Skip to main content

Advertisement

Springer Nature Link
Log in

Unsupervised learning with physics informed graph networks for partial differential equations

  • Published:
Applied Intelligence Aims and scope Submit manuscript

Abstract

Natural physical phenomena are commonly expressed using partial differential equations (PDEs), in domains such as fluid dynamics, electromagnetism, and atmospheric science. These equations typically require numerical solutions under given boundary conditions. There is a burgeoning interest in the exploration of neural network methodologies for solving PDEs, mainly based on automatic differentiation methods to learn the PDE-solving process, which means that the model needs to be retrained when the boundary conditions of PDE are changed. However, automatic differentiation requires substantial memory resources to facilitate the training regimen. Moreover, a learning objective that is tailored to the solution process often lacks the flexibility to extend to boundary conditions; thereby limiting the solution’s overall precision. The method proposed in this paper introduces a graph neural network approach, embedded with physical information, mainly for solving Poisson’s equation. An approach is introduced that reduces memory usage and enhances training efficiency through an unsupervised learning methodology based on numerical differentiation. Concurrently, by integrating boundary conditions directly into the neural network as supplementary physical information, this approach ensures that a singular model is capable of solving PDEs across a variety of boundary conditions. To address the challenges posed by more complex network inputs, the introduction of graph residual connections serves as a strategic measure to prevent network overfitting and to elevate the accuracy of the solutions provided. Experimental findings reveal that, despite having 30 times more training parameters than the Physics-Informed Neural Networks (PINN) model, the proposed model consumes 2.2% less memory than PINN. Additionally, generalization in boundary conditions has been achieved to a certain extent. This enables the model to solve partial differential equations with different boundary conditions, a capability that PINN currently lacks. To validate the solving capability of the proposed method, it has been applied to the model equation, the Sod shock tube problem, and the two-dimensional inviscid airfoil problem. In terms of the solution accuracy of the model equations, the proposed method outperforms PINN by 30% to four orders of magnitude. Compared to the traditional numerical method, the Finite Element Method (FEM), the proposed method also shows an order of magnitude improvement. Additionally, when compared to the improved version of PINN, TSONN, our method demonstrates certain advantages. The forward problem of the Sod shock tube, which PINN is currently unable to solve, is successfully handled by the proposed method. For the airfoil problem, the results are comparable to those of PINN.

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

Access this article

Log in via an institution

Subscribe and save

Springer+ Basic
$34.99 /Month
  • Get 10 units per month
  • Download Article/Chapter or eBook
  • 1 Unit = 1 Article or 1 Chapter
  • Cancel anytime
Subscribe now

Buy Now

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

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Algorithm 1
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

Explore related subjects

Discover the latest articles and news from researchers in related subjects, suggested using machine learning.

Data availability statement

The data that support the findings of this study are available from the corresponding author, [Jingyu Wang], upon reasonable request.

References

  1. Farlow SJ (2006) An introduction to differential equations and their applications, Courier Corporation

  2. Abbasi SF, Bilal M, Mukherjee T, Churm J, Pournik O, Epiphaniou G, Arvanitis TN (2024) Deep learning-based synthetic skin lesion image classification. In: Digital health and informatics innovations for sustainable health care systems, IOS Press, pp 1145–1150

  3. Temam R (2001) Navier-Stokes equations: theory and numerical analysis, Vol. 343, American Mathematical Soc

  4. Majda AJ, Bertozzi AL, Ogawa A (2002) Vorticity and incompressible flow. cambridge texts in applied mathematics. Appl Mech Rev 55(4) B77–B78. https://doi.org/10.1115/1.1483363

  5. Date AW (2005) Introduction to computational fluid dynamics. Cambridge University Press

    Book  MATH  Google Scholar 

  6. Dahlquist G, Björck Å (2008) Numerical methods in scientific computing, volume I, SIAM

  7. Yadav N, Yadav A, Kumar M, et al (2015) An introduction to neural network methods for differential equations, Vol. 1, Springer

  8. Meade AJ Jr, Fernandez AA (1994) Solution of nonlinear ordinary differential equations by feedforward neural networks. Math Comput Modell 20(9):19–44. https://doi.org/10.1016/0895-7177(94)00160-x

    Article  MathSciNet  MATH  Google Scholar 

  9. Meade AJ Jr, Fernandez AA (1994) The numerical solution of linear ordinary differential equations by feedforward neural networks. Math Comput Modell 19(12):1–25. https://doi.org/10.1016/0895-7177(94)90095-7

    Article  MathSciNet  MATH  Google Scholar 

  10. Lagaris IE, Likas A, Fotiadis DI (1998) Artificial neural networks for solving ordinary and partial differential equations. IEEE Trans Neural Netw 9(5):987–1000. https://doi.org/10.1109/72.718734

    Article  MATH  Google Scholar 

  11. LeCun Y, Bengio Y, Hinton G (2015) Deep learning. Nature 521(7553):436–444. https://doi.org/10.1038/nature14539

    Article  MATH  Google Scholar 

  12. Krizhevsky A, Sutskever I, Hinton G (2012) Imagenet classification with deep convolutional neural networks. Adv Neural Inf Process Syst 25(2). https://doi.org/10.1145/3065386

  13. Hinton G, Deng L, Yu D, Dahl GE, Mohamed A-r, Jaitly N, Senior A, Vanhoucke V, Nguyen P, Sainath TN et al (2012) Deep neural networks for acoustic modeling in speech recognition: The shared views of four research groups. IEEE Signal Process Mag 29(6):82–97. https://doi.org/10.1109/MSP.2012.2205597

    Article  Google Scholar 

  14. Wu Y, Schuster M, Chen Z, Le QV, Norouzi M, Macherey W, Krikun M, Cao Y, Gao Q, Macherey K, et al (2016) Google’s neural machine translation system: Bridging the gap between human and machine translation, arXiv:1609.08144, https://doi.org/10.48550/arXiv.1609.08144

  15. Stevens R, Taylor V, Nichols J, Maccabe AB, Yelick K, Brown D (2020) Ai for science: Report on the department of energy (doe) town halls on artificial intelligence (ai) for science, Tech. rep., Argonne National Lab.(ANL), Argonne, IL (United States)

  16. Weinan E (2021) The dawning of a new era in applied mathematics. Notices Am Math Soc 68(4):565–571. https://doi.org/10.1090/noti2411

    Article  MathSciNet  MATH  Google Scholar 

  17. Liu Z, Chen Y, Du Y, Tegmark M (2021) Physics-augmented learning: A new paradigm beyond physics-informed learning, arXiv:2109.13901, https://doi.org/10.48550/arXiv.2109.13901

  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. https://doi.org/10.1016/j.jcp.2018.10.045

    Article  MathSciNet  MATH  Google Scholar 

  19. Cai S, Mao Z, Wang Z, Yin M, Karniadakis GE (2021) Karniadakis, Physics-informed neural networks (pinns) for fluid mechanics: A review. Acta Mech Sinica 37(12):1727–1738. https://doi.org/10.1007/s10409-021-01148-1

    Article  MathSciNet  MATH  Google Scholar 

  20. Jagtap AD, Kharazmi E, Karniadakis GE (2020) Conservative physics-informed neural networks on discrete domains for conservation laws: Applications to forward and inverse problems. Comput Methods Appl Mech Eng 365:113028. https://doi.org/10.1016/j.cma.2020.113028

    Article  MathSciNet  MATH  Google Scholar 

  21. McGreivy N, Hakim A (2024) Weak baselines and reporting biases lead to overoptimism in machine learning for fluid-related partial differential equations. Nature Mach Intell 6(10):1256–1269. https://doi.org/10.1038/s42256-024-00897-5

    Article  MATH  Google Scholar 

  22. Grossmann TG, Komorowska UJ, Latz J, Schönlieb C-B (2024) Can physics-informed neural networks beat the finite element method? IMA J Appl Math. https://doi.org/10.1093/imamat/hxae011

    Article  MathSciNet  MATH  Google Scholar 

  23. Dong S, Li Z (2021) Local extreme learning machines and domain decomposition for solving linear and nonlinear partial differential equations. Comput Methods Appl Mech Eng 387:114129. https://doi.org/10.1016/j.cma.2021.114129

    Article  MathSciNet  MATH  Google Scholar 

  24. Song J, Cao W, Liao F, Zhang W (2025) Vw-pinns: A volume weighting method for pde residuals in physics-informed neural networks. Acta Mechanica Sinica 41(3):324140. https://doi.org/10.1007/s10409-024-24140-x

  25. Cao W, Zhang W (2023) Tsonn: Time-stepping-oriented neural network for solving partial differential equations, arXiv:2310.16491. https://doi.org/10.48550/arXiv.2310.16491

  26. Pioch F, Harmening JH, Müller AM, Peitzmann F-J, Schramm D, Oe Moctar (2023) Turbulence modeling for physics-informed neural networks: Comparison of different rans models for the backward-facing step flow. Fluids 8(2):43. https://doi.org/10.3390/fluids8020043

    Article  Google Scholar 

  27. Özbay AG, Hamzehloo A, Laizet S, Tzirakis P, Rizos G, Schuller B (2021) Poisson cnn: Convolutional neural networks for the solution of the poisson equation on a cartesian mesh. Data-Centric Eng 2:e6. https://doi.org/10.1017/dce.2021.7

    Article  Google Scholar 

  28. Wang S, Teng Y, Perdikaris P (2020) Understanding and mitigating gradient pathologies in physics-informed neural networks, arXiv:2001.04536. https://doi.org/10.1137/20m1318043

  29. Wang Y, Yao Y, Guo J, Gao Z (2024) A practical pinn framework for multi-scale problems with multi-magnitude loss terms. J Comput Phys 510:113112. https://doi.org/10.1016/j.jcp.2024.113112

    Article  MathSciNet  MATH  Google Scholar 

  30. Seo S, Meng C, Rambhatla S, Liu Y (2021) Physics-aware spatiotemporal modules with auxiliary tasks for meta-learning. In: Proceedings of the thirtieth international joint conference on artificial intelligence, IJCAI-2021, International Joint Conferences on Artificial Intelligence Organization, pp 2943–2949. https://doi.org/10.24963/ijcai.2021/405

  31. Seo S, Meng C, Rambhatla S, Liu Y (2020) Physics-aware spatiotemporal modules with auxiliary tasks for meta-learning, arXiv:2006.08831, https://doi.org/10.48550/arXiv.2006.08831

  32. Gao H, Zahr MJ, Wang J-X (2022) Physics-informed graph neural galerkin networks: A unified framework for solving pde-governed forward and inverse problems. Comput Methods Appl Mech Eng 390:114502. https://doi.org/10.1115/1.0004979v

    Article  MathSciNet  MATH  Google Scholar 

  33. He K, Zhang X, Ren S, Sun J (2016) Deep residual learning for image recognition. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp 770–778. https://doi.org/10.1109/cvpr.2016.90

  34. Lim L-H (2020) Hodge laplacians on graphs. Siam Rev 62(3):685–715. https://doi.org/10.1137/18m1223101

    Article  Google Scholar 

  35. Battaglia PW, Hamrick JB, Bapst V, Sanchez-Gonzalez A, Zambaldi V, Malinowski M, Tacchetti A, Raposo D, Santoro A, Faulkner R, et al (2018) Relational inductive biases, deep learning, and graph networks, arXiv:1806.01261. https://doi.org/10.48550/arXiv.1806.01261

  36. Hou TY, LeFloch PG (1994) Why nonconservative schemes converge to wrong solutions: error analysis. Math Comput 62(206):497–530. https://doi.org/10.2307/2153520

    Article  MathSciNet  MATH  Google Scholar 

  37. Cao W, Song J, Zhang W (2024) A solver for subsonic flow around airfoils based on physics-informed neural networks and mesh transformation. Phys Fluids 36(2). https://doi.org/10.1063/5.0188665

Download references

Acknowledgements

This work was supported by the National Key Project of China (Grant No. GJXM92579), the Sichuan Science and Technology Program (Project No. 2023YFG0329), and the Open Funding of the National Key Laboratory of Digital and Agile Aircraft Design.

Author information

Authors and Affiliations

  1. College of Computer Science, Sichuan University, Chengdu, 610065, China

    Lin Lu & Yiye Zou

  2. School of Aeronautics and Astronautics, Sichuan University, Chengdu, 610200, China

    Jingyu Wang

  3. College of Aerospace Science and Engineering, National University of Defense Technology, Changsha, 410000, China

    Shufan Zou

  4. National Key Laboratory of Fundamental Algorithms and Models for Engineering Simulation, Sichuan University, Chengdu, 610207, China

    Lin Lu, Yiye Zou, Jingyu Wang, Laiping Zhang & Xiaogang Deng

  5. National Key Laboratory of Digital and Agile Aircraft Design, Chengdu, China

    Jingyu Wang

Authors
  1. Lin Lu
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Yiye Zou
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Jingyu Wang
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Shufan Zou
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Laiping Zhang
    View author publications

    You can also search for this author inPubMed Google Scholar

  6. Xiaogang Deng
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

Lin Lu: Writing – original draft, Software, Methodology, Conceptualization. Jingyu Wang: Writing – review & editing, Methodology, Investigation, Formal analysis, Data curation. Shufan Zou: Writing – review & editing, Methodology, Formal analysis, Conceptualization. Laiping Zhang: Writing – review & editing, supervision, Project administration, Data curation. Xiaogang Deng: supervision, Resources, Validation, Formal analysis, Project administration.

Corresponding author

Correspondence to Jingyu Wang.

Ethics declarations

Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Additional information

Publisher's Note

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

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Lu, L., Zou, Y., Wang, J. et al. Unsupervised learning with physics informed graph networks for partial differential equations. Appl Intell 55, 617 (2025). https://doi.org/10.1007/s10489-025-06479-1

Download citation

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1007/s10489-025-06479-1

Keywords