Abstract
We propose a physics informed deep learning neural network, termed kεNet, for discovering the closure model of Reynolds stress in Reynolds-averaged Navier-Stokes (RANS) equations. The kεNet consists of a traditional physics uninformed neural network and several physics informed equations, e.g., the formula of Reynolds stress, the transport equations of turbulence kinetic energy k and turbulence dissipation rate ε. A low Reynolds number k-ε correction for turbulent channel flow is considered as an example. Based on the kεNet, all model parameters of the RANS are calibrated simultaneously with the specified mathematical formula, through training the neural network with data from direct numerical simulation. The gained turbulence model is utilized in OpenFOAM, an open source computational fluid dynamics (CFD) code, and predicts reasonable good results for a turbulent channel flow with Reτ = 5200 and 2000.
摘要
我们提出了一种基于物理信息的深度学习网络(kεNet), 可用于RANS方程中发现封闭的湍流模型. kεNet由一个传统的典型神经网络结构和若干个基于物理信息的方程组成, 如雷诺应力方程、k方程和ε方程. 以低雷诺数下的槽道流动的湍流模型的修正为例,通过训练基于物理信息的神经网络, 模型参数得到了修正. 修正后的湍流模型参数应用于OpenFOAM软件进行计算, 能够非常好地预测Reτ = 5200 和2000下的槽道流动.
References
J. Slotnick, A. Khodadoust, J. Alonso, D. Darmofal, W. Gropp, E. Lurie, and D. Mavriplis, CFD vision 2030 study: A path to revolutionary computational aerosciences, Langley Research Center, 2014.
P. Spalart and S. Allmaras, in A one-equation turbulence model for aerodynamic flows: Proceedings of 30th Aerospace Sciences Meeting and Exhibit, American Institute of Aeronautics and Astronautics, Reno, 1992.
B. E. Launder, and D. B. Spalding, The numerical computation of turbulent flows, Comput. Methods Appl. Mech. Eng. 3, 269 (1974).
F. R. Menter, Two-equation eddy-viscosity turbulence models for engineering applications, AIAA J. 32, 1598 (1994).
W. P. Jones, and B. E. Launder, The prediction of laminarization with a two-equation model of turbulence, Int. J. Heat Mass Transfer 15, 301 (1972).
C. K. G. Lam, and K. Bremhorst, A modified form of the k-ε model for predicting wall turbulence, J. Fluids Eng. 103, 456 (1981).
G. Dufour, J. B. Cazalbou, X. Carbonneau, and P. Chassaing, Assessing rotation/curvature corrections to eddy-viscosity models in the calculations of centrifugal-compressor flows, J. Fluids Eng. 130, (2008).
P. R. Spalart, and M. Shur, On the sensitization of turbulence models to rotation and curvature, Aerosp. Sci. Tech. 1, 297 (1997).
B. A. Pettersson Reif, P. A. Durbin, and A. Ooi, Modeling rotational effects in eddy-viscosity closures, Int. J. Heat Fluid Flow 20, 563 (1999).
Y. Huang, and H. Ma, Extended intrinsic mean spin tensor for turbulence modelling in non-inertial frame of reference, Appl. Math. Mech.-Engl. Ed. 29, 1463 (2008).
W. Zhang, Z. Ma, Y. C. Yu, and H. X. Chen, Applied new rotation correction k-ω SST model for turbulence simulation of centrifugal impeller in the rotating frame of reference, J. Hydrodyn. 22, 393 (2010).
F. Stern, R. V. Wilson, H. W. Coleman, and E. G. Paterson, Comprehensive approach to verification and validation of CFD simulations-Part 1: Methodology and procedures, J. Fluids Eng. 123, 793 (2001).
R. V. Wilson, F. Stern, H. W. Coleman, and E. G. Paterson, Comprehensive approach to verification and validation of CFD simulationsPart 2: Application for Rans simulation of a cargo/container ship, J. Fluids Eng. 123, 803 (2001).
J. Ling, A. Kurzawski, and J. Templeton, Reynolds averaged turbulence modelling using deep neural networks with embedded invariance, J. Fluid Mech. 807, 155 (2016).
H. Xiao, J. L. Wu, J. X. Wang, R. Sun, and C. J. Roy, Quantifying and reducing model-form uncertainties in Reynolds-averaged Navier-Stokes simulations: A data-driven, physics-informed Bayesian approach, J. Comput. Phys. 324, 115 (2016), arXiv: 1508.06315.
J. X. Wang, J. L. Wu, and H. Xiao, Physics-informed machine learning approach for reconstructing Reynolds stress modeling discrepancies based on DNS data, Phys. Rev. Fluids 2, 034603 (2017), arXiv: 1606.07987.
Z. Zhang, X. Song, S. Ye, Y. Wang, C. Huang, Y. An, and Y. Chen, Application of deep learning method to Reynolds stress models of channel flow based on reduced-order modeling of DNS data, J. Hydrodyn. 31, 58 (2018).
W. Zhang, L. Zhu, Y. Liu, and J. Kou, Machine learning methods for turbulence modeling in subsonic flows over airfoils (2018), arXiv: 1806.05904.
J. Weatheritt, and R. Sandberg, A novel evolutionary algorithm applied to algebraic modifications of the RANS stress-strain relationship, J. Comput. Phys. 325, 22 (2016).
M. Schoepplein, J. Weatheritt, R. Sandberg, M. Talei, and M. Klein, Application of an evolutionary algorithm to LES modelling of turbulent transport in premixed flames, J. Comput. Phys. 374, 1166 (2018).
W. Zhang, B. Zhu, H. Xu, and Y. Wang, Applying extended intrinsic mean spin tensor in evolution algorithm for RANS modelling of turbulent rotating channel flow, J. Hydrodyn. 31, 1255 (2019).
R. L. Thompson, L. E. B. Sampaio, F. A. V. de Bragança Alves, L. Thais, and G. Mompean, A methodology to evaluate statistical errors in DNS data of plane channel flows, Comput. Fluids 130, 1 (2016).
J. Wu, H. Xiao, R. Sun, and Q. Wang, Reynolds-averaged Navier-Stokes equations with explicit data-driven Reynolds stress closure can be ill-conditioned, J. Fluid Mech. 869, 553 (2019), arXiv: 1803.05581.
M. Raissi, A. Yazdani, and G. E. Karniadakis, Hidden fluid mechanics: Learning velocity and pressure fields from flow visualizations, Science 367, 1026 (2020).
X. Jin, S. Cai, H. Li, and G. E. Karniadakis, NSFnets (Navier-Stokes flow nets): Physics-informed neural networks for the incompressible Navier-Stokes equations, arXiv: 1808.08952 (2018).
K. Shukla, P. C. D. Leoni, J. Blackshire, D. Sparkman, and G. E. Karniadakis, Physics-informed neural network for ultrasound nondestructive quantification of surface breaking cracks, arXiv: 2005.03596.
M. Raissi, Z. Wang, M. S. Triantafyllou, and G. E. Karniadakis, Deep learning of vortex induced vibrations, arXiv: 1808.08952.
Z. Wei, Z. Hui, and W. Yong, Physics-informed neural network for discovering the turbulent flow, arXiv: 2005.03596v1.
W. Liu, J. Fang, S. Rolfo, C. Moulinec, and D. R. Emerson, An iterative machine-learning framework for RANS turbulence modeling, Int. J. Heat Fluid Flow 90, 108822 (2021).
E. Dow and Q. Wang, in Quantification of structural uncertainties in the k-ω turbulence model: Proceedings of 52nd AIAA/ASME/ASCE/AHS/ASC Structures, structural dynamics and materials conference 19th AIAA/ASME/AHS adaptive structures conference, Denver, 2011.
P. Ramachandran, Swish: A self-gated activation function, arXiv: 1710.05941v1.
X. Glorot, A. Bordes, and Y. Bengio, in Deep sparse rectifier neural networks: Proceedings of the Fourteenth International Conference on Artificial Intelligence and Statistics (PMLR, Fort Lauderdale, 2011), pp. 315–323.
D. P. Kingma and J. Ba, Adam: A method for stochastic optimization (2014), arXiv: 1412.6980.
Y. Bengio, A. Courville, and P. Vincent, Representation learning: A review and new perspectives (2012), arXiv: 1206.5538.
R. D. Moser and M. Lee, https://turbulence.oden.utexas.edu/ (2013).
A. Hellsten, in Some improvements in menters k-ω SST turbulence model: Proedings of 29th AIAA, Fluid Dynamics Conference (American Institute of Aeronautics and Astronautics, 1998).
Acknowledgements
This work was supported by the National Natural Science Foundation of China (Grant No. 91852117), the foundation of National Key Laboratory of Science and Technology on Aerodynamic Design and Research (Grant No. 614220121040106), and Shanghai Rising-Star Program (Grant No. 19QC1400200)
Author information
Authors and Affiliations
Contributions
Bing Zhu designed the conceptualization of this research and do the funding acquisition of this research. Longfeng Hou did the technical investigation and the original draft writing of this work. Ying Wang offered the resources and did the editing of this article.
Corresponding author
Rights and permissions
About this article
Cite this article
Hou, L., Zhu, B. & Wang, Y. kεNet: discovering the turbulence model and applying for low Reynolds number turbulent channel flow. Acta Mech. Sin. 39, 322326 (2023). https://doi.org/10.1007/s10409-022-22326-x
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s10409-022-22326-x