Advertisement

Towards Prediction of Turbulent Flows at High Reynolds Numbers Using High Performance Computing Data and Deep Learning

  • Mathis Bode
  • Michael GaudingEmail author
  • Jens Henrik Göbbert
  • Baohao Liao
  • Jenia Jitsev
  • Heinz Pitsch
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 11203)

Abstract

In this paper, deep learning (DL) methods are evaluated in the context of turbulent flows. Various generative adversarial networks (GANs) are discussed with respect to their suitability for understanding and modeling turbulence. Wasserstein GANs (WGANs) are then chosen to generate small-scale turbulence. Highly resolved direct numerical simulation (DNS) turbulent data is used for training the WGANs and the effect of network parameters, such as learning rate and loss function, is studied. Qualitatively good agreement between DNS input data and generated turbulent structures is shown. A quantitative statistical assessment of the predicted turbulent fields is performed.

Keywords

Turbulence High Reynolds number Deep learning Wasserstein generative adversarial networks Direct numerical simulation 

Notes

Acknowledgment

The authors gratefully acknowledge the computing time granted for the project JHPC55 by the JARA-HPC Vergabegremium and provided on the JARA-HPC Partition part of the supercomputer JURECA at Forschungszentrum Jülich. Also, the computing time granted for the projects HFG00/HFG02 on the supercomputer JUQUEEN [8] at Forschungszentrum Jülich is acknowledged. MG acknowledges financial support by Labex EMC3, under the grant VAVIDEN.

References

  1. 1.
    Arjovsky, M., Chintala, S., Bottou, L.: Wasserstein GAN. arXiv:1701.07875v3 (2017)
  2. 2.
    Boschung, J., Hennig, F., Gauding, M., Pitsch, H., Peters, N.: Generalised higher-order kolmogorov scales. J. Fluid Mech. 794, 233–251 (2016)MathSciNetCrossRefGoogle Scholar
  3. 3.
    Frisch, U.: Turbulence - The Legacy of A.N. Kolmogorov. Cambridge University Press, Cambridge (1995)CrossRefGoogle Scholar
  4. 4.
    Gauding, M., Danaila, L., Varea, E.: High-order structure functions for passive scalar fed by a mean gradient. Int. J. Heat Fluid Flow 67, 86–93 (2017)CrossRefGoogle Scholar
  5. 5.
    Gauding, M., Goebbert, J.H., Hasse, C., Peters, N.: Line segments in homogeneous scalar turbulence. Phys. Fluids 27(9), 095102 (2015)CrossRefGoogle Scholar
  6. 6.
    Gauding, M., Wick, A., Peters, N., Pitsch, H.: Generalized scale-by-scale energy budget equations for large-eddy simulations of scalar turbulence at various Schmidt numbers. J. Turbul. 15, 857–882 (2013)CrossRefGoogle Scholar
  7. 7.
    Goodfellow, I.J., et al.: Generative adversarial networks. arXiv:1406.2661 (2014)
  8. 8.
    Jülich Supercomputing Centre: JUQUEEN: IBM Blue Gene/Q supercomputer system at the Jülich supercomputing centre. J. Large-Scale Res. Facil. 1 (2015)Google Scholar
  9. 9.
    von Karman, T., Howarth, L.: On the statistical theory of isotropic turbulence. Proc. R. Soc. Lond. A: Math. Phys. Eng. Sci. 164(917), 192–215 (1938)CrossRefGoogle Scholar
  10. 10.
    Kolmogorov, A.N.: Dissipation of energy in locally isotropic turbulence. Dokl. Akad. Nauk SSSR 32, 16–18 (1941)MathSciNetzbMATHGoogle Scholar
  11. 11.
    Kolmogorov, A.N.: The local structure of turbulence in incompressible viscous fluid for very large Reynolds numbers. Dokl. Akad. Nauk SSSR 30, 299–303 (1941)MathSciNetGoogle Scholar
  12. 12.
    Peters, N., Boschung, J., Gauding, M., Goebbert, J.H., Hill, R.J., Pitsch, H.: Higher-order dissipation in the theory of homogeneous isotropic turbulence. J. Fluid Mech. 803, 250–274 (2016)MathSciNetCrossRefGoogle Scholar
  13. 13.
    Pope, S.B.: Turbulent Flows. Cambridge University Press, Cambridge (2000)CrossRefGoogle Scholar
  14. 14.
    Ronneberger, O., Fischer, P., Brox, T.: U-Net: convolutional networks for biomedical image segmentation. arXiv:1505.04597 (2015)Google Scholar
  15. 15.
    Sreenivasan, K.R.: The passive scalar spectrum and the Obukhov-Corrsin constant. Phys. Fluids 8, 189 (1996)MathSciNetCrossRefGoogle Scholar

Copyright information

© Springer Nature Switzerland AG 2018

Authors and Affiliations

  1. 1.Institute for Combustion TechnologyRWTH Aachen UniversityAachenGermany
  2. 2.CORIA – CNRS UMR 6614Saint Etienne du RouvrayFrance
  3. 3.Jülich Supercomputing Centre, FZ JülichJülichGermany

Personalised recommendations