Learning physics-constrained subgrid-scale closures in the small-data regime for stable and accurate LES

Abstract

We demonstrate how incorporating physics constraints into convolutional neural networks (CNNs) enables learning subgrid-scale (SGS) closures for stable and accurate large-eddy simulations (LES) in the small-data regime (i.e., when the availability of high-quality training data is limited). Using several setups of forced 2D turbulence as the testbeds, we examine the {\it a priori} and {\it a posteriori} performance of three methods for incorporating physics: 1) data augmentation (DA), 2) CNN with group convolutions (GCNN), and 3) loss functions that enforce a global enstrophy-transfer conservation (EnsCon). While the data-driven closures from physics-agnostic CNNs trained in the big-data regime are accurate and stable, and outperform dynamic Smagorinsky (DSMAG) closures, their performance substantially deteriorate when these CNNs are trained with 40x fewer samples (the small-data regime). We show that CNN with DA and GCNN address this issue and each produce accurate and stable data-driven closures in the small-data regime. Despite its simplicity, DA, which adds appropriately rotated samples to the training set, performs as well or in some cases even better than GCNN, which uses a sophisticated equivariance-preserving architecture. EnsCon, which combines structural modeling with aspect of functional modeling, also produces accurate and stable closures in the small-data regime. Overall, GCNN+EnCon, which combines these two physics constraints, shows the best {\it a posteriori} performance in this regime. These results illustrate the power of physics-constrained learning in the small-data regime for accurate and stable LES.

23 pages, 9 figures

References in corpus (26)

• Physics-Constrained Deep Learning for High-dimensional Surrogate Modeling and Uncertainty Quantification without Labeled Data
• Machine learning accelerated computational fluid dynamics
• Geometric Deep Learning: Grids, Groups, Graphs, Geodesics, and Gauges
• Perspectives on Machine Learning-augmented Reynolds-averaged and Large Eddy Simulation Models of Turbulence
• Scientific multi-agent reinforcement learning for wall-models of turbulent flows
• Bridging observation, theory and numerical simulation of the ocean using Machine Learning
• A Perspective on Machine Learning Methods in Turbulence Modelling
• Physics-guided deep learning framework for predictive modeling of the Reynolds stress anisotropy
• Stable a posteriori LES of 2D turbulence using convolutional neural networks: Backscattering analysis and generalization to higher Re via transfer learning
• Physical consistency of subgrid-scale models for large-eddy simulation of incompressible turbulent flows
• Physics-constrained, low-dimensional models for MHD: First-principles and data-driven approaches
• Incorporating Symmetry into Deep Dynamics Models for Improved Generalization
• Data-driven subgrid-scale modeling of forced Burgers turbulence using deep learning with generalization to higher Reynolds numbers via transfer learning
• Turbulence closure modeling with data-driven techniques: physical compatibility and consistency considerations
• Sparse nonlinear models of chaotic electroconvection
• Reduced-Order Models for Coupled Dynamical Systems: Data-driven Methods and the Koopman Operator
• Data-driven turbulence modeling in separated flows considering physical mechanism analysis
• Physical invariance in neural networks for subgrid-scale scalar flux modeling
• Energy and enstrophy dissipation in steady state 2-d turbulence
• Interpreting neural network models of residual scalar flux
• Invariant Data-Driven Subgrid Stress Modeling in the Strain-Rate Eigenframe for Large Eddy Simulation
• Data-driven Identification of 2D Partial Differential Equations using extracted physical features
• Frame invariant neural network closures for Kraichnan turbulence
• A machine learning framework for LES closure terms
• Inserting machine-learned virtual wall velocity for large-eddy simulation of turbulent channel flows
• Ensemble Kalman Inversion for Sparse Learning of Dynamical Systems from Time-Averaged Data

Cited by in corpus (16)

• Deep Reinforcement Learning for Turbulence Modeling in Large Eddy Simulations
• A posteriori learning for quasi-geostrophic turbulence parametrization
• Explaining the physics of transfer learning a data-driven subgrid-scale closure to a different turbulent flow
• Deep learning-enhanced ensemble-based data assimilation for high-dimensional nonlinear dynamical systems
• Generative data-driven approaches for stochastic subgrid parameterizations in an idealized ocean model
• Learning Closed-form Equations for Subgrid-scale Closures from High-fidelity Data: Promises and Challenges
• A Multifidelity deep operator network approach to closure for multiscale systems
• Discretize first, filter next: learning divergence-consistent closure models for large-eddy simulation
• A recursive neural-network-based subgrid-scale model for large eddy simulation: application to homogeneous isotropic turbulence
• Subgrid parameterizations of ocean mesoscale eddies based on Germano decomposition
• Reduced Data-Driven Turbulence Closure for Capturing Long-Term Statistics
• Towards Physically Consistent Deep Learning For Climate Model Parameterizations
• Guided Unconditional and Conditional Generative Models for Super-Resolution and Inference of Quasi-Geostrophic Turbulence
• A probabilistic approach to drift estimation from stochastic data
• A new data-driven energy-stable Evolve-Filter-Relax model for turbulent flow simulation
• Addressing A Posteriori Performance Degradation in Neural Network Subgrid Stress Models