Physics-informed neural networks for solving forward and inverse flow problems via the Boltzmann-BGK formulation. (English) Zbl 07516434
Summary: The Boltzmann equation with the Bhatnagar-Gross-Krook collision model (Boltzmann-BGK equation) has been widely employed to describe multiscale flows, i.e., from the hydrodynamic limit to free molecular flow. In this study, we employ physics-informed neural networks (PINNs) to solve forward and inverse problems via the Boltzmann-BGK formulation (PINN-BGK), enabling PINNs to model flows in both the continuum and rarefied regimes. In particular, the PINN-BGK is composed of three sub-networks, i.e., the first for approximating the equilibrium distribution function, the second for approximating the non-equilibrium distribution function, and the third one for encoding the Boltzmann-BGK equation as well as the corresponding boundary/initial conditions. By minimizing the residuals of the governing equations and the mismatch between the predicted and provided boundary/initial conditions, we can approximate the Boltzmann-BGK equation for both continuous and rarefied flows. For forward problems, the PINN-BGK is utilized to solve various benchmark flows given boundary/initial conditions, e.g., Kovasznay flow, Taylor-Green flow, cavity flow, and micro Couette flow for Knudsen number up to 5. For inverse problems, we focus on rarefied flows in which accurate boundary conditions are difficult to obtain. We employ the PINN-BGK to infer the flow field in the entire computational domain given a limited number of interior scattered measurements on the velocity without using the (unknown) boundary conditions. Results for the two-dimensional micro Couette and micro cavity flows with Knudsen numbers ranging from 0.1 to 10 indicate that the PINN-BGK can infer the velocity field in the entire domain with good accuracy. Finally, we also present some results on using transfer learning to accelerate the training process. Specifically, we can obtain a three-fold speedup compared to the standard training process (e.g., Adam plus L-BFGS-B) for the two-dimensional flow problems considered in our work.
MSC:
| 76Mxx | Basic methods in fluid mechanics |
| 76Pxx | Rarefied gas flows, Boltzmann equation in fluid mechanics |
| 68Txx | Artificial intelligence |
Keywords:
Boltzmann-BGK equation; physics-informed neural networks; rarefied flows; velocity slip; inverse problems; transfer learningReferences:
| [1] | Akkutlu, I. Y.; Efendiev, Y.; Vasilyeva, M.; Wang, Y., Multiscale model reduction for shale gas transport in poroelastic fractured media, J. Comput. Phys., 353, 356-376 (2018) · Zbl 1380.76032 |
| [2] | Jin, Z.; Firoozabadi, A., Flow of methane in shale nanopores at low and high pressure by molecular dynamics simulations, J. Chem. Phys., 143, 10, Article 104315 pp. (2015) |
| [3] | Greene, J., Transitioning from the art to the science of thin films: 1964 to 2003, J. Vac. Sci. Technol. A, 21, 5, S71-S73 (2003) |
| [4] | Redman, A. L.; Bailleres, H.; Perré, P.; Carr, E.; Turner, I., A relevant and robust vacuum-drying model applied to hardwoods, Wood Sci. Technol., 51, 4, 701-719 (2017) |
| [5] | Karniadakis, G. E.; Beskok, A.; Aluru, N., Microflows and Nanoflows: Fundamentals and Simulation (2005), Springer: Springer New York · Zbl 1115.76003 |
| [6] | Qiao, R.; He, P., Modulation of electroosmotic flow by neutral polymers, Langmuir, 23, 10, 5810-5816 (2007) |
| [7] | Succi, S.; Benzi, R.; Biferale, L.; Sbragaglia, M.; Toschi, F., Lattice kinetic theory as a form of supra-molecular dynamics for computational microfluidics, Bull. Pol. Acad. Sci.-Tech., 151-158 (2007) · Zbl 1203.76014 |
| [8] | Bhatnagar, P. L.; Gross, E. P.; Krook, M., A model for collision processes in gases. I. Small amplitude processes in charged and neutral one-component systems, Phys. Rev., 94, 3, 511 (1954) · Zbl 0055.23609 |
| [9] | Chen, H.; Orszag, S. A.; Staroselsky, I., Macroscopic description of arbitrary Knudsen number flow using Boltzmann-BGK kinetic theory, J. Fluid Mech., 574, 495 (2007) · Zbl 1133.76346 |
| [10] | Meng, J.; Zhang, Y., Accuracy analysis of high-order lattice Boltzmann models for rarefied gas flows, J. Comput. Phys., 230, 3, 835-849 (2011) · Zbl 1283.76059 |
| [11] | Luo, L., Some recent results on discrete velocity models and ramifications for lattice Boltzmann equation, Comput. Phys. Commun., 129, 63-74 (2000) · Zbl 0983.76072 |
| [12] | Lallemand, P.; Luo, L., Lattice Boltzmann equation with overset method for moving objects in two-dimensional flows, J. Comput. Phys., 407, Article 109223 pp. (2020) · Zbl 07504698 |
| [13] | Luo, L., Theory of the lattice Boltzmann method: lattice Boltzmann models for nonideal gases, Phys. Rev. E, 62, 4, 4982 (2000) |
| [14] | Guo, Z.; Shu, C., Lattice Boltzmann Method and Its Applications in Engineering (2013), World Scientific Publishing · Zbl 1278.76001 |
| [15] | Succi, S., The Lattice Boltzmann Equation: For Fluid Dynamics and Beyond (2001), Oxford University Press: Oxford University Press New York · Zbl 0990.76001 |
| [16] | Aidun, C. K.; Clausen, J. R., Lattice Boltzmann method for complex flows, Annu. Rev. Fluid Mech., 42, 439-472 (2010) · Zbl 1345.76087 |
| [17] | Geier, M.; Lenz, S.; Schönherr, M.; Krafczyk, M., Under-resolved and large eddy simulations of a decaying Taylor-Green vortex with the cumulant lattice Boltzmann method, Theor. Comput. Fluid Dyn., 35, 2, 169-208 (2021) |
| [18] | Geier, M.; Pasquali, A.; Schönherr, M., Parametrization of the cumulant lattice Boltzmann method for fourth order accurate diffusion part I: derivation and validation, J. Comput. Phys., 348, 862-888 (2017) · Zbl 1380.76119 |
| [19] | Geier, M.; Pasquali, A.; Schönherr, M., Parametrization of the cumulant lattice Boltzmann method for fourth order accurate diffusion part II: application to flow around a sphere at drag crisis, J. Comput. Phys., 348, 889-898 (2017) · Zbl 1380.76120 |
| [20] | Lallemand, P.; Luo, L.; Krafczyk, M.; Yong, W., The lattice Boltzmann method for nearly incompressible flows, J. Comput. Phys., 431, Article 109713 pp. (2021) · Zbl 07511443 |
| [21] | Mieussens, L., Discrete velocity model and implicit scheme for the BGK equation of rarefied gas dynamics, Math. Models Methods Appl. Sci., 10, 8, 1121-1149 (2000) · Zbl 1078.82526 |
| [22] | Xu, K.; Huang, J., A unified gas-kinetic scheme for continuum and rarefied flows, J. Comput. Phys., 229, 20, 7747-7764 (2010) · Zbl 1276.76057 |
| [23] | Xiao, T.; Liu, C.; Xu, K.; Cai, Q., A velocity-space adaptive unified gas kinetic scheme for continuum and rarefied flows, J. Comput. Phys., 415, Article 109535 pp. (2020) · Zbl 1440.76139 |
| [24] | Xu, K.; Huang, J., An improved unified gas-kinetic scheme and the study of shock structures, IMA J. Appl. Math., 76, 5, 698-711 (2011) · Zbl 1276.76063 |
| [25] | Titarev, V. A., Numerical method for computing two-dimensional unsteady rarefied gas flows in arbitrarily shaped domains, Comput. Math. Math. Phys., 49, 7, 1197-1211 (2009) · Zbl 1224.76119 |
| [26] | Guo, Z.; Xu, K.; Wang, R., Discrete unified gas kinetic scheme for all Knudsen number flows: low-speed isothermal case, Phys. Rev. E, 88, 3, Article 033305 pp. (2013) |
| [27] | Zhang, C.; Guo, Z., Discrete unified gas kinetic scheme for multiscale heat transfer with arbitrary temperature difference, Int. J. Heat Mass Transf., 134, 1127-1136 (2019) |
| [28] | Zhang, C.; Guo, Z.; Chen, S., An implicit kinetic scheme for multiscale heat transfer problem accounting for phonon dispersion and polarization, Int. J. Heat Mass Transf., 130, 1366-1376 (2019) |
| [29] | Meng, X.; Sun, H.; Guo, Z.; Yang, X., A multiscale study of density-driven flow with dissolution in porous media, Adv. Water Resour., 142, Article 103640 pp. (2020) |
| [30] | Dissanayake, M.; Phan-Thien, N., Neural-network-based approximations for solving partial differential equations, Commun. Numer. Methods Eng., 10, 3, 195-201 (1994) · Zbl 0802.65102 |
| [31] | Sirignano, J.; Spiliopoulos, K., DGM: a deep learning algorithm for solving partial differential equations, J. Comput. Phys., 375, 1339-1364 (2018) · Zbl 1416.65394 |
| [32] | Wu, K.; Xiu, D., Data-driven deep learning of partial differential equations in modal space, J. Comput. Phys., 408, Article 109307 pp. (2020) · Zbl 07505629 |
| [33] | Raissi, M.; Perdikaris, P.; Karniadakis, G. E., 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 (2019) · Zbl 1415.68175 |
| [34] | Baydin, A. G.; Pearlmutter, B. A.; Radul, A. A.; Siskind, J. M., Automatic differentiation in machine learning: a survey, J. Mach. Learn. Res., 18, 5595-5637 (2018) · Zbl 06982909 |
| [35] | Lu, L.; Meng, X.; Mao, Z.; Karniadakis, G. E., DeepXDE: a deep learning library for solving differential equations, SIAM Rev., 63, 1, 208-228 (2021) · Zbl 1459.65002 |
| [36] | Meng, X.; Li, Z.; Zhang, D.; Karniadakis, G. E., PPINN: parareal physics-informed neural network for time-dependent PDEs, Comput. Methods Appl. Mech., 370, Article 113250 pp. (2020) · Zbl 1506.65181 |
| [37] | Meng, X.; Karniadakis, G. E., A composite neural network that learns from multi-fidelity data: application to function approximation and inverse PDE problems, J. Comput. Phys., 401, Article 109020 pp. (2020) · Zbl 1454.76006 |
| [38] | Jin, X.; Cai, S.; Li, H.; Karniadakis, G. E., NSFnets (Navier-Stokes flow nets): physics-informed neural networks for the incompressible Navier-Stokes equations, J. Comput. Phys., 426, Article 109951 pp. (2021) · Zbl 07510065 |
| [39] | Mao, Z.; Jagtap, A. D.; Karniadakis, G. E., Physics-informed neural networks for high-speed flows, Comput. Methods Appl. Mech., 360, Article 112789 pp. (2020) · Zbl 1442.76092 |
| [40] | Raissi, M.; Yazdani, A.; Karniadakis, G. E., Hidden fluid mechanics: learning velocity and pressure fields from flow visualizations, Science, 367, 6481, 1026-1030 (2020) · Zbl 1478.76057 |
| [41] | Yu, H.; Chen, J.; Zhu, Y.; Wang, F.; Wu, H., Multiscale transport mechanism of shale gas in micro/nano-pores, Int. J. Heat Mass Transf., 111, 1172-1180 (2017) |
| [42] | Yang, X.; Zhou, W.; Liu, X.; Yan, Y., A multiscale approach for simulation of shale gas transport in organic nanopores, Energy, 210, Article 118547 pp. (2020) |
| [43] | Han, J.; Ma, C.; Ma, Z.; E, W., Uniformly accurate machine learning-based hydrodynamic models for kinetic equations, Proc. Natl. Acad. Sci. USA, 116, 44, 21983-21991 (2019) · Zbl 1431.76100 |
| [44] | Bandyopadhyay, P. R., Rough-wall turbulent boundary layers in the transition regime, J. Fluid Mech., 180, 231-266 (1987) |
| [45] | Gu, X.; Emerson, D. R., A high-order moment approach for capturing non-equilibrium phenomena in the transition regime, J. Fluid Mech., 636, 177 (2009) · Zbl 1183.76850 |
| [46] | Bhattacharya, D. K.; Lie, G. C., Nonequilibrium gas flow in the transition regime: a molecular-dynamics study, Phys. Rev. A, 43, 2, 761 (1991) |
| [47] | Barisik, M.; Beskok, A., Surface-gas interaction effects on nanoscale gas flows, Microfluid. Nanofluid., 13, 5, 789-798 (2012) |
| [48] | Celebi, A. T.; Beskok, A., Molecular and continuum transport perspectives on electroosmotic slip flows, J. Phys. Chem. C, 122, 17, 9699-9709 (2018) |
| [49] | Tao, S.; Guo, Z., Boundary condition for lattice Boltzmann modeling of microscale gas flows with curved walls in the slip regime, Phys. Rev. E, 91, 4, Article 043305 pp. (2015) |
| [50] | Singh, S.; Jiang, F.; Tsuji, T., Impact of the kinetic boundary condition on porous media flow in the lattice Boltzmann formulation, Phys. Rev. E, 96, 1, Article 013303 pp. (2017) |
| [51] | Liu, Z.; Mu, Z.; Wu, H., A new curved boundary treatment for LBM modeling of thermal gaseous microflow in the slip regime, Microfluid. Nanofluid., 23, 2, 27 (2019) |
| [52] | Fang, W.; Tang, Y.; Ban, C.; Kang, Q.; Qiao, R.; Tao, W., Atomic layer deposition in porous electrodes: a pore-scale modeling study, Chem. Eng. J., 378, Article 122099 pp. (2019) |
| [53] | Shan, X.; He, X., Discretization of the velocity space in the solution of the Boltzmann equation, Phys. Rev. Lett., 80, 1, 65 (1998) |
| [54] | Shan, X.; Yuan, X.; Chen, H., Kinetic theory representation of hydrodynamics: a way beyond the Navier-Stokes equation, J. Fluid Mech., 550, 413-441 (2006) · Zbl 1097.76061 |
| [55] | Guo, Z.; Shi, B.; Wang, N., Lattice BGK model for incompressible Navier-Stokes equation, J. Comput. Phys., 165, 1, 288-306 (2000) · Zbl 0979.76069 |
| [56] | Chen, S.; Xu, K., A comparative study of an asymptotic preserving scheme and unified gas-kinetic scheme in continuum flow limit, J. Comput. Phys., 288, 52-65 (2015) · Zbl 1351.76230 |
| [57] | Meng, X.; Guo, Z., Localized lattice Boltzmann equation model for simulating miscible viscous displacement in porous media, Int. J. Heat Mass Transf., 100, 767-778 (2016) |
| [58] | Guo, Z.; Zhao, T. S., Lattice Boltzmann model for incompressible flows through porous media, Phys. Rev. E, 66, 3, Article 036304 pp. (2002) |
| [59] | Guo, Z.; Zheng, C.; Shi, B., An extrapolation method for boundary conditions in lattice Boltzmann method, Phys. Fluids, 14, 6, 2007-2010 (2002) · Zbl 1185.76156 |
| [60] | Kovasznay, L., Laminar flow behind a two-dimensional grid, Proc. Camb. Philos. Soc., 44, 1, 58-62 (1948) · Zbl 0030.22902 |
| [61] | Qian, Y.; d’Humières, D.; Lallemand, P., Lattice BGK models for Navier-Stokes equation, Europhys. Lett., 17, 6, 479 (1992) · Zbl 1116.76419 |
| [62] | Kingma, D. P.; Ba, J., Adam: a method for stochastic optimization (2014), arXiv preprint |
| [63] | Byrd, R. H.; Lu, P.; Nocedal, J.; Zhu, C., A limited memory algorithm for bound constrained optimization, SIAM J. Sci. Comput., 16, 5, 1190-1208 (1995) · Zbl 0836.65080 |
| [64] | Yang, L.; Hanneke, S.; Carbonell, J., A theory of transfer learning with applications to active learning, Mach. Learn., 90, 2, 161-189 (2013) · Zbl 1260.68352 |
| [65] | Lu, J.; Behbood, V.; Hao, P.; Zuo, H.; Xue, S.; Zhang, G., Transfer learning using computational intelligence: a survey, Knowl.-Based Syst., 80, 14-23 (2015) |
| [66] | McClenny, L.; Braga-Neto, U., Self-adaptive physics-informed neural networks using a soft attention mechanism (2020), arXiv preprint |
| [67] | Shen, J., Dynamics of regularized cavity flow at high Reynolds numbers, Appl. Math. Lett., 2, 4, 381-384 (1989) · Zbl 0703.76024 |
| [68] | Shen, J., Hopf bifurcation of the unsteady regularized driven cavity flow, J. Comput. Phys., 95, 1, 228-245 (1991) · Zbl 0725.76059 |
| [69] | Lee, T.; Lin, C., A characteristic Galerkin method for discrete Boltzmann equation, J. Comput. Phys., 171, 1, 336-356 (2001) · Zbl 1017.76043 |
| [70] | Yang, L.; Shu, C.; Yang, W.; Wu, J., An improved three-dimensional implicit discrete velocity method on unstructured meshes for all Knudsen number flows, J. Comput. Phys., 396, 738-760 (2019) · Zbl 1452.76214 |
| [71] | Galant, D., Gauss quadrature rules for the evaluation of \(2 \pi^{- 1 / 2} \int_0^\infty \exp(- x^2) f(x) d x\), Math. Comput., 23, 674 (1969), s31-s39 |
| [72] | Sone, Y.; Takata, S.; Ohwada, T., Numerical analysis of the plane Couette flow of a rarefied gas on the basis of the linearized Boltzmann equation for hard sphere molecules, Eur. J. Mech. B, Fluids, 9, 273-288 (1990) · Zbl 0696.76089 |
| [73] | Hinton, G.; Srivastava, N.; Swersky, K., Neural networks for machine learning lecture 6a overview of mini-batch gradient descent, Cited on, 14, 8 (2012) |
| [74] | Li, M.; Zhang, T.; Chen, Y.; Smola, A. J., Efficient mini-batch training for stochastic optimization, (Proceedings of the 20th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (2014)), 661-670 |
| [75] | Chen, S.; Zhang, C.; Zhu, L.; Guo, Z., A unified implicit scheme for kinetic model equations. Part I. Memory reduction technique, Sci. Bull., 62, 2, 119-129 (2017) |
| [76] | Scarselli, F.; Tsoi, A. C., Universal approximation using feedforward neural networks: a survey of some existing methods, and some new results, Neural Netw., 11, 1, 15-37 (1998) |
| [77] | Elsken, T.; Metzen, J. H.; Hutter, F., Neural architecture search: a survey, J. Mach. Learn. Res., 20, 55, 1-21 (2019) · Zbl 1485.68229 |
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.
