Unsupervised deep learning for super-resolution reconstruction of turbulence

Abstract Recent attempts to use deep learning for super-resolution reconstruction of turbulent flows have used supervised learning, which requires paired data for training. This limitation hinders more practical applications of super-resolution reconstruction. Therefore, we present an unsupervised learning model that adopts a cycle-consistent generative adversarial network (CycleGAN) that can be trained with unpaired turbulence data for super-resolution reconstruction. Our model is validated using three examples: (i) recovering the original flow field from filtered data using direct numerical simulation (DNS) of homogeneous isotropic turbulence; (ii) reconstructing full-resolution fields using partially measured data from the DNS of turbulent channel flows; and (iii) generating a DNS-resolution flow field from large-eddy simulation (LES) data for turbulent channel flows. In examples (i) and (ii), for which paired data are available for supervised learning, our unsupervised model demonstrates qualitatively and quantitatively similar performance as that of the best supervised learning model. More importantly, in example (iii), where supervised learning is impossible, our model successfully reconstructs the high-resolution flow field of statistical DNS quality from the LES data. Furthermore, we find that the present model has almost universal applicability to all values of Reynolds numbers within the tested range. This demonstrates that unsupervised learning of turbulence data is indeed possible, opening a new door for the wide application of super-resolution reconstruction of turbulent fields.

[1]  George Em Karniadakis,et al.  Hidden fluid mechanics: Learning velocity and pressure fields from flow visualizations , 2020, Science.

[2]  Yoshua Bengio,et al.  Generative Adversarial Nets , 2014, NIPS.

[3]  K. Taira,et al.  Super-resolution reconstruction of turbulent flows with machine learning , 2018, Journal of Fluid Mechanics.

[4]  Koji Fukagata,et al.  Synthetic turbulent inflow generator using machine learning , 2018, Physical Review Fluids.

[5]  Yi Li,et al.  Data exploration of turbulence simulations using a database cluster , 2007, Proceedings of the 2007 ACM/IEEE Conference on Supercomputing (SC '07).

[6]  Nils Thuerey,et al.  A Multi-Pass GAN for Fluid Flow Super-Resolution , 2019, PACMCGIT.

[7]  Yuji Hattori,et al.  Searching for turbulence models by artificial neural network , 2016, 1607.01042.

[8]  Jian Sun,et al.  Deep Residual Learning for Image Recognition , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[9]  Karthik Duraisamy,et al.  A paradigm for data-driven predictive modeling using field inversion and machine learning , 2016, J. Comput. Phys..

[10]  Kai Fukami,et al.  Assessment of supervised machine learning methods for fluid flows , 2020 .

[11]  Michele Milano,et al.  Neural network modeling for near wall turbulent flow , 2002 .

[12]  Guigang Zhang,et al.  Deep Learning , 2016, Int. J. Semantic Comput..

[13]  Katepalli R. Sreenivasan,et al.  A perspective on machine learning in turbulent flows , 2020 .

[14]  L. Biferale,et al.  Synchronization to Big Data: Nudging the Navier-Stokes Equations for Data Assimilation of Turbulent Flows , 2019, Physical Review X.

[15]  Jian Sun,et al.  Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification , 2015, 2015 IEEE International Conference on Computer Vision (ICCV).

[16]  Yi Li,et al.  A public turbulence database cluster and applications to study Lagrangian evolution of velocity increments in turbulence , 2008, 0804.1703.

[17]  Prakash Vedula,et al.  Subgrid modelling for two-dimensional turbulence using neural networks , 2018, Journal of Fluid Mechanics.

[18]  Steven L. Brunton,et al.  Robust principal component analysis for modal decomposition of corrupt fluid flows , 2019, Physical Review Fluids.

[19]  Sergey Ioffe,et al.  Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift , 2015, ICML.

[20]  Changhoon Lee,et al.  Prediction of turbulent heat transfer using convolutional neural networks , 2019, Journal of Fluid Mechanics.

[21]  Michael Chertkov,et al.  Embedding Hard Physical Constraints in Neural Network Coarse-Graining of 3D Turbulence , 2020, 2002.00021.

[22]  Omer San,et al.  A neural network approach for the blind deconvolution of turbulent flows , 2017, Journal of Fluid Mechanics.

[23]  Petros Koumoutsakos,et al.  Machine Learning for Fluid Mechanics , 2019, Annual Review of Fluid Mechanics.

[24]  Xi-yun Lu,et al.  Deep learning methods for super-resolution reconstruction of turbulent flows , 2020, Physics of Fluids.

[25]  P. Holmes,et al.  The Proper Orthogonal Decomposition in the Analysis of Turbulent Flows , 1993 .

[26]  Alexei A. Efros,et al.  Unpaired Image-to-Image Translation Using Cycle-Consistent Adversarial Networks , 2017, 2017 IEEE International Conference on Computer Vision (ICCV).

[27]  Simon Osindero,et al.  Conditional Generative Adversarial Nets , 2014, ArXiv.

[28]  Jinlong Wu,et al.  Physics-informed machine learning approach for reconstructing Reynolds stress modeling discrepancies based on DNS data , 2016, 1606.07987.

[29]  Karthik Duraisamy,et al.  Turbulence Modeling in the Age of Data , 2018, Annual Review of Fluid Mechanics.

[30]  J. Rabault,et al.  Performing particle image velocimetry using artificial neural networks: a proof-of-concept , 2017 .

[31]  Jean Rabault,et al.  Artificial neural networks trained through deep reinforcement learning discover control strategies for active flow control , 2018, Journal of Fluid Mechanics.

[32]  Petros Koumoutsakos,et al.  Efficient collective swimming by harnessing vortices through deep reinforcement learning , 2018, Proceedings of the National Academy of Sciences.

[33]  Aaron C. Courville,et al.  Improved Training of Wasserstein GANs , 2017, NIPS.

[34]  M. P. Brenner,et al.  Perspective on machine learning for advancing fluid mechanics , 2019, Physical Review Fluids.

[35]  J. Smagorinsky,et al.  GENERAL CIRCULATION EXPERIMENTS WITH THE PRIMITIVE EQUATIONS , 1963 .

[36]  R. Goodman,et al.  Application of neural networks to turbulence control for drag reduction , 1997 .

[37]  P. Moin,et al.  Turbulence statistics in fully developed channel flow at low Reynolds number , 1987, Journal of Fluid Mechanics.

[38]  Max Welling,et al.  Semi-Supervised Classification with Graph Convolutional Networks , 2016, ICLR.

[39]  Claus-Dieter Munz,et al.  Deep Neural Networks for Data-Driven Turbulence Models , 2018, J. Comput. Phys..

[40]  J. Templeton,et al.  Reynolds averaged turbulence modelling using deep neural networks with embedded invariance , 2016, Journal of Fluid Mechanics.

[41]  Donghyun You,et al.  Data-driven prediction of unsteady flow over a circular cylinder using deep learning , 2018, Journal of Fluid Mechanics.

[42]  Changhoon Lee,et al.  Deep unsupervised learning of turbulence for inflow generation at various Reynolds numbers , 2020, J. Comput. Phys..

[43]  Hossein Azizpour,et al.  Predictions of turbulent shear flows using deep neural networks , 2019, Physical Review Fluids.

[44]  Diego Donzis,et al.  Dissipation, enstrophy and pressure statistics in turbulence simulations at high Reynolds numbers , 2012, Journal of Fluid Mechanics.

[45]  Kunihiko Taira,et al.  Machine-learning-based spatio-temporal super resolution reconstruction of turbulent flows , 2020, Journal of Fluid Mechanics.

[46]  Myoungkyu Lee,et al.  A Web services accessible database of turbulent channel flow and its use for testing a new integral wall model for LES , 2016 .

[47]  Chao Xu,et al.  Particle Image Velocimetry Based on a Deep Learning Motion Estimator , 2020, IEEE Transactions on Instrumentation and Measurement.

[48]  A. W. Vreman.  An eddy-viscosity subgrid-scale model for turbulent shear flow: Algebraic theory and applications , 2004 .

[49]  R. Vinuesa,et al.  Prediction of wall-bounded turbulence from wall quantities using convolutional neural networks , 2019, Journal of Physics: Conference Series.

[50]  J. Nathan Kutz,et al.  Deep learning in fluid dynamics , 2017, Journal of Fluid Mechanics.

[51]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[52]  Chao Ma,et al.  Artificial neural network approach to large-eddy simulation of compressible isotropic turbulence. , 2019, Physical review. E.

[53]  Paris Perdikaris,et al.  Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations , 2019, J. Comput. Phys..

[54]  Timo Aila,et al.  A Style-Based Generator Architecture for Generative Adversarial Networks , 2018, 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[55]  Kurt Hornik,et al.  Multilayer feedforward networks are universal approximators , 1989, Neural Networks.

[56]  Parviz Moin,et al.  Direct simulations of turbulent flow using finite-difference schemes , 1991 .

[57]  Yingzheng Liu,et al.  Super-resolution reconstruction of turbulent velocity fields using a generative adversarial network-based artificial intelligence framework , 2019 .

[58]  L. Biferale,et al.  Reconstruction of turbulent data with deep generative models for semantic inpainting from TURB-Rot database , 2020, Physical Review Fluids.