FlowDNN：一种用于快速精确流场预测的物理启发深度神经网络

陈东林; 高翔; 徐传福; 王思齐; 陈世钊; 方建滨; 王铮

doi:10.1631/FITEE.2000435

Your Location：

Home >

Browse articles >

FlowDNN：一种用于快速精确流场预测的物理启发深度神经网络

常规文章 | Updated：2022-04-19

- FlowDNN：一种用于快速精确流场预测的物理启发深度神经网络
  Enhanced Publication
- FlowDNN: a physics-informed deep neural network for fast and accurate flow prediction
- 信息与电子工程前沿（英文） 2022年23卷第2期页码：207-219
- Affiliations：
  
  1.College of Computer, National University of Defense Technology, Changsha410073, China
  2.State Key Laboratory of High Performance Computing, National University of Defense Technology, Changsha410073, China
  3.School of Computing, University of Leeds, LeedsLS29JT, UK
- Author bio：
  
  E-mail: chendonglin14@nudt.edu.cn;
  E-mail: gaoxiang12@nudt.edu.cn;
  ‡Corresponding author
- Funds：
  
  National Natural Science Foundation of China(61772542;61972408;12102467);Foundation of the State Key Laboratory of High Performance Computing, China(201901-11;202001-03)
- DOI：10.1631/FITEE.2000435
  中图分类号： TP391
- 收稿：2020-08-28，
  
  录用：2021-05-04，
  
  纸质出版：2022-02-0
- Accepted：
Scan QR Code
陈东林, 高翔, 徐传福, 等. FlowDNN：一种用于快速精确流场预测的物理启发深度神经网络[J]. 信息与电子工程前沿（英文）, 2022,23(2):207-219.

Donglin CHEN, Xiang GAO, Chuanfu XU, et al. FlowDNN: a physics-informed deep neural network for fast and accurate flow prediction[J]. Frontiers of Information Technology & Electronic Engineering, 2022, 23(2): 207-219.
陈东林, 高翔, 徐传福, 等. FlowDNN：一种用于快速精确流场预测的物理启发深度神经网络[J]. 信息与电子工程前沿（英文）, 2022,23(2):207-219. DOI： 10.1631/FITEE.2000435.

Donglin CHEN, Xiang GAO, Chuanfu XU, et al. FlowDNN: a physics-informed deep neural network for fast and accurate flow prediction[J]. Frontiers of Information Technology & Electronic Engineering, 2022, 23(2): 207-219. DOI： 10.1631/FITEE.2000435.

摘要

对于与流场相关的设计优化问题，例如飞机和汽车空气动力学设计，计算流体力学（CFD）模拟通常用于预测流场并分析性能。虽然CFD模拟十分重要，但它的迭代计算非常需要计算资源且极其耗时。昂贵的模拟开销限制了大范围设计空间的探索，并阻碍了实时的交互式设计。在本文中，我们提出FlowDNN模型，它是一种新颖的深度神经网络，可从CFD结果中高效地学习流场表示。FlowDNN根据给定的流动条件和几何形状可以直接预测预期的流场结果，从而极大地节省计算时间。FlowDNN首次结合了流体力学的基本守恒定律和注意力机制进行定常流场预测。这样做不仅可以提高预测准确性，而且可以维持预测流场的物理一致性，这对于CFD模拟至关重要。本文设计了多种指标以评估FlowDNN预测的整体流场和关键区域的结果（如流场快速变化的边界层）。实验结果表明，FlowDNN明显优于其他方法且具有更短的推理时间和更准确的结果。它与最新的GPU并行求解器相比，生成流场的速度提升14 000倍以上，同时保持预测误差在5%以内。

Abstract

For flow-related design optimization problems

e.g.

aircraft and automobile aerodynamic design

computational fluid dynamics (CFD) simulations are commonly used to predict flow fields and analyze performance. While important

CFD simulations are a resource-demanding and time-consuming iterative process. The expensive simulation overhead limits the opportunities for large design space exploration and prevents interactive design. In this paper

we propose FlowDNN

a novel deep neural network (DNN) to efficiently learn flow representations from CFD results. FlowDNN saves computational time by directly predicting the expected flow fields based on given flow conditions and geometry shapes. FlowDNN is the first DNN that incorporates the underlying physical conservation laws of fluid dynamics with a carefully designed attention mechanism for steady flow prediction. This approach not only improves the prediction accuracy

but also preserves the physical consistency of the predicted flow fields

which is essential for CFD. Various metrics are derived to evaluate FlowDNN with respect to the whole flow fields or regions of interest (RoIs) (e.g.

boundary layers where flow quantities change rapidly). Experiments show that FlowDNN significantly outperforms alternative methods with faster inference and more accurate results. It speeds up a graphics processing unit (GPU) accelerated CFD solver by more than

while keeping the prediction error under 5%.

关键词

Keywords

references

Ahmed MYM , Qin N , 2009 . Surrogate-based aerodynamic design optimization: use of surrogates in aerodynamic design optimization . Int Conf on Aerospace Sciences & Aviation Technology , p. 1 - 26 . doi: 10.21608/ASAT.2009.23442 http://doi.org/10.21608/ASAT.2009.23442

Amodio M , Krishnaswamy S , 2019 . TraVeLGAN: image-to-image translation by transformation vector learning . IEEE/CVF Conf on Computer Vision and Pattern Recognition , p. 8975 - 8984 . doi: 10.1109/CVPR.2019.00919 http://doi.org/10.1109/CVPR.2019.00919

Balabanov VO , Giunta AA , Golovidov O , et al. , 1999 . Reasonable design space approach to response surface approximation . J Aircr , 36 ( 1 ): 308 - 315 . doi: 10.2514/2.2438 http://doi.org/10.2514/2.2438

Bhatnagar S , Afshar Y , Pan S , et al. , 2019 . Prediction of aerodynamic flow fields using convolutional neural networks . Comput Mech , 64 ( 2 ): 525 - 545 . doi: 10.1007/s00466-019-01740-0 http://doi.org/10.1007/s00466-019-01740-0

Blazek J , 2015 . Computational Fluid Dynamics: Principles and Applications ( 3 rd Ed. ). Butterworth-Heinemann , Oxford, UK , p. 466 .

Constantin P , Foias C , 1988 . Navier–Stokes Equations . The University of Chicago Press , Chicago, IL, USA , p. 199 .

Daberkow DD , Mavris DN , 1998 . New approaches to conceptual and preliminary aircraft design: a comparative assessment of a neural network formulation and a response surface methodology . World Aviation Congress & Exposition, Article 15 . doi: 10.4271/985509 http://doi.org/10.4271/985509

Ernst MH , 1981 . Nonlinear model-Boltzmann equations and exact solutions . Phys Rep , 78 ( 1 ): 1 - 171 . doi: 10.1016/0370-1573(81)90002-8 http://doi.org/10.1016/0370-1573(81)90002-8

Farrashkhalvat M , Miles JP , 2003 . Basic Structured Grid Generation: with an Introduction to Unstructured Grid Generation . Elsevier , Amsterdam, the Netherlands , p. 190 - 226 . doi: 10.1016/B978-075065058-8/50008-3 http://doi.org/10.1016/B978-075065058-8/50008-3

Frankle J , Carbin M , 2019 . The lottery ticket hypothesis: finding sparse, trainable neural networks . https://arxiv.org/abs/1803.03635v5 https://arxiv.org/abs/1803.03635v5

Geneva N , Zabaras N , 2019 . Quantifying model form uncertainty in Reynolds-averaged turbulence models with Bayesian deep neural networks . J Comput Phys , 383 : 125 - 147 . doi: 10.1016/j.jcp.2019.01.021 http://doi.org/10.1016/j.jcp.2019.01.021

Guastoni L , Guemes A , Ianiro A , et al. , 2020 . Convolutional-network models to predict wall-bounded turbulence from wall quantities . https://arxiv.org/abs/2006.12483 https://arxiv.org/abs/2006.12483

Guo XX , Li W , Iorio F , 2016 . Convolutional neural networks for steady flow approximation . Proc 22 nd ACM SIGKDD Int Conf on Knowledge Discovery and Data Mining , p. 481 - 490 . doi: 10.1145/2939672.2939738 http://doi.org/10.1145/2939672.2939738

Hamdan MKA , Rover DT , Darr MJ , et al. , 2019 . Mass estimation from images using deep neural network and sparse ground truth . http://arxiv.org/abs/1908.04387 http://arxiv.org/abs/1908.04387

Hu J , Shen L , Sun G , 2018 . Squeeze-and-excitation networks . IEEE/CVF Conf on Computer Vision and Pattern Recognition , p. 7132 - 7141 . doi: 10.1109/CVPR.2018.00745 http://doi.org/10.1109/CVPR.2018.00745

Isola P , Zhu JY , Zhou TH , et al. , 2017 . Image-to-image translation with conditional adversarial networks . IEEE Conf on Computer Vision and Pattern Recognition , p. 5967 - 5976 . doi: 10.1109/CVPR.2017.632 http://doi.org/10.1109/CVPR.2017.632

Kim T , Cha M , Kim H , et al. , 2017 . Learning to discover cross-domain relations with generative adversarial networks . Proc 34 th Int Conf on Machine Learning , p. 1857 - 1865 .

Lee S , You D , 2019 . Data-driven prediction of unsteady flow over a circular cylinder using deep learning . J Fluid Mech , 879 : 217 - 254 . doi: 10.1017/jfm.2019.700 http://doi.org/10.1017/jfm.2019.700

Li DL , Xu CF , Wang YX , et al. , 2016 . Parallelizing and optimizing large-scale 3D multi-phase flow simulations on the Tianhe-2 supercomputer . Concurr Comput , 28 ( 5 ): 1678 - 1692 . doi: 10.1002/cpe.3717 http://doi.org/10.1002/cpe.3717

Ling JL , Kurzawski A , Templeton J , 2016 . Reynolds averaged turbulence modelling using deep neural networks with embedded invariance . J Fluid Mech , 807 : 155 - 166 . doi: 10.1017/jfm.2016.615 http://doi.org/10.1017/jfm.2016.615

Liu Z , Sun MJ , Zhou TH , et al. , 2019 . Rethinking the value of network pruning . https://arxiv.org/abs/1810.05270 https://arxiv.org/abs/1810.05270

Long J , Shelhamer E , Darrell T , 2015 . Fully convolutional networks for semantic segmentation . IEEE Conf on Computer Vision and Pattern Recognition , p. 3431 - 3440 . doi: 10.1109/CVPR.2015.7298965 http://doi.org/10.1109/CVPR.2015.7298965

Molchanov P , Tyree S , Karras T , et al. , 2017 . Pruning convolutional neural networks for resource efficient inference . Int Conf on Learning Representations .

Odena A , Dumoulin V , Olah C , 2016 . Deconvolution and checkerboard artifacts . Distill , 1 ( 10 ): e3 . doi: 10.23915/distill.00003 http://doi.org/10.23915/distill.00003

Park J , Woo S , Lee JY , et al. , 2018 . BAM: bottleneck attention module . https://arxiv.org/abs/1807.06514v1 https://arxiv.org/abs/1807.06514v1

Raissi M , Perdikaris P , Karniadakis GE , 2017 . Physics informed deep learning (part I): data-driven solutions of nonlinear partial differential equations . https://arxiv.org/abs/1711.10561 https://arxiv.org/abs/1711.10561

Raissi M , Perdikaris P , Karniadakis GE , 2019 . Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations . J Comput Phys , 378 : 686 - 707 . doi: 10.1016/j.jcp.2018.10.045 http://doi.org/10.1016/j.jcp.2018.10.045

Ronneberger O , Fischer P , Brox T , 2015 . U-Net: convolutional networks for biomedical image segmentation . Medical Image Computing and Computer-Assisted Intervention , p. 234 - 241 . doi: 10.1007/978-3-319-24574-4_28 http://doi.org/10.1007/978-3-319-24574-4_28

Srinivasan PA , Guastoni L , Azizpour H , et al. , 2019 . Predictions of turbulent shear flows using deep neural networks . Phys Rev Fluids , 4 : 054603 . doi: 10.1103/PhysRevFluids.4.054603 http://doi.org/10.1103/PhysRevFluids.4.054603

Thuerey N , Weissenow K , Prantl L , et al. , 2020 . Deep learning methods for Reynolds-averaged Navier-ÍCStokes simulations of airfoil flows . AIAA J , 58 ( 1 ): 25 - 36 . doi: 10.2514/1.J058291 http://doi.org/10.2514/1.J058291

Wang R , Kashinath K , Mustafa M , et al. , 2020 . Towards physics-informed deep learning for turbulent flow prediction . Proc 26 th ACM SIGKDD Int Conf on Knowledge Discovery & Data Mining , p. 1457 - 1466 . doi: 10.1145/3394486.3403198 http://doi.org/10.1145/3394486.3403198

Woo S , Park J , Lee JY , et al. , 2018 . CBAM: convolutional block attention module . European Conf on Computer Vision , p. 3 - 9 . doi: 10.1007/978-3-030-01234-2_1 http://doi.org/10.1007/978-3-030-01234-2_1

Zhou ZW , Siddiquee MMR , Tajbakhsh N , et al. , 2018 . UNet++: a nested U-Net architecture for medical image segmentation . 4 th Int Workshop on Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support , p. 3 - 11 . doi: 10.1007/978-3-030-00889-5_1 http://doi.org/10.1007/978-3-030-00889-5_1

Zhu JY , Park T , Isola P , et al. , 2017 . Unpaired image-to-image translation using cycle-consistent adversarial networks . IEEE Int Conf on Computer Vision , p. 2242 - 2251 . doi: 10.1109/ICCV.2017.244 http://doi.org/10.1109/ICCV.2017.244

浏览量

Downloads

CSCD

文章被引用时，请邮件提醒。

Submit

工具集

关联资源

Associative affinity network learning for multi-object tracking

Deep learning compact binary codes for fingerprint indexing