基于机器学习的非结构网格阵面推进生成技术初探

王年华; 鲁鹏; 常兴华; 张来平

doi:10.6052/0459-1879-20-402

基于机器学习的非结构网格阵面推进生成技术初探

doi: 10.6052/0459-1879-20-402

王年华^*,2), ,,
鲁鹏^†,**,
常兴华^††,
张来平^††

^*中国空气动力研究与发展中心空气动力学国家重点实验室,四川绵阳 621000
^††西南科技大学信息工程学院,四川绵阳 621010
^**重庆文理学院智能制造工程学院,重庆 402160
^†††国防科技创新研究院无人系统技术研究中心,北京 100071

基金项目: 1) 国家重点研发计划(2016YFB0200701);空气动力学国家重点实验室创新基金(SKLA190104);国家重大专项(GJXM92579)

详细信息

作者简介:
2) 王年华,助理研究员,主要研究方向: 计算流体力学. E-mail: nhwang@skla.cardc.cn

通讯作者:
王年华

中图分类号: V211.3
计量
- 文章访问数: 2089
- HTML全文浏览量: 444
- PDF下载量: 715
- 被引次数: 0
出版历程
- 收稿日期: 2020-11-27
- 刊出日期: 2021-03-10

PRELIMINARY INVESTIGATION ON UNSTRUCTURED MESH GENERATION TECHNIQUE BASED ON ADVANCING FRONT METHOD AND MACHINE LEARNING METHODS

Wang Nianhua^{*,2)
, ,},
Lu Peng^†,**,
Chang Xinghua^††,
Zhang Laiping^††

^*State Key Laboratory of Aerodynamics,China Aerodynamics Research and Development Center,Mianyang 621000,Sichuan,China
^††School of Information Engineering,Southwest University of Science and Technology,Mianyang 621010,Sichuan,China
^**School of Intelligent Manufacturing Engineering,Chongqing University of Arts and Science,Chongqing 402160,China
^†††Unmanned Systems Research Center,National Innovation Institute of Defense Technology,Beijing 100071,China

摘要

摘要: 网格生成和自适应是制约计算流体力学未来发展的瓶颈问题之一,网格生成自动化和智能化仍是一个需要持续研究的领域.随着高性能计算算力的提升和大数据时代的到来,以机器学习为代表的人工智能方法已经成功应用于包括流体力学在内的多个领域,革命性地推动了这些领域的发展.本文首先简要综述机器学习方法在非结构网格生成领域的研究进展,分析基于机器学习进行非结构网格生成的关键问题;其次,设计非结构网格样本数据格式并实现了样本数据集的自动提取,通过结合人工神经网络和阵面推进法,初步发展了一种基于人工神经网络的二维非结构网格阵面推进生成方法;最后,采用新发展的方法生成了几个典型二维各向同性非结构三角形网格(二维圆柱、二维NACA0012翼型和30p30n三段翼型),进一步采用合并法生成了相应的三角形/四边形混合网格,并测试了网格质量和生成耗时,结果显示本文方法生成的网格质量可以达到商业软件的水平,且生成效率较传统阵面推进法提高30%.
- 机器学习 /
- 人工神经网络 /
- 计算流体力学 /
- 网格生成 /
- 阵面推进法
Abstract: Mesh generation and adaptation are bottleneck problems restricting future development of computational fluid dynamics (CFD). Automatic and intelligent mesh generation is still worth continuous investigation. With the rapid progress in high-performance computing power and big data technology, artificial intelligence, represented by machine learning, has been successfully applied to multiple fields including fluid dynamics, which has revolutionarily boosted the development of these fields. This paper reviews briefly the application of machine learning in the unstructured mesh generation in CFD and analyzes the key issues in the mesh generation based on machine learning. Meanwhile, the sample data format is designed and the automatic extraction of unstructured mesh sample data sets is realized. By integrating the advancing front (AFT) method with the artificial neural network, a novel two-dimensional triangular grid generation method is developed based on machine learning. Finally, several isotropic unstructured grids and hybrid grids (2D cylinder, 2D NACA0012 airfoil, and 30p30n three-element airfoil) are generated and mesh quality and elapsed time are counted, it indicates that mesh quality is generally equivalent to commercial software and the efficiency is 30% higher than the traditional AFT method.
- machine learning /
- artificial neural network /
- computational fluid dynamics /
- mesh generation /
- advancing front method

HTML全文

参考文献(53)

[1]	张涵信, 沈孟育. 计算流体力学—差分方法的原理和应用. 北京：国防工业出版社, 2003 (Zhang Hanxin, Computational Fluid Dynamics, Principles and Application of Finite Difference Methods. Beijing: National Defense Industry Press, 2003 (in Chinese))
[2]	Zhang SH, Li Q, Zhang LP, et al. The history of CFD in China. Acta Aerodynamica Sinica, 2016,34(2):157-174
[3]	Slotnick J, Khodadoust A, Alonso J, et al. CFD vision 2030 study: A path to revolutionary computational aerosciences. 2014, NASA/CR-2014-218178
[4]	Chawner JR, Taylor NJ. Progress in geometry modeling and mesh generation toward the CFD vision 2030. AIAA Aviation Forum, June, 2019, Texas
[5]	Baker TJ. Mesh generation: art or science. Progress in Aerospace Science, 2005,41(1):29-63
[6]	Roy CJ, Tinoco EN. Summary of data from the sixth AIAA CFD drag prediction workshop: case 1 code verification. AIAA Paper 2017-1206, 2017
[7]	Tinoco EN, Brodersen OP, Keye S, et al. Summary of data from the sixth AIAA CFD drag prediction workshop: CRM cases 2 to 5. AIAA Paper 2017-1208, 2017
[8]	Rumsey CL, Slotnick JP. Overview and summary of the second AIAA high-lift prediction workshop. Journal of Aircraft, 2015,52(4):1006-1025
[9]	Rumsey CL, Slotnick JP, Sclafani AJ. Overview and summary of the third AIAA high-lift prediction workshop. Journal of Aircraft, 2019,56(2):621-644
[10]	王运涛, 刘刚, 陈作斌. 第一届航空CFD可信度研讨会总结. 空气动力学学报, 2019,37(2):84-99 (Wang Yuntao, Liu Gang, Chen Zuobin. Summary of the first aeronautical computational fluid dynamics credibility workshop. Acta Aerodynamica Sinica, 2019,37(2):84-99 (in Chinese))
[11]	王年华, 常兴华, 赵钟, 等. 基于HyperFLOW平台的客机标模CHN-T1气动性能预测及可信度研究. 空气动力学学报, 2019,37(2):139-147 (Wang Nianhua, Chang Xinghua, Zhao Zhong, et al. Aerodynamic performance prediction and credibility study of the transport model CHN-T1 based on HyperFLOW solver. Acta Aerodynamica Sinica, 2019,37(2):139-147 (in Chinese))
[12]	张天娇, 钱炜祺, 周宇, 等. 人工智能与空气动力学结合的初步思考. 航空工程进展, 2019,10(1):1-11 (Zhang Tianjiao, Qian Weiqi, Zhou Yu, et al. Preliminary thoughts on the combination of artificial intelligence and aerodynamics. Advances in Aeronautical Science and Engineering, 2019,10(1):1-11 (in Chinese))
[13]	Kou JQ, Zhang WW. Multi-kernel neural networks for nonlinear unsteady aerodynamic reduced-order modeling. Aerospace Science and Technology, 2017,67:309-326
[14]	尹明朗, 寇家庆, 张伟伟, 等. 一种高泛化能力的神经网络气动力降阶模型. 空气动力学学报, 2017,35(2):205-213 (Yin Minglang, Kou Jiaqing, Zhang Weiwei, et al. A reduced-order aerodynamic model with high generalization capability based on neural network. Acta Aerodynamica Sinica, 2017,35(2):205-213 (in Chinese))
[15]	Kou JQ, Zhang WW. Layered reduced-order models for nonlinear aerodynamics and aeroelasticity. Journal of Fluids and Structures, 2017,68:174-193
[16]	蔡声泽, 许超, 高琪, 等. 基于深度神经网络的粒子图像测速算法. 空气动力学学报, 2019,37(3):455-461 (Cai Shengze, Xu Chao, Gao Qi, et al. Particle image velocimetry based on a deep neural network. Acta Aerodynamica Sinica, 2019,37(3):455-461 (in Chinese))
[17]	Deng L, Wang YQ, Liu Y, et al. A CNN-based vortex identification method. Journal of Visualization, 2019,22(1):65-78
[18]	Sekar V, Jiang QH, Shu C, et al. Fast flow field prediction over airfoils using deep learning approach. Physics of Fluids, 2019,31(5):057103
[19]	Duraisamy K, Iaccarino G, Xiao H. Turbulence modeling in the age of data. Annual Review of Fluid Mechanics, 2019,51:357-377
[20]	闫重阳, 张宇飞, 陈海昕. 基于离散伴随的流场反演在湍流模拟中的应用. 航空学报, 2021,42(4):624695 (Yan Chongyang, Zhang Yufei, Chen Haixin. Applications of field inversion based on discrete adjoint method in turbulence modeling. Acta Aeronautica et Astronautica Sinica, 2021,42(4):624695 (in Chinese))
[21]	Zhu LY, Zhang WW, Kou JQ, et al. Machine learning methods for turbulence modeling in subsonic flows around airfoils. Physics of Fluids, 2019,31(1):015105
[22]	Brunton SL, Noack BR, Koumoutsakos P. Machine learning for fluid dynamics. Annual Review of Fluid Mechanics, 2020,52(1):1-31
[23]	Lecun Y, Bengio Y, Hinton G. Deep learning. Nature, 2015,521(7553):436-444
[24]	张来平, 常兴华, 赵钟, 等. 计算流体力学网格生成技术. 北京: 科学出版社, 2017 (Zhang Laiping, Chang Xinghua, Zhao Zhong, et al. Mesh generation techniques in computational fluid dynamics. Beijing: Science Press, 2017 (in Chinese))
[25]	梁义, 陈建军, 陈立岗, 等. 并行平面 Delaunay 网格生成. 浙江大学学报 (工学版), 2008,42(4):558-564 (Liang Yi, Chen Jianjun, Chen Ligang, et al. Parallel planar Delaunay mesh generation. Journal of Zhejiang University (Engineering Science), 2008,42(4):558-564 (in Chinese))
[26]	陈建军, 黄争舸, 杨永健等. 复杂外形的非结构四面体网格生成算法. 空气动力学学报, 2010,28(4):400-404 (Chen Jianjun, Huang Zhengke, Yang Yongjian, et al. Unstructured tetrahedral mesh generation for complex configurations. Acta Aerodynamica Sinica, 2010,28(4):400-404 (in Chinese))
[27]	陈建军, 郑建靖, 季廷炜等. 前沿推进曲面四边形网格生成算法. 计算力学学报, 2011,28(5):779-784 (Chen Jianjun, Zheng Jianjing, Ji Tingwei, et al. Advancing front quadrilateral surface mesh generation. Chinese Journal of Computational Mechanics, 2011,28(5):779-784 (in Chinese))
[28]	陈建军, 郑耀, 陈立岗, 等. 非结构化四边形网格生成新算法. 中国图象图形学报, 2008,13(9):1796-1803 (Chen Jianjun, Zheng Yao, Chen Ligang, et al. A new unstructured quadrilateral mesh generation algorithm. Journal of Image and Graphics, 2008,13(9):1796-1803 (in Chinese))
[29]	陈建军, 曹建, 徐彦, 等. 适应复杂外形的三维黏性混合网格生成算法. 计算力学学报, 2014, ( 3):363-370 (Chen Jianjun, Cao Jian, Xu Yan, et al. Hybrid mesh generation algorithm for viscous computations of complex aerodynamics configurations. Chinese Journal of Computational Mechanics, 2014, ( 3):363-370 (in Chinese))
[30]	Steger OL, Chaussee DS. Generation of body-fitted coordinates using hyperbolic partial differential equations. Journal on Scientific and Statistical Computing, 1980,1(4):431-437
[31]	Matsuno K. Hyperbolic upwind method for prismatic grid generation. AIAA paper 2000-1003, 2000
[32]	赵钟, 张来平, 赫新. 基于“各向异性”四面体网格聚合的复杂外形混合网格生成方法. 空气动力学学报, 2013,31(1):34-39 (Zhao Zhong, Zhang Laiping, He Xin. Hybrid grid generation technique for complex geometries based on agglomeration of anisotropic tetrahedrons. Acta Aerodynamica Sinica, 2013,31(1):34-39 (in Chinese))
[33]	甘洋科, 刘剑飞. 黏性边界层网格自动生成. 力学学报, 2017,49(5):1029-1041 (Gan Keyang, Liu Jianfei. Automatic viscous boundary layer mesh generation. Chinese Journal of Theoretical and Applied Mechanics, 2017,49(5):1029-1041 (in Chinese))
[34]	Chawner JR, Michal TR, Slotnick JP, et al. Summary of the 1st AIAA geometry and mesh generation workshop (GMGW-1) and future plans// 2018 AIAA Aerospace Sciences Meeting, 2018:0128
[35]	Ahn CH, Lee SS, Lee HJ, et al. A self-organizing neural network approach for automatic mesh generation. IEEE Transactions on Magnetics, 1991,27(5):4201-4204
[36]	Alfonzetti S, Coco S, Cavalieri S, et al. Automatic mesh generation by the Let-It-Grow neural network. IEEE Transactions on Magnetics, 1996,32(3):1349-1352
[37]	Alfonzetti S. A finite element mesh generator based on an adaptive neural network. IEEE Transactions on Magnetics, 1998,34(5):3363-3366
[38]	Alfonzetti S. A neural network generator for tetrahedral meshes. IEEE Transactions on Magnetics, 2003,39(3):1650-1653
[39]	Alfonzetti S, Dilettoso E, Salerno N. An optimized generator of finite element meshes based on a neural network. IEEE Transactions on Magnetics, 2008,44(6):1278-1281
[40]	Lu HQ, Wu YZ, Chen SC. A new method based on SOM network to generate coarse meshes for overlapping unstructured multigrid algorithm. Applied Mathematics and Computation, 2003,140:353-360
[41]	陈先华. Let-It-Grow神经网络在网格剖分中的应用. [硕士论文]. 大连: 大连理工大学, 2008 (Chen Xianhua. Application of Let-It-Grow neural networks in triangulation. [Master Thesis]. Dalian: Dalian University of Technology, 2008 (in Chinese))
[42]	张伟, 姜献峰, 陈丽能, 等. 基于神经网络的三角形网格智能重建. 工程图学学报, 2004,1(1):71-75 (Zhang Wei, Jiang Xianfeng, Chen Lineng, et al. Intelligent reconstruction of triangle mesh based on neural network. Journal of Engineering Graphics, 2004,1(1):71-75 (in Chinese))
[43]	Jilani H, Bahreininejad A, Ahmadi MT. Adaptive finite element mesh triangulation using self-organizing neural networks. Advances in Engineering Software, 2009,40:1097-1103
[44]	杨占华, 杨艳. SOM神经网络算法的研究与进展. 计算机工程, 2006,32(16):201-202 (Yang Zhanhua, Yang Yan. Research and development of self-organizing maps algorithms. Computer Engineering, 2006,32(16):201-202 (in Chinese))
[45]	Lowther DA, Dyck DN. A density driven mesh generator guided by a neural network. IEEE Transactions on Magnetics, 1993,29(2):1927-1930
[46]	Chedid R, Najjar N. Automatic finite-element mesh generation using artificial neural networks-Part I: Prediction of mesh density. IEEE Transactions on Magnetics, 1996,32(5):5173-5178
[47]	Zhang Z, Wang Y, Jimack PK, et al. Meshing net: A new mesh generation method based on deep learning. International Conference on Computational Science, 2020, arXiv: 2004.07016
[48]	Yao S, Yan B, Chen B, et al. An ANN-based element extraction method for automatic mesh generation. Expert Systems with Applications, 2005,29(1):193-206
[49]	Danglade F, Pernot JP, Veron P. On the use of machine learning to defeature CAD models for simulation. Computer-Aided Design and Applications, CAD Solutions LLC and Taylor & Francis Online, 2013,11(3):358-368
[50]	Stadler D, Kosel F, Celic D, et al. Mesh deformation based on artificial neural networks. International Journal of Computational Fluid Dynamics, 2011,25(8):439-448
[51]	高翔. 非结构CFD并行网格变形算法及其应用. [博士论文]. 长沙: 国防科技大学, 2018 (Gao Xiang. Parallel unstructured mesh deformation algorithms and their applications in CFD. [PhD Thesis]. Changsha: National University of Defense Technology, 2018 (in Chinese))
[52]	黄尚坤. 基于卷积神经网络的网格质量判别技术研究. [硕士论文]. 绵阳: 西南科技大学, 2019 (Huang Shangkun. Research on grid quality discrimination technology based on convolutional neural network. [Master Thesis]. Mianyang: Southwest University of Science and Technology, 2019 (in Chinese))
[53]	Hagan MT, Menhaj M. Training feed-forward networks with the Marquardt algorithm. IEEE Transactions on Neural Networks, 1994,5(6):989-993