数据驱动的升力体飞行器表面热流快速预测

杜文聪; 陈智

doi:10.7638/kqdlxxb-2025.0109

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 >

数据驱动的升力体飞行器表面热流快速预测

研究专栏——高速飞行“新热障”问题研究进展（第二十七届中国科协年会学术论文专题） | 更新时间：2026-05-28

- 数据驱动的升力体飞行器表面热流快速预测
- Data-driven rapid prediction of surface heat flux for lifting-body aircraft
- 空气动力学学报 2025年43卷第7期页码：93-107
- 作者机构：
  
  中国航天空气动力技术研究院,北京,100074
- 作者简介：
  
  杜文聪（2000—），男，河南商丘人，硕士研究生，研究方向：人机与环境工程. E-mail：1131318603@qq.com
  陈智*（1983—），男，湖北武汉人，研究员，研究方向：复杂流动热环境模拟. E-mail：chenzhi@live.cn
- 基金信息：
  
  稳定支持项目（800009000199C1652442）
- DOI：10.7638/kqdlxxb-2025.0109
  中图分类号： V19;V411.3;TP181
- 纸质出版：2025
- 稿件说明：
移动端阅览
杜文聪, 陈智. 数据驱动的升力体飞行器表面热流快速预测[J]. 空气动力学学报, 2025, 43(7): 93−107.

DU Wencong, CHEN Zhi. Data-driven rapid prediction of surface heat flux for lifting-body aircraft[J]. 2025, 43(7): 93-107.
杜文聪, 陈智. 数据驱动的升力体飞行器表面热流快速预测[J]. 空气动力学学报, 2025, 43(7): 93−107. DOI： 10.7638/kqdlxxb-2025.0109

DU Wencong, CHEN Zhi. Data-driven rapid prediction of surface heat flux for lifting-body aircraft[J]. 2025, 43(7): 93-107. DOI： 10.7638/kqdlxxb-2025.0109.

摘要

为实现高速升力体气动热的快速预测、提升预测模型的泛化能力，本文针对一类三维升力体外形，基于参数化建模方法构建了包含15种外形、36种飞行条件，合计540个算例的升力体外形气动热数据集，并提出了一种基于切片和插值策略的数据预处理方式，将飞行器的外形及物面信息转化为神经网络能够识别的矩阵形式。然后，结合Transformer、多层感知机与UNet设计了一种编码器-解码器架构的网络模型D-TMU，并使用深度过参数化卷积替代了网络中的传统卷积。该模型能够在给定飞行器的几何形状、压强分布及飞行条件的基础上，直接预测其表面热流分布，避免了复杂的迭代计算。结果表明：D-TMU模型在测试集中的总体误差为1.21%，高热流区域误差为1.19%，关键点误差为0.97%，单个算例的平均预测时间仅为0.03 s，加速比可达6个数量级。这表明本文模型能够有效捕捉飞行器外形的全局特征，并充分学习局部特征之间的相关性，具有较高的预测精度和速度。同时，在应用于未参与训练的一般升力体外形时，模型仍具备一定的泛化能力。

Abstract

This study developed a rapid method for aerodynamic heating prediction in high-speed lifting bodies. A dataset comprising 540 simulation cases was constructed through parametric modeling

encompassing 15 three-dimensional configurations and 36 flight conditions. To facilitate efficient neural network training

a geometry-aware preprocessing method based on surface slicing and interpolation was proposed

converting complex surface and geometry information into matrix representations. Subsequently

a D-TMU model based on deep learning was introduced. The model employed an encoder–decoder architecture integrating transformer modules

multilayer perceptrons

and UNet structures

with depthwise over-parameterized convolutions replacing standard convolution layers. D-TMU directly predicted surface heat flux distributions without iterative computations from input geometry

pressure distribution

and flight conditions. Validation results demonstrated the predictive performance of the model

with mean prediction errors of 1.21% on the test set

1.19% in high heat flux regions

and 0.97% at key points. The average inference time per case was 0.03 seconds

representing computational acceleration a speedup of approximately six orders of magnitude compared over traditional CFD methods. These results indicated that D-TMU effectively captured global geometric characteristics and local feature interactions

maintaining achieving high predictive accuracy and computational efficiency. In addition

the model exhibited promising generalization capability for lifting body configurations beyond the training set.

关键词

Keywords

references

AN F X, LI L, SU W, et al. Key issues in hypersonic vehicle aerodynamic design[J]. Scientia Sinica (Physica, Mechanica Astronomica), 2021, 51(10): 6−25(in Chinese). doi: 10.1360/SSPMA-2021-0135

YU C Z, LIU W H. Research status of aeroheating prediction technology for hypersionic aircraft[J]. Aeronautical Science Technology, 2021, 32(2): 14−21 (in Chinese). doi: 10.19452/j.issn1007-5453.2021.02.002

TANG Z G, YUAN X X, QIAN W Q, et al. Research progress on the fusion of data obtained by high-speed wind tunnels, CFD and model flight[J]. Acta Aerodynamica Sinica, 2023, 41(8): 44−58(in Chinese). doi: 10.7638/kqdlxxb-2023.0096

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(in Chinese). doi: 10.7638/kqdlxxb-2023.0128

WANG Z, SONG S F, WANG X, et al. Summary and prospect of data-driven aerotheraml modeling prediction methods[J]. Physics of Gases, 2024, 9(4): 39−55(in Chinese). doi: 10.19527/j.cnki.2096-1642.1068

ZHANG Z C, GAO T Y, ZHANG L, et al. Aeroheating agent model based on radial basis function neural network[J]. Acta Aeronautica et Astronautica Sinica, 2021, 42(4): 524167 (in Chinese). doi: 10.7527/S1000-6893.2020.24167

WANG Z, WANG Z Y, WANG X, et al. A data-driven aeroheating prediction model[J]. Acta Aerodynamica Sinica, 2023, 41(5): 12−19(in Chinese). doi: 10.7638/kqdlxxb-2022.0010

REN H J, WANG S, YUAN X X, et al. A flight test based deep learning method for transition heat flux prediction in hypersonic flow[J]. Physics of Fluids, 2022, 34(5): 054106. doi: 10.1063/5.0093438

HE J Y, LYU H Q, LI X D, et al. Aerodynamic thermal prediction method of deformed structure based on conditional diffusion[J]. Acta Aerodynamica Sinica, 2025(X): 1−14(in Chinese). doi: 10.7638/kqdlxxb-2024.0176

WANG Z, ZHANG W W, WANG X, et al. High precision aerodynamic heat prediction method based on data augmentation and transfer learning[J]. Aerospace Science and Technology, 2024, 155: 109663. doi: 10.1016/j.ast.2024.109663

WANG Z, WANG X, ZHANG W W, et al. ED-ResNet: Euler equation embedding double-series residual neural network for aerothermal modeling[J]. AIAA Journal, 2025, 63(1): 134−147. doi: 10.2514/1.j064012

JIANG T R, GUO L, SUN G P, et al. PDI-HFP: An intelligent method for heat flux prediction on hypersonic aircraft based on projection depth images[J]. Engineering Applications of Artificial Intelligence, 2024, 127: 107366. doi: 10.1016/j.engappai.2023.107366

LI T, GUO L, YANG Z G, et al. An automatic shape-aware method for predicting heat flux of supersonic aircraft based on a deep learning approach[J]. Physics of Fluids, 2022, 34(7): 077103. doi: 10.1063/5.0098341

YUAN J C, ZONG W G, ZENG L, et al. CNN-based method for predicting aerodynamic heating[J]. Acta Aerodynamica Sinica, 2024, 42(1): 13−25(in Chinese). doi: 10.7638/kqdlxxb-2023.0072

RONNEBERGER O, FISCHER P, BROX T. U-Net: Convolutional networks for biomedical image segmentation[M]//Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015. Cham: Springer International Publishing, 2015: 234–241. doi: 10.1007/978-3-319-24574-4_28

VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[C]// 31st Conference on Neural Information Processing Systems. Long Beach, California, USA: Curran Associates Inc, 2017: 5998–6008. doi: 10.48550/arXiv.1706.03762

CAO J, LI Y, SUN M, et al. DO-conv: Depthwise over-parameterized convolutional layer[J]. IEEE Trans Image Process, 2022, 31: 3726−3736. doi: 10.1109/tip.2022.3175432

ZHANG L, CHENG X L, AI B C. Normal grid rule for hypersonic heat flux numerical simulation[J]. Mechanics in Engineering, 2014, 36(6): 722−727,741 (in Chinese). doi: 10.6052/1000-0879-14-046

CHEN Z, PAN H L, ZHANG L. Numerical study on the supersonic near-wake flow structures of a cylinder at an angle of attack[J]. Computers Fluids, 2021, 228: 105043. doi: 10.1016/j.compfluid.2021.105043

CHEN Z, MIAO W B, LI Q, et al. Influence of the gas model on the boundary layer characteristic parameters and transition prediction for the Mars entry capsule[J]. Scientia Sinica (Technologica), 2022, 52(2): 292−302 (in Chinese). doi: 10.1360/SST-2021-0476

LIOU M S. A sequel to AUSM: AUSM+[J]. Journal of Computational Physics, 1996, 129(2): 364−382. doi: 10.1006/jcph.1996.0256

YOON S, JAMESON A. Lower-upper symmetric-Gauss-seidel method for the Euler and navier-stokes equations[J]. AIAA Journal, 1988, 26(9): 1025−1026. doi: 10.2514/3.10007

MILLER C G. Experimental and predicted heating distributions for biconics at incidence in air at Mach 10: NASA-TP-2334[R]. NASA, 1984.

DOSOVITSKIY A, BEYER L, KOLESNIKOV A, et al. An Image is worth 16×16 words: Transformers for image recognition at scale[C]// International Conference on Learning Representations. 2021. doi: 10.48550/arXiv.2010.11929

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

基于稀疏神经核的多保真度代理模型

基于柔性智能蒙皮的翼型流场预测和重建方法

基于深度神经网络的高速翼型流场降阶模型