Lattice Boltzmann method based on deep neural network

CHEN Xinyang; NIE Zisen; JIANG Zichao; YANG Gengchao; YAO Qinghe

doi:10.13471/j.cnki.acta.snus.2020.07.09.2020B076

您当前的位置：

首页 >

文章列表页 >

Lattice Boltzmann method based on deep neural network

Research articles | 更新时间：2023-11-01

- Lattice Boltzmann method based on deep neural network
- Acta Scientiarum Naturalium Universitatis Sunyatseni Vol. 60, Issue 5, Pages: 39-49(2021)
- 作者机构：
  
  中山大学航空航天学院，广东广州 510006
- 作者简介：
- 基金信息：
- DOI：10.13471/j.cnki.acta.snus.2020.07.09.2020B076
  CLC： TP183
- Published：25 September 2021，
  
  Published Online：11 November 2020，
  
  Received：09 July 2020，
  
  Accepted：24 September 2020
扫描看全文
陈辛阳,聂滋森,蒋子超等.基于深度神经网络的格子玻尔兹曼算法[J].中山大学学报(自然科学版),2021,60(05):39-49.

CHEN Xinyang,NIE Zisen,JIANG Zichao,et al.Lattice Boltzmann method based on deep neural network[J].Acta Scientiarum Naturalium Universitatis Sunyatseni,2021,60(05):39-49.
陈辛阳,聂滋森,蒋子超等.基于深度神经网络的格子玻尔兹曼算法[J].中山大学学报(自然科学版),2021,60(05):39-49. DOI： 10.13471/j.cnki.acta.snus.2020.07.09.2020B076.

CHEN Xinyang,NIE Zisen,JIANG Zichao,et al.Lattice Boltzmann method based on deep neural network[J].Acta Scientiarum Naturalium Universitatis Sunyatseni,2021,60(05):39-49. DOI： 10.13471/j.cnki.acta.snus.2020.07.09.2020B076.

摘要

格子玻尔兹曼算法（LBM，Lattice Boltzmann method）相较于传统计算流体力学方法具有程序结构简单，对复杂边界和非线性问题适应性强以及便于并行计算等诸多优点。然而，其作为一种显式算法，在计算过程中的迭代次数较多，进而消耗大量计算资源。利用深度学习在预测与回归方面的强大能力，基于LBM设计了一个由卷积层与卷积长短期记忆层组成的人工神经网络预测模型并将其命名为C-LBM（compressed LBM）。该模型能等效替代多个普通LBM迭代。对于方腔环流问题，模型完成训练后，对测试集均方差在

https://html.publish.founderss.cn/rc-pub/api/common/picture?pictureId=49469718&type=

https://html.publish.founderss.cn/rc-pub/api/common/picture?pictureId=49469696&type=

10.66800022

2.53999996

以下，对泛化算例在

https://html.publish.founderss.cn/rc-pub/api/common/picture?pictureId=49469744&type=

https://html.publish.founderss.cn/rc-pub/api/common/picture?pictureId=49469722&type=

5.41866684

2.53999996

以下，精度得到了保障。相较于串行LBM程序，C-LBM模型计算效率提升约15倍。

Abstract

Compared to the traditional computational fluid dynamics techniques， the Lattice Boltzmann method has the advantages of simple structure of program， strong adaptability to complex boundaries as well as nonlinear problems， and high parallelism. However， since LBM is an explicit algorithm， its calculation usually involves many iteration steps， and thereby consumes a huge amount of computing resources. This study takes advantage of deep learning in prediction and regression to accelerate LBM calculations innovatively. We establish a prediction model （compressed LBM or C-LBM）， which involves an artificial neural network composed of convolution layers and convolution long-term and short-term memory layers. The prediction model is an equivalent substitution of multiple ordinary LBM iterations. For the two dimensional driven cavity circulation problem， the mean square error of C-LBM is less than

https://html.publish.founderss.cn/rc-pub/api/common/picture?pictureId=49469754&type=

https://html.publish.founderss.cn/rc-pub/api/common/picture?pictureId=49469763&type=

12.19200039

2.96333337

for the test set， and is less than

https://html.publish.founderss.cn/rc-pub/api/common/picture?pictureId=49469806&type=

https://html.publish.founderss.cn/rc-pub/api/common/picture?pictureId=49469794&type=

6.01133299

2.96333337

for the generalized calculation example. The calculation efficiency of C-LBM is about 15 times higher than that of the serial LBM.

关键词

格子玻尔兹曼算法数据驱动模型深度学习算法加速

Keywords

Lattice Boltzmann methoddata-driven modeldeep learningalgorithm acceleration

references

ZHANG W， YANG Y， XU J. Theory and application of Lattice Boltzmann method［J］. Machinery Modern， 2003，4：4-6.

FAN Z， QIU F， KAUFMAN A，et al. GPU cluster for high performance computing［C］// The 2004 ACM/IEEE Conference on Supercomputing， 2004：47.

WANG Y， HE Y， ZHAO T， et al. Implicit-Explicit finite-difference Lattice Boltzmann method for compressible flows［J］. International Journal of Modern Physics C， 2007， 18（12）：1961-1983.

XIE X， MOHEBUJJAMAN M， REBHOLZ L G， et al. Data-driven filtered reduced order modeling of fluid flows， SIAM ［J］.Journal on Scientific Computing， 2017，40（3）：B834-B857.

WITT T D， LESSIG C， FIUME E. Fluid simulation using laplacian eigenfunctions ［J］.ACM Transactions on Graphics ， 2012，31（1）： 1-11.

SILVA W A， BARTELS R E . Development of reduced-order models for aeroelastic analysis and flutter prediction using the CFL3Dv6.0 Code ［J］.Journal of Fluids and Structures， 2004，19（6）： 729-745.

VAROL Y， OZTO P，AVCI E，et al. Estimation of thermal and flow fields due to natural convection using support vector machines （SVM） in a porous cavity with discrete heat sources ［J］. Letters in Heat and Mass Transfer， 2008，35（8）：928-936.

LADICKÝ L， JEONG S， SOLENTHALER B，et al. Data-driven fluid simulations using regression forests ［J］. ACM Transactions on Graphics， 2015， 34（6）： 1-9.

SCHMIDHUBER J. Deep learning in neural networks： An overview ［J］. Neural Networks， 2014，61：85-117.

GUO X， LI W， IORIO F. Convolutional neural networks for steady flow approximation ［C］// The 22nd ACM SIGKDD. International Conference on Knowledge Discovery and Data Mining， 2016：481-490.

JONATHAN T， KRISTOFER S， PABLO S，et al. Accelerating eulerian fluid simulation with convolutional networks ［C］// International Conference on Machine Learning， 2017：3424-3433.

OLIVER H. Lat-Net： Compressing Lattice Boltzmann flow simulations using deep neural networks［J/OL］. （2017-05-25）［2020-09-29］.http：//arxiv.org/abs/1705.09036http://arxiv.org/abs/1705.09036.

GOODFELLOW I， POUGET-ABADIE J， MIRZA M， et al. Generative adversarial nets ［C］// The 27th International Conference on Neural Information Processing Systems， 2014： 2672-2680.

GREGOR K， DANIHELKAI I， GRAVES A， et al.DRAW： A recurrent neural network for image generation ［J/OL］.（2015-04-02）［2020-09-29］.https：//arxiv.org/abs/1502.04623https://arxiv.org/abs/1502.04623.

JIANWEI Y， ANITHA K， DHRUV B， et al. LR-GAN： Layered recursive generative adversarial networks for image generation［J/OL］.（2017-04-02）［2020-09-29］.http：//arxiv.org/abs/1703.01560http://arxiv.org/abs/1703.01560.

SHI X， CHEN Z， WANG H， et al. Convolutional LSTM Network： A machine learning approach for precipitation nowcasting ［J/OL］. （2015-09-19）［2020-09-29］.http：//arxiv.org/abs/1506.04214http://arxiv.org/abs/1506.04214.

HE K， ZHANG X， REN S， et al.Deep residual learning for image recognition ［C］// IEEE Conference on Computer Vision & Pattern Recognition， 2016： 770-778.

QIAN Y-H，DHUMIERES D， LALLEMAND P. Lattice BGK model for Navier-Stokes equation ［J］. Europhysics Letters，1992， 17 （6）： 479-484.

GUO Z， ZHENG C， Theory and applications of Lattice Boltzmann method ［M］.Beijing： Press Science， 2009.

RUMELHART D E， MaClelland J L. Learning internal representations by error propagation ［M］. Parallel Distributed Processing： Explorations in the Microstructure of Cognition： Foundations. MIT Press， 1987：318-362.

DIEDERIK P K， JIMMY B. Adam： a method for stochastic optimization ［J/OL］.（2017-01-30）［2020-09-29］. https：//arxiv.org/abs/1412.6980https://arxiv.org/abs/1412.6980.

SHANKAR P N， DESHPANDE M D. Fluid mechanics in the driven cavity ［J］. Annual Review of Fluid Mechanics， 2000，32 （1）： 93-136.

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Automatic detection of newly increased construction land based on deep learning and multi-source remote sensing data

Style transfer based on cross-layer correlation perceptual loss

Kernel principal component analysis network method for face recognition

Overview of person re-identification in unconstrained environments

Related Author

Zisen NIE

Zichao JIANG

Gengchao YANG

Qinghe YAO

Xinyang CHEN

LUO Weiling

ZHANG Honghui

ZHANG Zerui

Related Institution

School of Geographic Science and Planning， Sun Yat-sen University

Guangdong Guodi Planning Science Technology Co.， Ltd.

School of Mathematics， Sun Yat-sen University

Guangdong Province Key Laboratory of Computational Science， Sun Yat-sen University

School of Electronics and Information Engineering, Sun Yat-sen University

Postal code：510275
Tel：020-84112585，84113223 Email：xuebaozr@mail.sysu.edu.cn
Technical support is provided by Beijing Founder electronics co., LTD 京ICP备09064830号-19 京公网安备11010802024621
It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.

⁰