Skip to main content
SpringerLink
Log in

Navier–stokes Generative Adversarial Network: a physics-informed deep learning model for fluid flow generation

  • Original Article
  • Published:
Neural Computing and Applications Aims and scope Submit manuscript

Abstract

Numerical simulation in Computational Fluid Dynamics mainly relies on discretizing the governing equations in time or space to obtain numerical solutions, which is expensive and time-consuming. Deep learning significantly reduces the computational cost due to its great nonlinear curve fitting capability, however, the data-driven models is agnostic to latent relationships between input and output. In this paper, a novel deep learning named Navier–Stokes Generative Adversarial Network integrated with physical information is proposed. The Navier–Stokes Equation is added to the loss function of the generator in the form of residuals, which means physics loss in this paper. Then, the proposed model is trained in the framework of generative adversarial network. Experimental results show that proposed model significantly outperform similar models, mean absolute error are decreased by 62.29, 78.42 and 78.61% on pressure and velocity components. What’s more, effectiveness of introducing physics loss is also verified.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10
Fig. 11
Fig. 12

Similar content being viewed by others

References

  1. Bai Xd, Wang Y, Zhang W (2020) Applying physics informed neural network for flow data assimilation. J Hydrodyn. pp. 1–9

  2. Baydin AG, Pearlmutter BA, Radul AA, Siskind JM (2018) Automatic differentiation in machine learning: a survey. J Mach Learn Res. Vol. 18

  3. Booz J, Yu W, Xu G, Griffith D, Golmie N (2019) A deep learning-based weather forecast system for data volume and recency analysis. In: 2019 International Conference on Computing, Networking and Communications (ICNC), pp. 697–701. IEEE

  4. Erichson NB, Muehlebach M, Mahoney MW (2019) Physics-informed autoencoders for lyapunov-stable fluid flow prediction. arXiv preprint arXiv:1905.10866

  5. Han Y, Chun Q, Jin H (2021) Wind-induced vibration performance of early chinese hall-style timber buildings. J Wood Sci 67(1):1–18

    Article  Google Scholar 

  6. Hornik K, Stinchcombe M, White H (1989) Multilayer feedforward networks are universal approximators. Neural Netw 2(5):359–366

    Article  Google Scholar 

  7. Huang G, Jafari AH (2021) Enhanced balancing gan: minority-class image generation. Neural Comput Appl. pp. 1–10

  8. Islama M, Nasrinb S (2020) Dusty fluid flow past between two parallel riga plates embedded in a porous medium

  9. Isola P, Zhu JY, Zhou T, Efros AA (2016) Image-to-image translation with conditional adversarial networks

  10. Kim B, Azevedo VC, Thuerey N, Kim T, Gross M, Solenthaler B (2019) Deep fluids: a generative network for parameterized fluid simulations. In: Computer Graphics Forum, vol. 38, pp. 59–70. Wiley Online Library

  11. Lye KO, Mishra S, Ray D (2020) Deep learning observables in computational fluid dynamics. J Comput Phys 410:109339

    Article  MathSciNet  Google Scholar 

  12. Major R, Kopernik M, Kuźmińska A, Imbir G, Plutecka H, Pomorska M, Ciach T, Lackner JM (2021) In vitro haemocompatibility assessment of acrylic acid deposited on solid, polyurethane substrate. Colloids Surf, B 199:111562

    Article  Google Scholar 

  13. Mao S, Rajan D, Chia LT (2021) Deep residual pooling network for texture recognition. Pattern Recogn 112:107817

    Article  Google Scholar 

  14. Noé F, Olsson S, Köhler J, Wu H (2019) Boltzmann generators: sampling equilibrium states of many-body systems with deep learning. Science 365:6457

    Article  Google Scholar 

  15. Obukhov AD, Krasnyanskiy MN (2021) Neural network method for automatic data generation in adaptive information systems. Neural Comput Appl pp. 1–23

  16. Raissi M, Perdikaris P, Karniadakis G (2017) Physics informed deep learning (part ii): Data-driven, discovery of nonlinear partial differential equations,”, arxiv e-prints, p. arXiv preprint arXiv:1711.10566

  17. Raissi M, Perdikaris P, Karniadakis GE (2017) Physics informed deep learning (part i): Data-driven solutions of nonlinear partial differential equations. arXiv preprint arXiv:1711.10561

  18. Raissi M, Perdikaris P, Karniadakis GE (2019) Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. J Comput Phys 378:686–707

    Article  MathSciNet  Google Scholar 

  19. Raissi M, Yazdani A, Karniadakis GE (2018) Hidden fluid mechanics: A navier–stokes informed deep learning framework for assimilating flow visualization data. arXiv preprint arXiv:1808.04327

  20. Raissi M, Yazdani A, Karniadakis GE (2020) Hidden fluid mechanics: learning velocity and pressure fields from flow visualizations. Science 367(6481):1026–1030

    Article  MathSciNet  Google Scholar 

  21. Ramachandran P, Zoph B, Le QV (2017) Searching for activation functions. arXiv preprint arXiv:1710.05941

  22. Rao C, Sun H, Liu Y (2020) Physics-informed deep learning for incompressible laminar flows. Theor Appl Mech Lett 10(3):207–212

    Article  Google Scholar 

  23. Salimans T, Kingma DP (2016) Weight normalization: a simple reparameterization to accelerate training of deep neural networks. arXiv preprint arXiv:1602.07868

  24. Saurav S, Saini R, Singh S (2021) Emnet: a deep integrated convolutional neural network for facial emotion recognition in the wild. Appl Intell. pp. 1–28

  25. Shams EA, Rizaner A, Ulusoy AH (2021) A novel context-aware feature extraction method for convolutional neural network-based intrusion detection systems. Neural Comput Appl. pp. 1–19

  26. Shanmuganathan V, Yesudhas HR, Khan MS, Khari M, Gandomi AH (2020) R-cnn and wavelet feature extraction for hand gesture recognition with emg signals. Neural Comput Appl 32(21):16723–16736

    Article  Google Scholar 

  27. Shorten C, Khoshgoftaar TM (2019) A survey on image data augmentation for deep learning. J Big Data 6(1):1–48

    Article  Google Scholar 

  28. Subramaniam A, Wong ML, Borker RD, Nimmagadda S, Lele SK (2020) Turbulence enrichment using physics-informed generative adversarial networks. arXiv e-prints pp. arXiv–2003

  29. Wandel N, Weinmann M, Klein R (2020) Fast fluid simulations in 3d with physics-informed deep learning. arXiv preprint arXiv:2012.11893

  30. Xu C, Wang H, Wu S, Lin Z (2021) Treelstm with tag-aware hypernetwork for sentence representation. Neurocomputing 434:11–20

    Article  Google Scholar 

  31. Xu H, Zhang W, Wang Y (2020) Explore missing flow dynamics by physics-informed deep learning: the parameterised governing systems. arXiv preprint arXiv:2008.12266

  32. Yang Z, Wu JL, Xiao H (2019) Enforcing deterministic constraints on generative adversarial networks for emulating physical systems. arXiv preprint arXiv:1911.06671

  33. Yi X, Duan Z, Li R, Zhang J, Li T, Zheng Y (2020) Predicting fine-grained air quality based on deep neural networks. IEEE Transactions on Big Data

Download references

Acknowledgements

This work is supported by Natural Science Foundation of Shanghai under Grant 19ZR1417700, Transforming Systems through Partnership, Newton Fund under Grant TSPC1086, High Performance Computing Center of Shanghai University.

Author information

Authors and Affiliations

  1. School of Computer Engineering and Science, Shanghai University, Shanghai, 200444, China

    Pin Wu, Kaikai Pan, Lulu Ji, Siquan Gong, Weibing Feng & Wenyan Yuan

  2. Data Science Institute, Department of Computing, Imperial College London, London, UK

    Christopher Pain

  3. Department of Earth Science and Engineering, Imperial College London, London, UK

    Christopher Pain

Authors
  1. Pin Wu
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Kaikai Pan
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Lulu Ji
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Siquan Gong
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Weibing Feng
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Wenyan Yuan
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Christopher Pain
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Pin Wu.

Ethics declarations

Conflict of interest

The authors declare that they have no conflict of interest.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Wu, P., Pan, K., Ji, L. et al. Navier–stokes Generative Adversarial Network: a physics-informed deep learning model for fluid flow generation. Neural Comput & Applic 34, 11539–11552 (2022). https://doi.org/10.1007/s00521-022-07042-6

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00521-022-07042-6

Keywords

Search

Navigation