Hostname: page-component-857557d7f7-nbs69 Total loading time: 0 Render date: 2025-12-08T21:38:00.641Z Has data issue: false hasContentIssue false
  • English
  • Français

Symbolic identification of tensor equations in multidimensional physical fields

Published online by Cambridge University Press:  02 December 2025

Tianyi Chen Open the ORCID record for Tianyi Chen [Opens in a new window] ,
Hao Yang ,
Wenjun Ma Open the ORCID record for Wenjun Ma [Opens in a new window]  and
Jun Zhang Open the ORCID record for Jun Zhang [Opens in a new window]
Tianyi Chen
Affiliation:
School of Aeronautic Science and Engineering, Beihang University, Beijing 100191, PR China
Hao Yang
Affiliation:
School of Aeronautic Science and Engineering, Beihang University, Beijing 100191, PR China
Wenjun Ma
Affiliation:
School of Aeronautic Science and Engineering, Beihang University, Beijing 100191, PR China
Jun Zhang*
Affiliation:
School of Aeronautic Science and Engineering, Beihang University, Beijing 100191, PR China
*
Corresponding author: Jun Zhang, jun.zhang@buaa.edu.cn
Get access Rights & Permissions [Opens in a new window]

Abstract

Recently, data-driven methods have shown great promise for discovering governing equations from simulation or experimental data. However, most existing approaches are limited to scalar equations, with few capable of identifying tensor relationships. In this work, we propose a general data-driven framework for identifying tensor equations, referred to as symbolic identification of tensor equations (SITE). The core idea of SITE – representing tensor equations using a host–plasmid structure – is inspired by the multidimensional gene expression programming approach. To improve the robustness of the evolutionary process, SITE adopts a genetic information retention strategy. Moreover, SITE introduces two key innovations beyond conventional evolutionary algorithms. First, it incorporates a dimensional homogeneity check to restrict the search space and eliminate physically invalid expressions. Second, it replaces traditional linear scaling with a tensor linear regression technique, greatly enhancing the efficiency of numerical coefficient optimization. We validate SITE using two benchmark scenarios, where it accurately recovers target equations from synthetic data, showing robustness to noise and flexible expressive capability. Furthermore, SITE is applied to identify constitutive relations directly from molecular simulation data, which are generated without reliance on macroscopic constitutive models. It adapts to both compressible and incompressible flow conditions and successfully identifies the corresponding macroscopic forms, highlighting its potential for data-driven discovery of tensor equation.

JFM classification

Mathematical Foundations: Machine learning

Information

Type
JFM Papers
Information
Journal of Fluid Mechanics , Volume 1024 , 10 December 2025 , A34
Check for updates
Copyright
© The Author(s), 2025. Published by Cambridge University Press

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

Article purchase

Temporarily unavailable

References

Bakarji, J., Callaham, J., Brunton, S.L. & Kutz, J.N. 2022 Dimensionally consistent learning with Buckingham Pi. Nat. Comput. Sci. 2 (12), 834–844.10.1038/s43588-022-00355-5CrossRefGoogle ScholarPubMed
Bird, G.A. 1994 Molecular Gas Dynamics and the Direct Simulation of Gas Flows. Oxford University Press.10.1093/oso/9780198561958.001.0001CrossRefGoogle Scholar
Bongard, J. & Lipson, H. 2007 Automated reverse engineering of nonlinear dynamical systems. Proc. Natl Acad. Sci. USA 104 (24), 9943–9948.10.1073/pnas.0609476104CrossRefGoogle ScholarPubMed
Brunton, S.L., Noack, B.R. & Koumoutsakos, P. 2020 Machine learning for fluid mechanics. Annu. Rev. Fluid Mech. 52 (2020), 477–508.10.1146/annurev-fluid-010719-060214CrossRefGoogle Scholar
Brunton, S.L., Proctor, J.L. & Kutz, J.N. 2016 Discovering governing equations from data by sparse identification of nonlinear dynamical systems. Proc. Natl Acad. Sci. USA 113 (15), 3932–3937.10.1073/pnas.1517384113CrossRefGoogle ScholarPubMed
Byrd, R.H., Lu, P., Nocedal, J. & Zhu, C. 1995 A limited memory algorithm for bound constrained optimization. Siam J. Sci. Comput. 16 (5), 1190–1208.10.1137/0916069CrossRefGoogle Scholar
Chen, Y., Luo, Y., Liu, Q., Xu, H. & Zhang, D. 2022 Symbolic genetic algorithm for discovering open-form partial differential equations (SGA-PDE). Phys. Rev. Res. 4 (2), 023174.10.1103/PhysRevResearch.4.023174CrossRefGoogle Scholar
Chen, Z., Liu, Y. & Sun, H. 2021 Physics-informed learning of governing equations from scarce data. Nat. Commun. 12 (1), 6136.10.1038/s41467-021-26434-1CrossRefGoogle ScholarPubMed
Cohen, B.G., Beykal, B. & Bollas, G.M. 2024 Physics-informed genetic programming for discovery of partial differential equations from scarce and noisy data. J. Comput. Phys. 514, 113261.10.1016/j.jcp.2024.113261CrossRefGoogle ScholarPubMed
Cranmer, M. 2023 Interpretable Machine Learning for Science with PySR and SymbolicRegression.jl. arXiv:2305.01582.Google Scholar
Du, M., Chen, Y., Nie, L., Lou, S. & Zhang, D. 2024 Physics-constrained robust learning of open-form partial differential equations from limited and noisy data. Phys. Fluids 36 (5), 057123.10.1063/5.0204187CrossRefGoogle Scholar
Ferreira, C. 2001 Gene expression programming: a new adaptive algorithm for solving problems. Complex Syst. 13 (2), 87–129.Google Scholar
Fukami, K., Goto, S. & Taira, K. 2024 Data-driven nonlinear turbulent flow scaling with Buckingham Pi variables. J. Fluid Mech. 984, R4.10.1017/jfm.2024.211CrossRefGoogle Scholar
Goldberg, D.E. & Richardson, J. 1987 Genetic algorithms with sharing for multimodal function optimization. In Genetic algorithms and their applications: Proceedings of the Second International Conference on Genetic Algorithms, vol. 4149, pp. 414–425. Lawrence Erlbaum. Lawrence Erlbaum.Google Scholar
Golden, M., Grigoriev, R.O., Nambisan, J. & Fernandez-Nieves, A. 2023 Physically informed data-driven modeling of active nematics. Sci. Adv. 9 (27), eabq6120.10.1126/sciadv.abq6120CrossRefGoogle ScholarPubMed
Gurevich, D.R., Golden, M.R., Reinbold, P.A.K. & Grigoriev, R.O. 2024 Learning fluid physics from highly turbulent data using sparse physics-informed discovery of empirical relations (SPIDER). J. Fluid Mech. 996, A25.10.1017/jfm.2024.813CrossRefGoogle Scholar
Hadjiconstantinou, N.G., Garcia, A.L., Bazant, M.Z. & He, G. 2003 Statistical error in particle simulations of hydrodynamic phenomena. J. Comput. Phys. 187 (1), 274–297.10.1016/S0021-9991(03)00099-8CrossRefGoogle Scholar
Jasak, H., Jemcov, A. & Tukovic, Z. 2007 Openfoam: A C++ library for complex physics simulations. In International workshop on coupled methods in numerical dynamics, vol. 1000, pp. 1–20, Dubrovnik, Croatia.Google Scholar
Keijzer, M. 2003 Improving symbolic regression with interval arithmetic and linear scaling. In European Conference on Genetic Programming, pp. 70–82. Springer.10.1007/3-540-36599-0_7CrossRefGoogle Scholar
Kingma, D.P. & Ba, J. 2017, Adam: A Method for Stochastic Optimization. arXiv:1412.6980.Google Scholar
Koza, J.R. 1994 Genetic programming as a means for programming computers by natural selection. Stat. Comput. 4, 87–112.10.1007/BF00175355CrossRefGoogle Scholar
Launder, B.E., Reece, G. Jr & Rodi, W. 1975 Progress in the development of a Reynolds-stress turbulence closure. J. Fluid Mech. 68 (3), 537–566.10.1017/S0022112075001814CrossRefGoogle Scholar
Lennon, K.R., McKinley, G.H. & Swan, J.W. 2023 Scientific machine learning for modeling and simulating complex fluids. Proc. Natl Acad. Sci. USA 120 (27), e2304669120.10.1073/pnas.2304669120CrossRefGoogle ScholarPubMed
Li, H., Zhao, Y., Wang, J. & Sandberg, R.D. 2021 Data-driven model development for large-eddy simulation of turbulence using gene-expression programing. Phys. Fluids 33 (12), 125127.10.1063/5.0076693CrossRefGoogle Scholar
Loiseau, J.-C. & Brunton, S.L. 2018 Constrained sparse Galerkin regression. J. Fluid Mech. 838, 42–67.10.1017/jfm.2017.823CrossRefGoogle Scholar
Lu, L., Zeng, Z., Jiang, Y., Zhu, Y. & Hu, P. 2024 Weak collocation regression method: fast reveal hidden stochastic dynamics from high-dimensional aggregate data. J. Comput. Phys. 502, 112799.10.1016/j.jcp.2024.112799CrossRefGoogle Scholar
Ma, Q., Yang, C., Chen, S., Feng, K., Cui, Z. & Zhang, J. 2024 a Effect of thermal fluctuations on spectra and predictability in compressible decaying isotropic turbulence. J. Fluid Mech. 987, A29.10.1017/jfm.2024.342CrossRefGoogle Scholar
Ma, W., Zhang, J., Feng, K., Xing, H. & Wen, D. 2024 b Dimensional homogeneity constrained gene expression programming for discovering governing equations. J. Fluid Mech. 985, A12.10.1017/jfm.2024.272CrossRefGoogle Scholar
Maslyaev, M., Hvatov, A. & Kalyuzhnaya, A. 2019 Data-driven partial derivative equations discovery with evolutionary approach, proceedings. In Computational Science–ICCS 2019: 19th International Conference, Faro, Portugal, June 12–14, 2019, Proceedings, Part V 19, pp. 635–641. Springer.10.1007/978-3-030-22750-0_61CrossRefGoogle Scholar
Reese, J.M., Gallis, M.A. & Lockerby, D.A. 2003 New directions in fluid dynamics: non-equilibrium aerodynamic and microsystem flows. Phil. Trans. R. Soc. Lond. A 361 (1813), 2967–2988.10.1098/rsta.2003.1281CrossRefGoogle ScholarPubMed
Reinbold, P.A.K., Kageorge, L.M., Schatz, M.F. & Grigoriev, R.O. 2021 Robust learning from noisy, incomplete, high-dimensional experimental data via physically constrained symbolic regression. Nat. Commun. 12 (1), 3219.10.1038/s41467-021-23479-0CrossRefGoogle ScholarPubMed
Reissmann, M., Fang, Y., Ooi, A.S.H. & Sandberg, R.D. 2025 Constraining genetic symbolic regression via semantic backpropagation. Genet. Program. Evol. Mach. 26 (1), 12.10.1007/s10710-025-09510-zCrossRefGoogle Scholar
Schmidt, M. & Lipson, H. 2009 Distilling free-form natural laws from experimental data. Science 324 (5923), 81–85.10.1126/science.1165893CrossRefGoogle ScholarPubMed
Udrescu, S.-M. & Tegmark, M. 2020 Ai Feynman: a physics-inspired method for symbolic regression. Sci. Adv. 6 (16), eaay2631.10.1126/sciadv.aay2631CrossRefGoogle ScholarPubMed
Wang, H. et al. 2023 Scientific discovery in the age of artificial intelligence. Nature 620 (7972), 47–60.10.1038/s41586-023-06221-2CrossRefGoogle ScholarPubMed
Wang, L., Ji, T., Huang, Y., Zhou, H. & Xie, F. 2025 Physics-based discovery of governing equations from scarce and noisy data. Phys. Rev. E 111 (6), 065314.10.1103/d4tm-92vbCrossRefGoogle ScholarPubMed
Weatheritt, J. & Sandberg, R. 2016 A novel evolutionary algorithm applied to algebraic modifications of the RANS stress–strain relationship. J. Comput. Phys. 325, 22–37.10.1016/j.jcp.2016.08.015CrossRefGoogle Scholar
Zhang, B., Lei, J., Sun, G., Ding, S. & Guo, J. 2025 Cartesian tensor-based sparse regression for data-driven discovery of high-dimensional invariant governing equations. Phys. Fluids 37 (7), 077191.10.1063/5.0275717CrossRefGoogle Scholar
Zhang, J. & Ma, W. 2020 Data-driven discovery of governing equations for fluid dynamics based on molecular simulation. J. Fluid Mech. 892, A5.10.1017/jfm.2020.184CrossRefGoogle Scholar
Zhao, Y., Akolekar, H.D., Weatheritt, J., Michelassi, V. & Sandberg, R.D. 2020 RANS turbulence model development using CFD-driven machine learning. J. Comput. Phys. 411, 109413.10.1016/j.jcp.2020.109413CrossRefGoogle Scholar