☆ 4.7 Article

A priori assessment of nonlocal data-driven wall modeling in large eddy simulation

PHYSICS OF FLUIDS (2023)

Related references

Note: Only part of the references are listed.

Article Physics, Multidisciplinary

Frame invariant neural network closures for Kraichnan turbulence

Suraj Pawar et al.

Summary: Numerical simulations of geophysical and atmospheric flows rely on imperfect parameterization models for subgrid scale processes, leading to inaccurate predictions. To improve accuracy and generalization, a frame invariant closure approach is proposed, embedding physical symmetries directly into the neural network structure. The framework accurately predicts subgrid scale source terms, respects physical symmetries, and demonstrates numerical stability in coarse-grid simulations, opening up possibilities for connecting physics-based theories and data-driven modeling.

PHYSICA A-STATISTICAL MECHANICS AND ITS APPLICATIONS (2023)

Add to Collection

Review Mathematics, Applied

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

Yifei Guan et al.

Summary: This study demonstrates 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. Three methods for incorporating physics were examined: data augmentation (DA), CNN with group convolutions (GCNN), and loss functions that enforce a global enstrophy-transfer conservation (EnsCon). The results show that DA, GCNN, and EnsCon can all produce accurate and stable data-driven closures in the small-data regime, with GCNN+EnsCon showing the best performance.

PHYSICA D-NONLINEAR PHENOMENA (2023)

Add to Collection

Article Mechanics

Log-law recovery through reinforcement-learning wall model for large eddy simulation

Aurelien Vadrot et al.

Summary: This paper focuses on the use of reinforcement learning (RL) for near-wall turbulence modeling. A new RL wall model (WM) called VYBA23 is developed, which uses agents dispersed in the flow near the wall. The model is trained on a single Reynolds number and does not rely on high-fidelity data. The results show potential for developing RLWMs that can recover physical laws and for extending this type of ML models to more complex flows in the future.

PHYSICS OF FLUIDS (2023)

Add to Collection

Article Physics, Fluids & Plasmas

Survey of machine-learning wall models for large-eddy simulation

Aurelien Vadrot et al.

Summary: This survey investigated the use of data-driven machine learning techniques for wall modeling in large-eddy simulations (LES). Three machine learning wall models were implemented and compared with an equilibrium wall model in LES of half-channel flow at various Reynolds numbers. Results showed promise in data-driven ML wall models, although some models had errors at certain Reynolds numbers.

PHYSICAL REVIEW FLUIDS (2023)

Add to Collection

Article Computer Science, Artificial Intelligence

Generalization techniques of neural networks for fluid flow estimation

Masaki Morimoto et al.

Summary: This paper explores techniques to promote the practical use of neural networks in fluid flow estimation, focusing on challenges such as interpretability of machine-learned results, bulking out of training data, and generalizability of neural networks. The study demonstrates methods to enhance interpretability and generalizability, as well as techniques to increase training data for fluid flow problems, indicating promising results for applications of machine learning in fluid dynamics.

NEURAL COMPUTING & APPLICATIONS (2022)

Add to Collection

Article Computer Science, Interdisciplinary Applications

Stable a posteriori LES of 2D turbulence using convolutional neural networks: Backscattering analysis and generalization to higher Re via transfer learning

Yifei Guan et al.

Summary: Developing machine learning-based data-driven subgrid-scale models for large eddy simulation is of growing interest. Prior tests show that deep convolutional neural networks accurately predict the inter-scale transfers, and transfer learning enhances the stability and accuracy of the model for different flow conditions.

JOURNAL OF COMPUTATIONAL PHYSICS (2022)

Add to Collection

Article Mechanics

Development of subgrid-scale model for LES of Burgers turbulence with large filter size

A. Golsa Tabe Jamaat et al.

Summary: In this study, a neural network is implemented to establish the relationship between the resolved-scale flow field and the subgrid-scale stress for the Burgers equation. The training datasets are obtained by filtering the direct numerical simulation results and different sets of input parameters are considered. The a priori and a posteriori tests show that the neural network models have high correlation coefficients with the filtered DNS data and can produce stable results without any stabilization techniques.

PHYSICS OF FLUIDS (2022)

Add to Collection

Article Multidisciplinary Sciences

Scientific multi-agent reinforcement learning for wall-models of turbulent flows

H. Jane Bae et al.

Summary: Researchers propose a multi-agent reinforcement learning approach to discover wall models for large-eddy simulations, solving the challenge of capturing near-wall dynamics in turbulent flow simulations.

NATURE COMMUNICATIONS (2022)

Add to Collection

Article Nanoscience & Nanotechnology

Deep learning method for the super-resolution reconstruction of small-scale motions in large-eddy simulation

Qingyi Zhao et al.

Summary: This paper proposes a super-resolution reconstruction model for the subgrid scale turbulent flow field in large-eddy simulation (LES) using a meta-learning deep convolutional neural network (MLDCNN). The model achieves good results in reconstructing the LES flow field and improves the prediction accuracy of LES.

AIP ADVANCES (2022)

Add to Collection

Article Nanoscience & Nanotechnology

Investigation of nonlocal data-driven methods for subgrid-scale stress modeling in large eddy simulation

Bo Liu et al.

Summary: A nonlocal subgrid-scale stress (SGS) model based on convolutional neural network (CNN) is developed and trained with direct numerical simulation (DNS) data. The results show that the nonlocal model performs better and is more robust than local data-driven models in large eddy simulation (LES) of complex anisotropic turbulent flows.

AIP ADVANCES (2022)

Add to Collection

Article Mechanics

Grid-point and time-step requirements for direct numerical simulation and large-eddy simulation

Xiang I. A. Yang et al.

Summary: This paper revisits the grid-point and time-step requirements for DNS and LES of turbulent boundary layers and establishes more general requirements. By considering the relationship between local grid spacing and the Kolmogorov length scale, accurate requirements for different simulation methods are proposed. According to the estimated costs, the computational costs of different simulation methods also vary.

PHYSICS OF FLUIDS (2021)

Add to Collection

Article Mechanics

Convolutional-network models to predict wall-bounded turbulence from wall quantities

Luca Guastoni et al.

Summary: Two convolutional neural network models were trained to predict velocity-fluctuation fields in turbulent open-channel flow, with one directly predicting fluctuations and the other reconstructing flow fields using a linear combination of orthonormal basis functions. Both models outperformed traditional methods in predicting nonlinear interactions in the flow.

JOURNAL OF FLUID MECHANICS (2021)

Add to Collection

Article Mechanics

Toward neural-network-based large eddy simulation: application to turbulent channel flow

Jonghwan Park et al.

Summary: The study developed a subgrid-scale model using a fully connected neural network and conducted both a priori and a posteriori tests to evaluate its prediction performance. Results showed that while the NN-based SGS model can provide good predictions under certain conditions, it also exhibited stability issues in some cases.

JOURNAL OF FLUID MECHANICS (2021)

Add to Collection

Article Mechanics

Data-driven subgrid-scale modeling of forced Burgers turbulence using deep learning with generalization to higher Reynolds numbers via transfer learning

Adam Subel et al.

Summary: Recent research has shown that deep neural networks trained using properly pre-conditioned data can generate stable and accurate a posteriori LES models, and transfer learning can enable accurate and stable generalization to flows with higher Reynolds numbers.

PHYSICS OF FLUIDS (2021)

Add to Collection

Article Physics, Fluids & Plasmas

Data assimilation empowered neural network parametrizations for subgrid processes in geophysical flows

Suraj Pawar et al.

Summary: This passage discusses the increasing use of machine learning approaches in representing subgrid-scale processes in geophysical flows to improve forecasting capability and accelerate numerical simulations. By utilizing data assimilation to correct physics-based models and incorporating neural networks for unresolved flow dynamics, significant improvements have been achieved in predicting chaotic dynamics and handling the non-Gaussian statistics of subgrid-scale processes.

PHYSICAL REVIEW FLUIDS (2021)

Add to Collection

Article Mechanics

Convolutional neural networks for fluid flow analysis: toward effective metamodeling and low dimensionalization

Masaki Morimoto et al.

Summary: This study focuses on the application of convolutional neural networks in fluid flow analysis, investigating the influence of various operations on performance and the placement of additional scalar inputs on the model. Two types of CNN-based fluid flow analyses, CNN metamodeling and CNN autoencoder, are discussed, providing insights into the design of CNN architecture for practical fluid flow analysis.

THEORETICAL AND COMPUTATIONAL FLUID DYNAMICS (2021)

Add to Collection

Article Physics, Fluids & Plasmas

Wall model based on neural networks for LES of turbulent flows over periodic hills

Zhideng Zhou et al.

Summary: A data-driven wall model for turbulent flows over periodic hills is developed using feedforward neural network (FNN) and wall-resolved large-eddy simulation (WRLES) data. The FNN predictions show good agreement with WRLES data for the wall shear stresses, but some discrepancies are observed near the crest of the hill. Overall, the correlation coefficients between FNN predictions and WRLES predictions are larger than 0.7 at most streamwise locations.

PHYSICAL REVIEW FLUIDS (2021)

Add to Collection

Article Physics, Fluids & Plasmas

Effect of Wall Boundary Conditions on a Wall-Modeled Large-Eddy Simulation in a Finite-Difference Framework

H. Jane Bae et al.

Summary: The type of boundary condition used has an impact on statistics near the wall, with augmenting eddy viscosity at the wall improving predictions in the near-wall region and providing a simple solution to the well-known log-layer mismatch in WMLES.

FLUIDS (2021)

Add to Collection

Article Mechanics

Artificial neural network-based nonlinear algebraic models for large eddy simulation of turbulence

Chenyue Xie et al.

PHYSICS OF FLUIDS (2020)

Add to Collection

Article Mechanics

Deconvolutional artificial neural network models for large eddy simulation of turbulence

Zelong Yuan et al.

PHYSICS OF FLUIDS (2020)

Add to Collection

Article Physics, Fluids & Plasmas

Closure Learning for Nonlinear Model Reduction Using Deep Residual Neural Network

Xuping Xie et al.

FLUIDS (2020)

Add to Collection

Article Physics, Fluids & Plasmas

Predictive large-eddy-simulation wall modeling via physics-informed neural networks

X. I. A. Yang et al.

PHYSICAL REVIEW FLUIDS (2019)

Add to Collection

Article Mechanics

Wall-modeled large-eddy simulations of spanwise rotating turbulent channels-Comparing a physics-based approach and a data-based approach

Xinyi L. D. Huang et al.

PHYSICS OF FLUIDS (2019)

Add to Collection

Article Computer Science, Interdisciplinary Applications

Subgrid-scale model for large-eddy simulation of isotropic turbulent flows using an artificial neural network

Zhideng Zhou et al.

COMPUTERS & FLUIDS (2019)

Add to Collection

Article Mechanics

Dynamic slip wall model for large-eddy simulation

Hyunji Jane Bae et al.

JOURNAL OF FLUID MECHANICS (2019)

Add to Collection

Article Computer Science, Interdisciplinary Applications

Deep neural networks for data-driven LES closure models

Andrea Beck et al.

JOURNAL OF COMPUTATIONAL PHYSICS (2019)

Add to Collection

Article Mechanics

Investigations of data-driven closure for subgrid-scale stress in large-eddy simulation

Zhuo Wang et al.

PHYSICS OF FLUIDS (2018)

Add to Collection

Review Computer Science, Artificial Intelligence

Deep Convolutional Neural Networks for Image Classification: A Comprehensive Review

Waseem Rawat et al.

NEURAL COMPUTATION (2017)

Add to Collection

Article Physics, Fluids & Plasmas

Searching for turbulence models by artificial neural network

Masataka Gamahara et al.

PHYSICAL REVIEW FLUIDS (2017)

Add to Collection

Article Physics, Fluids & Plasmas

Space-time characteristics of wall-pressure and wall shear-stress fluctuations in wall-modeled large eddy simulation

George Ilhwan Park et al.

PHYSICAL REVIEW FLUIDS (2016)

Add to Collection

Article Mechanics

Integral wall model for large eddy simulations of wall-bounded turbulent flows

X. I. A. Yang et al.

PHYSICS OF FLUIDS (2015)

Add to Collection

Article Mechanics

A dynamic slip boundary condition for wall-modeled large-eddy simulation

S. T. Bose et al.

PHYSICS OF FLUIDS (2014)

Add to Collection

Article Mechanics

Estimating wall-shear-stress fluctuations given an outer region input

Romain Mathis et al.

JOURNAL OF FLUID MECHANICS (2013)

Add to Collection

Article Mechanics

Large eddy simulations of turbulent channel and boundary layer flows at high Reynolds number with mean wall shear stress boundary condition

Jungil Lee et al.

PHYSICS OF FLUIDS (2013)

Add to Collection

Article Mechanics

Grid-point requirements for large eddy simulation: Chapman's estimates revisited

Haecheon Choi et al.

PHYSICS OF FLUIDS (2012)

Add to Collection

Article Mechanics

Wall-modeling in large eddy simulation: Length scales, grid resolution, and accuracy

Soshi Kawai et al.

PHYSICS OF FLUIDS (2012)

Add to Collection

Article Thermodynamics

A hybrid RANS-LES approach with delayed-DES and wall-modelled LES capabilities

Mikhail L. Shur et al.

INTERNATIONAL JOURNAL OF HEAT AND FLUID FLOW (2008)

Add to Collection

Article Environmental Sciences

Large-eddy simulation of neutral atmospheric boundary layer flow over heterogeneous surfaces: Blending height and effective surface roughness

E Bou-Zeid et al.

WATER RESOURCES RESEARCH (2004)

Add to Collection

Article Thermodynamics

The inner-outer layer interface in large-eddy simulations with wall-layer models

U Piomelli et al.

INTERNATIONAL JOURNAL OF HEAT AND FLUID FLOW (2003)

Add to Collection

Article Mechanics

Dynamic wall modeling for large-eddy simulation of complex turbulent flows

M Wang et al.

PHYSICS OF FLUIDS (2002)

Add to Collection

Article Mechanics

Large eddy simulation wall-modeling based on suboptimal control theory and linear stochastic estimation

F Nicoud et al.

PHYSICS OF FLUIDS (2001)

Add to Collection

Letter Mechanics

An approach to wall modeling in large-eddy simulations

NV Nikitin et al.

PHYSICS OF FLUIDS (2000)

Add to Collection

Article Thermodynamics

Approximate wall boundary conditions in the large-eddy simulation of high reynolds number flow

W Cabot et al.

FLOW TURBULENCE AND COMBUSTION (2000)

Add to Collection

© Peeref 2019-2024. All rights reserved.