Deep learning at scale for subgrid modeling in turbulent flows

  title={Deep learning at scale for subgrid modeling in turbulent flows},
  author={Mathis Bode and Michael Gauding and Konstantin Kleinheinz and Heinz Pitsch},
  booktitle={ISC Workshops},
Modeling of turbulent flows is still challenging. One way to deal with the large scale separation due to turbulence is to simulate only the large scales and model the unresolved contributions as done in large-eddy simulation (LES). This paper focuses on two deep learning (DL) strategies, regression and reconstruction, which are data-driven and promising alternatives to classical modeling concepts. Using three-dimensional (3-D) forced turbulence direct numerical simulation (DNS) data, subgrid… 

Using Physics-Informed Super-Resolution Generative Adversarial Networks for Subgrid Modeling in Turbulent Reactive Flows

This work presents a novel subgrid modeling approach based on a generative adversarial network (GAN), which is trained with unsupervised deep learning (DL) using adversarial and physics-informed losses to improve the generalization capability, especially extrapolation, of the network.

Automated Dissipation Control for Turbulence Simulation with Shell Models

This work constructs a strongly simplified representation of turbulence by using the Gledzer-OhkitaniYamada shell model and proposes an approach that aims to reconstruct statistical properties of turbulence such as the self-similar inertial-range scaling, where it could achieve encouraging experimental results.

Applying Physics-Informed Enhanced Super-Resolution Generative Adversarial Networks to Turbulent Premixed Combustion and Engine-like Flame Kernel Direct Numerical Simulation Data

The recently developed PIESRGAN modeling approach to turbulent premixed combustion can very accurately reproduce direct numerical simulation data on much coarser meshes, which is hardly possible with classical subfilter models, and enables studying statistical processes more ef ficiently due to the smaller computing cost.

Sparse identification of multiphase turbulence closures for coupled fluid–particle flows

Abstract In this work, model closures of the multiphase Reynolds-averaged Navier–Stokes (RANS) equations are developed for homogeneous, fully developed gas–particle flows. To date, the majority of

Multiphase turbulence modeling using sparse regression and gene expression programming

This work proposes an approach that blends sparse regression and gene expression programming (GEP) to generate closed-form algebraic models from simulation data to capture the physics of turbulence closure modeling.

Applying Physics-Informed Enhanced Super-Resolution Generative Adversarial Networks to Finite-Rate-Chemistry Flows and Predicting Lean Premixed Gas Turbine Combustors

The modeling approach of PIESRGAN is modified to accurately account for the challenges in the context of laminar finite-rate-chemistry flows and a reduced PiesRGAN-based model is presented that solves only the major species on a reconstructed �áeld and employs PIERSGAN lookup for the remaining species, utilizing staggering in time.

Applying Physics-Informed Enhanced Super-Resolution Generative Adversarial Networks to Turbulent Non-Premixed Combustion on Non-Uniform Meshes and Demonstration of an Accelerated Simulation Workflow

This paper extends the methodology to use physics-informed enhanced super-resolution generative adversarial networks (PIESR-GANs) for LES subfilter modeling in turbulent flows with finite-rate chemistry

Deep Learning for Efficient Reconstruction of High-Resolution Turbulent DNS Data

A novel deep learning framework SR-DNS Net is introduced, which aims to mitigate this inherent trade-off between solution fidelity and computational complexity by leveraging deep learning techniques used in image super-resolution.



Neural Networks for Data-Based Turbulence Models

This work defines the exact closure terms including the discretization operators and generates training data from direct numerical simulations of decaying homogeneous isotropic turbulence to construct a stable and accurate LES model from the learned closure terms.

Large-Eddy Simulation of an Idealized Tropical Cyclone

©2009 American Meteorological Society A dvances in computing have allowed presentday numerical weather-prediction models to forecast and/or simulate tropical cyclones (TCs) with high resolution

A dynamic subgrid‐scale eddy viscosity model

One major drawback of the eddy viscosity subgrid‐scale stress models used in large‐eddy simulations is their inability to represent correctly with a single universal constant different turbulent

A neural network approach for the blind deconvolution of turbulent flows

The proposed blind deconvolution network performs exceptionally well in the a priori testing of two-dimensional Kraichnan, three-dimensional Kolmogorov and compressible stratified turbulence test cases, and shows promise in forming the backbone of a physics-augmented data-driven closure for the Navier–Stokes equations.

Energy Cascade in Large-Eddy Simulations of Turbulent Fluid Flows

Towards Prediction of Turbulent Flows at High Reynolds Numbers Using High Performance Computing Data and Deep Learning

In this paper, deep learning methods are evaluated in the context of turbulent flows and qualitatively good agreement between DNS input data and generated turbulent structures is shown.

Deep learning in fluid dynamics

  • J. Kutz
  • Computer Science
    Journal of Fluid Mechanics
  • 2017
Although neural networks have been applied previously to complex fluid flows, the article featured here is the first to apply a true DNN architecture, specifically to Reynolds averaged Navier Stokes turbulence models, suggesting that DNNs may play a critically enabling role in the future of modelling complex flows.

Predictions of turbulent shear flows using deep neural networks

The proposed machine-learning framework may underpin future applications aimed at developing accurate and efficient data-driven subgrid-scale models for large-eddy simulations of more complex wall-bounded turbulent flows, including channels and developing boundary layers.