Compressed Convolutional LSTM: An Efficient Deep Learning framework to Model High Fidelity 3D Turbulence

Abstract

High-fidelity modeling of turbulent flows is one of the major challenges in computational physics, with diverse applications in engineering, earth sciences and astrophysics, among many others. The rising popularity of high-fidelity computational fluid dynamics (CFD) techniques like direct numerical simulation (DNS) and large eddy simulation (LES) have made significant inroads into the problem. However, they remain out of reach for many practical three-dimensional flows characterized by extremely large domains and transient phenomena. Therefore designing efficient and accurate data-driven generative approaches to model turbulence is a necessity. We propose a novel training approach for dimensionality reduction and spatio-temporal modeling of the three-dimensional dynamics of turbulence using a combination of Convolutional autoencoder and the Convolutional LSTM neural networks. The quality of the emulated turbulent fields is assessed with rigorous physics-based statistical tests, instead of visual assessments. The results show significant promise in the training methodology to generate physically consistent turbulent flows at a small fraction of the computing resources required for DNS.

References in corpus (9)

• Convolutional LSTM Network: A Machine Learning Approach for Precipitation Nowcasting
• Deep learning for universal linear embeddings of nonlinear dynamics
• PDE-Net: Learning PDEs from Data
• A Deep Learning based Approach to Reduced Order Modeling for Turbulent Flow Control using LSTM Neural Networks
• Deep Learning the Physics of Transport Phenomena
• Nonlinear Model Reduction for Complex Systems using Sparse Optimal Sensor Locations from Learned Nonlinear Libraries
• From Deep to Physics-Informed Learning of Turbulence: Diagnostics
• NeuralDrop: DNN-based Simulation of Small-Scale Liquid Flows on Solids
• Massively parallel implementation in Python of a pseudo-spectral DNS code for turbulent flows

Cited by in corpus (32)

• Reduced-order modeling of advection-dominated systems with recurrent neural networks and convolutional autoencoders
• A new prediction method of unsteady wake flow by the hybrid deep neural network
• Multi-level Convolutional Autoencoder Networks for Parametric Prediction of Spatio-temporal Dynamics
• Probabilistic neural networks for fluid flow surrogate modeling and data recovery
• Solver-in-the-Loop: Learning from Differentiable Physics to Interact with Iterative PDE-Solvers
• Time-series learning of latent-space dynamics for reduced-order model closure
• Zermelo's problem: Optimal point-to-point navigation in 2D turbulent flows using Reinforcement Learning
• Stable a posteriori LES of 2D turbulence using convolutional neural networks: Backscattering analysis and generalization to higher Re via transfer learning
• Deep unsupervised learning of turbulence for inflow generation at various Reynolds numbers
• Incorporating Symmetry into Deep Dynamics Models for Improved Generalization
• Reconstruction of turbulent data with deep generative models for semantic inpainting from TURB-Rot database
• Synchronization to big-data: nudging the Navier-Stokes equations for data assimilation of turbulent flows
• Reservoir computing model of two-dimensional turbulent convection
• Long short-term memory embedded nudging schemes for nonlinear data assimilation of geophysical flows
• Nonlinear proper orthogonal decomposition for convection-dominated flows
• Embedding Hard Physical Constraints in Neural Network Coarse-Graining of 3D Turbulence
• Latent-space time evolution of non-intrusive reduced-order models using Gaussian process emulation
• FiniteNet: A Fully Convolutional LSTM Network Architecture for Time-Dependent Partial Differential Equations
• Robust deep learning for emulating turbulent viscosities
• Machine Learning Changes the Rules for Flux Limiters
• A Koopman-based framework for forecasting the spatiotemporal evolution of chaotic dynamics with nonlinearities modeled as exogenous forcings
• Data assimilation empowered neural network parameterizations for subgrid processes in geophysical flows
• Validation and parameterization of a novel physics-constrained neural dynamics model applied to turbulent fluid flow
• Learning Stable Galerkin Models of Turbulence with Differentiable Programming
• Conditionally Parameterized, Discretization-Aware Neural Networks for Mesh-Based Modeling of Physical Systems
• Data-Driven Modeling of Geometry-Adaptive Steady Heat Transfer based on Convolutional Neural Networks: Heat Conduction
• Predicting Localized Primordial Star Formation with Deep Convolutional Neural Networks
• Development of Deep Learning Methods for Inflow Turbulence Generation
• Deep learning based on PINN for solving 2 D0F vortex induced vibration of cylinder with high Reynolds number
• Deep learning based on mixed-variable physics informed neural network for solving fluid dynamics without simulation data
• Unsupervised Machine Learning to Teach Fluid Dynamicists to Think in 15 Dimensions
• Physics-Informed Machine Learning of the Lagrangian Dynamics of Velocity Gradient Tensor