Generative modeling of turbulence

被引:25
|
作者
Drygala, C. [1 ,2 ]
Winhart, B. [3 ]
di Mare, F. [3 ]
Gottschalk, H. [1 ,2 ]
机构
[1] Univ Wuppertal, Sch Math & Nat Sci, IMACM, Wuppertal, Germany
[2] Univ Wuppertal, Sch Math & Nat Sci, IZMD, Wuppertal, Germany
[3] Ruhr Univ Bochum, Dept Mech Engn, Chair Thermal Turbomachines & Aero Engines, Bochum, Germany
关键词
CIRCULAR-CYLINDER; MACHINE; FLOWS; PROOF; FIELD;
D O I
10.1063/5.0082562
中图分类号
O3 [力学];
学科分类号
08 ; 0801 ;
摘要
We present a mathematically well-founded approach for the synthetic modeling of turbulent flows using generative adversarial networks (GAN). Based on the analysis of chaotic, deterministic systems in terms of ergodicity, we outline a mathematical proof that GAN can actually learn to sample state snapshots from the invariant measure of the chaotic system. Based on this analysis, we study a hierarchy of chaotic systems starting with the Lorenz attractor and then carry on to the modeling of turbulent flows with GAN. As training data, we use fields of velocity fluctuations obtained from large-eddy simulations (LES). Two architectures are investigated in detail: we use a deep, convolutional GAN (DCGAN) to synthesize the turbulent flow around a cylinder. We furthermore simulate the flow around a low-pressure turbine stator using the pix2pixHD architecture for a conditional DCGAN being conditioned on the position of a rotating wake in front of the stator. The settings of adversarial training and the effects of using specific GAN architectures are explained. We thereby show that GAN are efficient in simulating turbulence in technically challenging flow problems on the basis of a moderate amount of training data. GAN training and inference times significantly fall short when compared with classical numerical methods, in particular, LES, while still providing turbulent flows in high resolution. We furthermore analyze the statistical properties of the synthesized and LES flow fields, which agree excellently. We also show the ability of the conditional GAN to generalize over changes of geometry by generating turbulent flow fields for positions of the wake that are not included in the training data.
引用
收藏
页数:19
相关论文
共 50 条
  • [1] Efficient channel modeling of structured light in turbulence using generative adversarial networks
    Briantcev, Dmitrii
    Cox, Mitchell A.
    Trichili, Abderrahmen
    Drozdov, Alice, V
    Ooi, Boon S.
    Alouini, Mohamed-Slim
    OPTICS EXPRESS, 2022, 30 (05): : 7238 - 7252
  • [2] Synthetic Lagrangian turbulence by generative diffusion models
    Li, T.
    Biferale, L.
    Bonaccorso, F.
    Scarpolini, M. A.
    Buzzicotti, M.
    NATURE MACHINE INTELLIGENCE, 2024, 6 (04) : 393 - 403
  • [3] TURBULENCE MODELING
    LUMLEY, JL
    JOURNAL OF APPLIED MECHANICS-TRANSACTIONS OF THE ASME, 1983, 50 (4B): : 1097 - 1103
  • [4] TURBULENCE MODELING
    HAINES, AB
    AERONAUTICAL JOURNAL, 1982, 86 (857): : 269 - 277
  • [5] TURBULENCE MODELING
    LI, GY
    PROCEEDINGS OF THE INSTITUTION OF CIVIL ENGINEERS-WATER MARITIME AND ENERGY, 1992, 96 (03): : 187 - 192
  • [6] Turbulence modeling
    Ravachol, M
    COMPUTATIONAL SCIENCE FOR THE 21ST CENTURY, 1997, : 380 - 391
  • [7] Generative Modeling by PRISM
    Sato, Taisuke
    LOGIC PROGRAMMING, 2009, 5649 : 24 - 35
  • [8] Lifelong generative modeling
    Ramapuram, Jason
    Gregorova, Magda
    Kalousis, Alexandros
    NEUROCOMPUTING, 2020, 404 : 381 - 400
  • [9] Generative modeling of genomes
    Tang, Lin
    NATURE METHODS, 2025, 22 (01) : 7 - 7
  • [10] Developed Turbulence: New Methods for Turbulence Modeling
    A. M. Mukhamedov
    Physical Mesomechanics, 2019, 22 : 181 - 187