耦合聚类的数据驱动稀薄流非线性本构计算方法
收稿日期: 2022-06-30
修回日期: 2022-07-27
录用日期: 2022-08-29
网络出版日期: 2022-09-22
基金资助
国家自然科学基金(U20B2007);国家数值风洞项目(NNW2019ZT3-A08);中央高校基本科研业务费(226-2022-00172)
Numerical method of data-driven rarefied nonlinear constitutive relations coupled with clustering
Received date: 2022-06-30
Revised date: 2022-07-27
Accepted date: 2022-08-29
Online published: 2022-09-22
Supported by
National Natural Science Foundation of China(U20B2007);National Numerical Wind Tunnel Project(NNW2019ZT3-A08);Fundamental Research Funds for the Central Universities(226-2022-00172)
临近空间环境下的稀薄流动具有典型非平衡特征，现有数值模拟方法在面对多尺度共存的复杂流动现象时难以同时高效、准确描述。结合近些年兴起的机器学习理论与建模方法，基于数据驱动非线性本构关系（DNCR）为稀薄非平衡流动快速求解提供了一种全新的物理建模思路。为进一步提升DNCR方法的泛化性能、模型训练与计算效率，本文在DNCR方法中首次引入高斯混合模型（GMM）与稀疏主成分分析（SPCA）在回归模型训练预测前对训练集与预测集数据进行聚类预处理，建立了耦合聚类模型的数据驱动非线性本构计算方法。相关算例预测结果表明：在DNCR可解释性方面，GMM与SPCA方法可以不依靠人工经验通过流场数据提取出不同流动中的主导物理量，提升了DNCR方法的可解释性与鲁棒性；在计算效率方面，GMM与SPCA方法能在大量流场样本下对数据点进行精准聚类，排除冗余信息的干扰，提高了方法的并行计算效率，降低了回归模型的训练预测时间，同时有利于后期添加新样本数据时模型的更新与维护；在计算精度方面，GMM与SPCA方法在简单特征流动预测中能够在保持现有预示精度的前提下提高计算效率，并在面对复杂流动特征耦合预测场景时有效提高DNCR方法预测精度。
尧少波 , 蒋励剑 , 赵文文 , 卢铮 , 吴昌聚 , 陈伟芳 . 耦合聚类的数据驱动稀薄流非线性本构计算方法[J]. 航空学报, 2022 , 43(S2) : 40 -53 . DOI: 10.7527/S1000-6893.2022.27708
Owing to the rarefied nonequilibrium effect encountered in near space environment， the existing numerical methods are difficult to describe the multi-scale flow phenomena efficiently and accurately at the same time. By integrating the machine learning methods proposed in recent years， Data-driven Nonlinear Constitutive Relations （DNCR） presented a new data-driven modeling approach for solving the nonequilibrium problem. To further improve the generalization capability and model training efficiency of the DNCR method， Gaussian Mixture Model （GMM） and Sparse Principal Component Analysis （SPCA） for preprocessing the data of training set and prediction set are introduced to the DNCR method for the first time in this paper. The prediction results of relevant cases show that the dominant parameters in different flow fields are extracted by GMM and SPCA without relying on artificial experience， which could improve the interpretability and robustness of DNCR. On the other hand， GMM and SPCA can accurately cluster the data points under a large number of flow field samples to eliminate the interference of redundant information and reduce the training and prediction time of the regression model， which is crucial to the updating and maintenance of the model when adding new sample data in future. For prediction accuracy， the coupled DNCR could improve the computational efficiency without losing accuracy in simple flows， and could further elevate the precision significantly in complex flows coupled with different flow characteristics.
| 1 | TSIEN H S. Superaerodynamics, mechanics of rarefied gases[J]. Journal of the Aeronautical Sciences, 1946, 13(12): 653-664. |
| 2 | BHATNAGAR P L, GROSS E P, KROOK M. A model for collision processes in Gases. I. Small amplitude processes in charged and neutral one-component systems[J]. Physical Review, 1954, 94(3): 511-525. |
| 3 | XU K, HUANG J C. A unified gas-kinetic scheme for continuum and rarefied flows[J]. AIP Conference Proceedings, 2011, 1333(1): 525-530. |
| 4 | LIU S, YU P, XU K, et al. Unified gas-kinetic scheme for diatomic molecular simulations in all flow regimes[J]. Journal of Computational Physics, 2014, 259: 96-113. |
| 5 | 周恒, 张涵信. 空气动力学的新问题[J]. 中国科学: 物理学 力学 天文学, 2015, 45(10): 109-113. |
| ZHOU H, ZHANG H X. New problems of aerodynamics[J]. Scientia Sinica (Physica, Mechanica ＆ Astronomica), 2015, 45(10): 109-113 (in Chinese). | |
| 6 | 张伟伟, 寇家庆, 刘溢浪. 智能赋能流体力学展望[J]. 航空学报, 2021, 42(4): 524689. |
| ZHANG W W, KOU J Q, LIU Y L. Prospect of artificial intelligence empowered fluid mechanics[J]. Acta Aeronautica et Astronautica Sinica, 2021, 42(4): 524689 (in Chinese). | |
| 7 | WANG J X, WU J L, XIAO H. Physics informed machine learning approach for reconstructing Reynolds stress modeling discrepancies based on DNS data[J]. Physical Review Fluids, 2017, 2(3): 1-22. |
| 8 | RABAULT J, KUCHTA M, JENSEN A, et al. Artificial neural networks trained through deep reinforcement learning discover control strategies for active flow control[J]. Journal of Fluid Mechanics, 2019, 865: 281-302. |
| 9 | SEKAR V, KHOO B C. Fast flow field prediction over airfoils using deep learning approach[J]. Physics of Fluids, 2019, 31(5): 57103. |
| 10 | LI Z, KOVACHKI N, AZIZZADENESHELI K, et al. Fourier neural operator for parametric partial differential equations[DB/OL]. arXiv preprint: 2010.08895, 2020. |
| 11 | BAR-SINAI Y, HOYER S, HICKEY J, et al. Learning data-driven discretizations for partial differential equations[J]. Proceedings of the National Academy of Sciences of the United States of America, 2019, 116(31): 15344-15349. |
| 12 | KOCHKOV D, SMITH J A, ALIEVA A, et al. Machine learning-accelerated computational fluid dynamics[J]. Proceedings of the National Academy of Sciences of the United States of America, 2021, 118(21): e2101784118. |
| 13 | RAISSI M, PERDIKARIS P, KARNIADAKIS G E. Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations[J]. Journal of Computational Physics, 2019, 378: 686-707. |
| 14 | HU L, XIANG Y, ZHAN J, et al. Aerodynamic data predictions based on multi-task learning[J]. Applied Soft Computing, 2021, 116: 108369. |
| 15 | ZHANG J, MA W. Data-driven discovery of governing equations for fluid dynamics based on molecular simulation[J]. Journal of Fluid Mechanics, 2020, 892: A5. |
| 16 | XING H Y, ZHANG J, MA W J, et al. Using gene expression programming to discover macroscopic governing equations hidden in the data of molecular simulations[J]. Physics of Fluids, 2022, 34(5): 057109. |
| 17 | ZHAO W W, JIANG L J, YAO S B, et al. Data-driven nonlinear constitutive relations for rarefied flow computations[J]. Advances in Aerodynamics, 2021, 3(1): 540-558. |
| 18 | 李廷伟, 张莽, 赵文文, 等. 面向稀薄流非线性本构预测的机器学习方法[J]. 航空学报, 2021, 42(4): 524386. |
| LI T W, ZHANG M, ZHAO W W, et al. Machine learning method for correction of rarefied nonlinear constitutive relations[J]. Acta Aeronautica et Astronautica Sinica, 2021, 42(4): 524386 (in Chinese). | |
| 19 | 蒋励剑,赵文文,陈伟芳,等. 旋转不变的数据驱动稀薄非线性本构计算方法[J/OL].航空学报, (2021-10-14)[2022-06-30]. . |
| JIANG L J, ZHAO W W, CHEN W F, et al. Data-driven rarefied nonlinear constitutive relations based on rotation in-variants[J/OL]. Acta Aeronautica et As-tronautica Sinica, (2021-10-14)[2022-06-30]. . | |
| 20 | SAFAVIAN S R, LANDGREBE D. A survey of decision tree classifier methodology[J]. IEEE Transactions on Systems, Man, and Cybernetics, 1991, 21(3): 660-674. |
| 21 | 孙吉贵, 刘杰, 赵连宇. 聚类算法研究[J]. 软件学报, 2008, 19(1): 48-61. |
| SUN J G, LIU J, ZHAO L Y. Clustering algorithms research[J]. Journal of Software, 2008, 19(1): 48-61 (in Chinese). | |
| 22 | 贺玲, 吴玲达, 蔡益朝. 数据挖掘中的聚类算法综述[J]. 计算机应用研究, 2007, 24(1): 10-13. |
| HE L, WU L D, CAI Y C. Survey of clustering algorithms in data mining[J]. Application Research of Computers, 2007, 24(1): 10-13 (in Chinese). | |
| 23 | SUNG H G. Gaussian mixture regression and classification[D]. Houston: Rice University, 2004: 23-47. |
| 24 | ZOU H, HASTIE T, TIBSHIRANI R. Sparse principal component analysis[J]. Journal of Computational and Graphical Statistics, 2006, 15(2): 265-286. |
| 25 | TSIEN H S. Superaerodynamics, mechanics of rarefied gases[M]. Collected Works of H.S. Tsien (1938-1956). Amsterdam: Elsevier, 2012: 406-429. |
| 26 | HARTIGAN J A, WONG M A. Algorithm AS 136: A K-means clustering algorithm[J]. Applied Statistics, 1979, 28(1): 100. |
| 27 | WOLD S, ESBENSEN K, GELADI P. Principal component analysis[J]. Chemometrics and Intelligent Laboratory Systems, 1987, 2(1-3): 37-52. |
| 28 | CALLAHAM J L, KOCH J V, BRUNTON B W, et al. Learning dominant physical processes with data-driven balance models[J]. Nature Communications, 2021, 12: 1016. |
| 29 | GEURTS P, ERNST D, WEHENKEL L. Extremely randomized trees[J]. Machine Learning, 2006, 63(1): 3-42. |
/
| 〈 |
|
〉 |
CJA