D, FANG F, et al. Model identification of reduced order fluid dynamics systems using deep learning[J]. International Journal for Numerical Methods in Fluids, 2018, 86(4): 255−268. doi: 10.1002/fld.4416

EDWARDS W S, TUCKERMAN L S, FRIESNER R A, et al. Krylov methods for the incompressible navier-stokes equations[J]. Journal of Computational Physics, 1994, 110(1): 82−102. doi: 10.1006/jcph.1994.1007

ROWLEY C W, COLONIUS T, MURRAY R M. Model reduction for compressible flows using POD and Galerkin projection[J]. Physica D: Nonlinear Phenomena, 2004, 189(1-2): 115−129. doi: 10.1016/j.physd.2003.03.001

HUI X Y, BAI J Q, WANG H, et al. Fast pressure distribution prediction of airfoils using deep learning[J]. Aerospace Science and Technology, 2020, 105: 105949. doi: 10.1016/j.ast.2020.105949

CHEN G, LI Y M. Advances and prospects of the reduced order model for unsteady flow and its application[J]. Advances in Mechanics, 2011, 41(6): 686−701 (in Chinese).

LEE S, HA J W, ZOKHIROVA M, et al. Background information of deep learning for structural engineering[J]. Archives of Computational Methods in Engineering, 2018, 25(1): 121−129. doi: 10.1007/s11831-017-9237-0

KUTZ J N. Deep learning in fluid dynamics[J]. Journal of Fluid Mechanics, 2017, 814: 1−4. doi: 10.1017/jfm.2016.803

TANG Z G, ZHU L Y, XIANG X H, et al. Some research progress and prospect of Intelligent Aerodynamics[J]. Acta Aerodynamica Sinica, 2023, 41(7): 1−35. doi: 10.7638/kqdlxxb-2023.0128

LIU X, FENG W H, LIAN F, et al. Aerodynamic data fusion method based on physics-informed neural network[J]. Acta Aerodynamica Sinica, 2023, 41(8): 87−96. doi: 10.7638/kqdlxxb-2023.0060

XIA Y, LIU M. Traffic flow prediction based on spatial-temporal attention convolutional neural network[J]. Journal of Southwest Jiaotong University, 2023, 58(2): 340−347 (in Chinese).

CHEN H, GUO M M, TIAN Y, et al. Progress of convolution neural networks in flow field reconstruction[J]. Chinese Journal of Theoretical and Applied Mechanics, 2022, 54(9): 2343−2360 (in Chinese). doi: 10.6052/0459-1879-22-130

ZHU X B, CAI Z, WU J J, et al. Convolutional neural network based combustion mode classification for condition monitoring in the supersonic combustor[J]. Acta Astronautica, 2019, 159: 349−357. doi: 10.1016/j.actaastro.2019.03.072

PARK S Y, AHN J. Deep neural network approach for fault detection and diagnosis during startup transient of liquid-propellant rocket engine[J]. Acta Astronautica, 2020, 177: 714−730. doi: 10.1016/j.actaastro.2020.08.019

YONEKURA K, HATTORI H. Framework for design optimization using deep reinforcement learning[J]. Structural and Multidisciplinary Optimization, 2019, 60(4): 1709−1713. doi: 10.1007/s00158-019-02276-w

ZHANG W W, KOU J Q, LIU Y L. Prospect of artificial intelligence empowered fluid mechanics[J]. Acta Aeronautica et Astronautica Sinica, 2021, 42(4): 524689 (in Chinese).

WANG Y X, HAN R K, LIU Z Y, et al. Progress of deep learning modeling technology for fluid mechanics[J]. Acta Aeronautica et Astronautica Sinica, 2021, 42(4): 524779 (in Chinese).

HILL W J, HUNTER W G. A review of response surface methodology: a literature survey[J]. Technometrics, 1966, 8(4): 571−590. doi: 10.1080/00401706.1966.10490404

PENG JZ, CHEN S H, AUBRY N, et al. Time-variant prediction of flow over an airfoil using deep neural network[J]. Physics of Fluids, 2020, 32(12): 123602. doi: 10.1063/5.0022222

WU M Y, WU Y, YUAN X Y, et al. Fast prediction of flow field around airfoils based on deep convolutional neural network[J]. Applied Sciences, 2022, 12(23): 12075. doi: 10.3390/app122312075

TAN Z W, LI R Z, ZHANG Y F. Flow field reconstruction of 2D hypersonic inlets based on a variational autoencoder[J]. Aerospace, 2023, 10(9): 825. doi: 10.3390/aerospace10090825

KONG C, CHANG J T, WANG Z A, et al. Data-driven super-resolution reconstruction of supersonic flow field by convolutional neural networks[J]. AIP Advances, 2021, 11(6): 065321. doi: 10.1063/5.0056569

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于放大因子输运方程模型的高速前体转捩特性研究

基于多支微射流的超声速射流主动控制

无人机气动/风能获取一体化设计

基于U-Net的磁浮飞行风洞气动力干扰剥离算法

基于稀疏神经核的多保真度代理模型