飞行器设计生成式模型专栏

基于径向基生成式对抗网络的多源数据融合方法

  • 胡登峰 ,
  • 向渝 ,
  • 张骏 ,
  • 杨佳辰 ,
  • 汪文勇
展开
  • 电子科技大学 计算机科学与工程学院,成都 611731
.E-mail: jcxiang@uestc.edu.cn

收稿日期: 2024-11-01

  修回日期: 2025-01-13

  录用日期: 2025-04-15

  网络出版日期: 2025-04-25

基金资助

国家自然科学基金(62250067);中央高校基本科研业务费专项资金(ZYGX2024Z009)

Multi-source data fusion method based on radial basis function generative adversarial network

  • Dengfeng HU ,
  • Yu XIANG ,
  • Jun ZHANG ,
  • Jiachen YANG ,
  • Wenyong WANG
Expand
  • School of Computer Science and Engineering,University of Electronic Science and Technology of China,Chengdu 611731,China

Received date: 2024-11-01

  Revised date: 2025-01-13

  Accepted date: 2025-04-15

  Online published: 2025-04-25

Supported by

National Natural Science Foundation of China(62250067);The Fundamental Research Funds for the Central Universities(ZYGX2024Z009)

摘要

飞行器气动数据可通过数值计算、风洞试验、飞行试验等途径获得。如何在低成本的基础上获得高保真度且不确定度较低的气动数据依然是一个挑战。在深度学习方法中,利用多层神经网络能够在更高维度上进行特征的表征与组合,从而有效地融合来自不同数据源的相关特征,提升模型的预测精度、可信度。然而,现有方法在融合不同数据集特征时,仍面临诸多技术难题。针对这一问题,提出了一种基于径向基生成式对抗网络(RBFGAN)的多源数据融合方法——RGAN-MSFM。该方法能够融合不同来源的气动数据集,挖掘多源气动数据特征之间的互补性、相关性。利用RBFGAN融合不同来源数据中的互补状态特征,生成新的融合数据集;构建功能网络和上下文网络,深入探索各数据源之间的相关性,从而实现多源数据特征的融合学习。基于包含计算流体力学(CFD)数据、风洞试验数据的数据集,以及包含CFD数据、飞行试验数据的数据集,设计了2组仿真实验。实验结果显示,采用该模型进行数据特征融合后,数据集的精度平均提高了30.2%、67.5%;与传统数据融合方法相比,不确定度分别降低了23.9%、27.6%。

本文引用格式

胡登峰 , 向渝 , 张骏 , 杨佳辰 , 汪文勇 . 基于径向基生成式对抗网络的多源数据融合方法[J]. 航空学报, 2025 , 46(10) : 631478 -631478 . DOI: 10.7527/S1000-6893.2025.31478

Abstract

Aerodynamic data of aircraft can be obtained through methods such as numerical simulations, wind tunnel tests, and flight tests. However, obtaining high-fidelity aerodynamic data with low uncertainty at a low cost remains a challenging problem. In deep learning methods, multi-layer neural networks can represent and combine features in higher dimensions, effectively integrating relevant features from different data sources to improve the model’s prediction accuracy and reliability. However, the existing methods still face numerous technical challenges when fusing features from different datasets. To address this issue, we propose a multi-source data fusion method based on Radial Basis Function Generative Adversarial Network (RBFGAN)—RGAN-MSFM. This method can fuse aerodynamic datasets from different sources, mining the complementarity and correlation between multi-source aerodynamic data features. The proposed method first uses RBFGAN to fuse complementary state features from different data sources to generate a new fused dataset. Then, functional and contextual networks are constructed to explore the correlation among the data sources, enabling fusion learning of multi-source data features. We designed two sets of simulation experiments based on datasets containing CFD and wind tunnel test data, as well as CFD and flight test data. The experimental results show that after data feature fusion using this model, the accuracy of the datasets improved by an average of 30.2% and 67.5%; compared with traditional data fusion methods, the uncertainty was reduced by 23.9% and 27.6%, respectively.

参考文献

1 唐志共, 袁先旭, 钱炜祺, 等. 高速空气动力学三大手段数据融合研究进展[J]. 空气动力学学报202341(8): 44-58.
  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 Sinica202341(8): 44-58 (in Chinese).
2 邓晨, 陈功, 王文正, 等. 基于不确定度和气动模型的气动数据融合算法[J]. 空气动力学学报202240(4): 117-123.
  DENG C, CHEN G, WANG W Z, et al. Aerodynamic data fusion algorithms based on aerodynamic model and uncertainly?[J]. Acta Aerodynamica Sinica202240(4): 117-123 (in Chinese).
3 POLOCZEK M, WANG J, FRAZIER P I. Multi-information source optimization?[DB/OL]. arXiv preprint: 1603.00389v2, 2017.
4 HE L, QIAN W Q, ZHAO T, et al. Multi-fidelity aerodynamic data fusion with a deep neural network modeling method[J]. Entropy202022(9): 1022.
5 唐志共, 朱林阳, 向星皓, 等. 智能空气动力学若干研究进展及展望[J]. 空气动力学学报202341(7): 1-35.
  TANG Z G, ZHU L Y, XIANG X H, et al. Some research progress and prospect of intelligent aerodynamics[J]. Acta Aerodynamica Sinica202341(7): 1-35 (in Chinese).
6 LIANG X Y, QIAN Y H, GUO Q, et al. AF: An association-based fusion method for multi-modal classification?[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence202244(12): 9236-9254.
7 ZHANG Z J, DURAISAMY K. Machine learning methods for data-driven turbulence modeling?[C]?∥22nd AIAA Computational Fluid Dynamics Conference. Reston: AIAA, 2015.
8 LECUN Y, BENGIO Y, HINTON G. Deep learning[J]. Nature2015521(7553): 436-444.
9 HU L W, ZHANG J, XIANG Y, et al. Neural networks-based aerodynamic data modeling: A comprehensive review?[J]. IEEE Access20208: 90805-90823.
10 MARCATO A, BOCCARDO G, MARCHISIO D L. A computational workflow to study particle transport in porous media: Coupling CFD and deep learning?[M]?∥30th European Symposium on Computer Aided Process Engineering. Amsterdam: Elsevier, 2020: 1759-1764.
11 PARISH E J, DURAISAMY K. A paradigm for data-driven predictive modeling using field inversion and machine learning?[J]. Journal of Computational Physics2016305: 758-774.
12 LI J H, LI B, XU J Z, et al. Fully connected network-based intra prediction for image coding[J]. IEEE Transactions on Image Processing201827(7): 3236-3247.
13 BELYAEV M, BURNAEV E, KAPUSHEV E, et al. Building data fusion surrogate models for spacecraft aerodynamic problems with incomplete factorial design of experiments?[J]. Advanced Materials Research20141016: 405-412.
14 王文正, 桂业伟, 何开锋, 等. 基于数学模型的气动力数据融合研究[J]. 空气动力学学报200927(5): 524-528.
  WANG W Z, GUI Y W, HE K F, et al. Aerodynamic data fusion technique exploration[J]. Acta Aerodynamica Sinica200927(5): 524-528 (in Chinese).
15 张骏, 张广博, 程艳青, 等. 一种气动大差异性数据多任务学习方法[J]. 空气动力学学报202240(6): 64-72.
  ZHANG J, ZHANG G B, CHENG Y Q, et al. A multi-task learning method for large discrepant aerodynamic data[J]. Acta Aerodynamica Sinica202240(6): 64-72 (in Chinese).
16 赵旋, 彭绪浩, 邓子辰, 等. 基于多源数据融合的翼型表面压强精细化重构方法[J]. 实验流体力学202236(3): 93-101.
  ZHAO X, PENG X H, DENG Z C, et al. Fine reconstruction method of airfoil surface pressure based on multi-source data fusion[J]. Journal of Experiments in Fluid Mechanics202236(3): 93-101 (in Chinese).
17 王旭, 宁晨伽, 王文正, 等. 面向飞行试验的多源气动数据智能融合方法[J]. 空气动力学学报202341(2): 12-20.
  WANG X, NING C J, WANG W Z, et al. Intelligent fusion method of multi-source aerodynamic data for flight tests[J]. Acta Aerodynamica Sinica202341(2): 12-20 (in Chinese).
18 LI K, KOU J Q, ZHANG W W. Deep learning for multifidelity aerodynamic distribution modeling from experimental and simulation data[J]. AIAA Journal202260(7): 4413-4427.
19 HU L W, XIANG Y, ZHANG J, et al. Aerodynamic data predictions based on multi-task learning[J]. Applied Soft Computing2022116: 108369.
20 GIROSI F, POGGIO T. Networks and the best approximation property[J]. Biological Cybernetics199063(3): 169-176.
21 POGGIO T, GIROSI F. Networks for approximation and learning[J]. Proceedings of the IEEE199078(9): 1481-1497.
22 HU L W, WANG W Y, XIANG Y, et al. Flow field reconstructions with GANs based on radial basis functions[J]. IEEE Transactions on Aerospace and Electronic Systems202258(4): 3460-3476.
23 GOODFELLOW I J, POUGET-ABADIE J, MIRZA M, et al. Generative adversarial nets?[J]. Advances in neural information processing systems2014, 27.
24 FASSHAUER G E, YE Q. Reproducing kernels of generalized Sobolev spaces via a Green function approach with distributional operators?[J]. Numerische Mathematik2011119: 585-611.
文章导航

/