1,020 Hits in 8.9 sec

A deep learning framework for turbulence modeling using data assimilation and feature extraction [article]

Atieh Alizadeh Moghaddam, Amir Sadaghiyani
2018 arXiv   pre-print
Turbulent problems in industrial applications are predominantly solved using Reynolds Averaged Navier Stokes (RANS) turbulence models.  ...  We propose the use of deep learning algorithms via convolution neural networks along with data from direct numerical simulations to extract the optimal set of features that explain the evolution of turbulent  ...  Currently the predominant tool for engineering problems with turbulent flows are Reynolds Averaged Navier Stokes (RANS) models.  ... 
arXiv:1802.06106v1 fatcat:gabblgeum5bzrm7bth7banwysu

The Potential of Machine Learning to Enhance Computational Fluid Dynamics [article]

Ricardo Vinuesa, Steven L. Brunton
2021 arXiv   pre-print
Machine learning is rapidly becoming a core technology for scientific computing, with numerous opportunities to advance the field of computational fluid dynamics.  ...  In each of these areas, it is possible to improve machine learning capabilities by incorporating physics into the process, and in turn, to improve the simulation of fluids to uncover new physical understanding  ...  As described below, LES denotes large-eddy simulation and RANS Reynolds-averaged Navier-Stokes. Images reproduced from Refs. [32, 89, 118] with permission of the publishers.  ... 
arXiv:2110.02085v1 fatcat:fu4b2dym6bd53ihnpqprdpwefu

Learning Incompressible Fluid Dynamics from Scratch – Towards Fast, Differentiable Fluid Models that Generalize [article]

Nils Wandel, Michael Weinmann, Reinhard Klein
2021 arXiv   pre-print
of the fluid simulation to traditional methods.  ...  Recent deep learning based approaches promise vast speed-ups but do not generalize to new fluid domains, require fluid simulation data for training, or rely on complex pipelines that outsource major parts  ...  INTRODUCTION Simulating the behavior of fluids by solving the incompressible Navier-Stokes equations is of great importance for a wide range of applications and accurate as well as fast fluid simulations  ... 
arXiv:2006.08762v3 fatcat:6vfzottjgzdodjzx6l74mjcd5i

Towards Physics-informed Deep Learning for Turbulent Flow Prediction [article]

Rui Wang, Karthik Kashinath, Mustafa Mustafa, Adrian Albert, Rose Yu
2020 arXiv   pre-print
Specifically, we introduce trainable spectral filters in a coupled model of Reynolds-averaged Navier-Stokes (RANS) and Large Eddy Simulation (LES), followed by a specialized U-net for prediction.  ...  We adopt a hybrid approach by marrying two well-established turbulent flow simulation techniques with deep learning.  ...  ACKNOWLEDGEMENT This research used resources of the National Energy Research Scientific Computing Center, a DOE Office of Science User Facility supported by the Office of Science of the U.S.  ... 
arXiv:1911.08655v4 fatcat:qnjvnynhbja6rjhcmrx77r3jfu

Explore missing flow dynamics by physics-informed deep learning: the parameterised governing systems [article]

Hui Xu, Wei Zhang, Yong Wang
2021 arXiv   pre-print
On the other hand, detailed flow features are described by the governing equations, e.g. the Navier-Stokes equations for viscous fluid, and can be resolved numerically, which is heavily dependent on the  ...  As a reliable way in practice, especially for turbulent flows, regional flow information such as velocity and its statistics, can be measured experimentally.  ...  NNW2019ZT4-B09) and the National Natural Science Foundation of China (Grant NOs. 91852106).  ... 
arXiv:2008.12266v3 fatcat:wjacvxatt5fn5dmuxksa77x6p4

Deep Learning Methods for Reynolds-Averaged Navier-Stokes Simulations of Airfoil Flows [article]

Nils Thuerey, Konstantin Weissenow, Lukas Prantl, Xiangyu Hu
2019 arXiv   pre-print
With this study we investigate the accuracy of deep learning models for the inference of Reynolds-Averaged Navier-Stokes solutions.  ...  In addition all source code is publicly available in order to ensure reproducibility and to provide a starting point for researchers interested in deep learning methods for physics problems.  ...  Mainali for their help with the deep learning experiments.  ... 
arXiv:1810.08217v2 fatcat:yvrbcusv2jbp3o4jj75ocw6bhy

DPM: A deep learning PDE augmentation method (with application to large-eddy simulation) [article]

Jonathan B. Freund, Jonathan F. MacArt, Justin Sirignano
2019 arXiv   pre-print
The approach is demonstrated and evaluated for turbulence predictions using large-eddy simulation (LES), a filtered version of the Navier--Stokes equation containing unclosed sub-filter-scale terms.  ...  Once trained, the deep learning PDE model (DPM) can make out-of-sample predictions for new physical parameters, geometries, and boundary conditions.  ...  Blue Waters is a joint effort of the University of Illinois at Urbana-Champaign and its National Center for Supercomputing Applications.  ... 
arXiv:1911.09145v1 fatcat:j2twqx5d3zd37e2ceva4vgmjny

Deep Surrogate for Direct Time Fluid Dynamics [article]

Lucas Meyer
2021 arXiv   pre-print
The recent years have seen a surge of deep learning surrogate models substituting these solvers to alleviate the simulation process.  ...  The ubiquity of fluids in the physical world explains the need to accurately simulate their dynamics for many scientific and engineering applications.  ...  In this paper, we focus on Deep Surrogates where a neural network is trained to provide a quality solution to the Navier-Stokes equations for a given domain, initial and boundary conditions.  ... 
arXiv:2112.10296v1 fatcat:qpnyviwhy5hmvarhyftdjg2rxm

Deep learning in fluid dynamics

J. Nathan Kutz
2017 Journal of Fluid Mechanics  
Fluid Mech., vol. 807, 2016, pp. 155–166) is the first to apply a true DNN architecture, specifically to Reynolds averaged Navier Stokes turbulence models.  ...  It was only a matter of time before deep neural networks (DNNs) – deep learning – made their mark in turbulence modelling, or more broadly, in the general area of high-dimensional, complex dynamical systems  ...  Ling, Kurzawski & Templeton (2016) have proposed using DNNs for Reynolds averaged Navier Stokes (RANS) models which are widely used because of their computational tractability in modelling the rich set  ... 
doi:10.1017/jfm.2016.803 fatcat:w6vqfv7j2jhjpalnp5nhjqrw6a

Deep learning for turbulent channel flow [article]

Rui Fang, David Sondak, Pavlos Protopapas, Sauro Succi
2018 arXiv   pre-print
One of the most popular reduced models is the Reynolds averaged Navier-Stokes (RANS) equations. The goal is to solve the RANS equations for the mean velocity and pressure field.  ...  Its performance is compared with classical turbulence models as well as a neural network model that does not preserve Galilean invariance.  ...  Introducing (2.3) and (2.4) into the Navier-Stokes equations and averaging results in the Reynolds averaged Navier-Stokes (RANS) equations, ∂u ∂t + ∇ · (u ⊗ u) = − 1 ρ ∇p + ν∇ 2 u − ∇ · R (2.5) ∇ ·  ... 
arXiv:1812.02241v1 fatcat:tvffwf2fdzdupf7mxe2fhskrbi

NSFnets (Navier-Stokes Flow nets): Physics-informed neural networks for the incompressible Navier-Stokes equations [article]

Xiaowei Jin, Shengze Cai, Hui Li, George Em Karniadakis
2020 arXiv   pre-print
We refer to these specific PINNs for the Navier-Stokes flow nets as NSFnets.  ...  We also perform a systematic study on the weights used in the loss function for the data/physics components and investigate a new way of computing the weights dynamically to accelerate training and enhance  ...  In this study, instead of using conventional compu- tational fluid dynamics (CFD) methods, we investigate the possibility of using neural networks (NNs) for solving the Navier-Stokes equations.  ... 
arXiv:2003.06496v1 fatcat:rfe34muognc4nlgaymyckd2zum

9 From the POD-Galerkin method to sparse manifold models [chapter]

2020 Applications  
The leading method for model reduction in fluids is Galerkin projection of the Navier-Stokes equations onto a linear subspace of modes obtained via proper orthogonal decomposition (POD).  ...  Neural networks have long been used for flow modeling and control [74, 122, 56, 54] , and recently deep neural networks have been used for Reynolds-averaged turbulence modeling [59] .  ...  How accurate is it?  ... 
doi:10.1515/9783110499001-009 fatcat:ubjntebufbculkiayjaog344om

An Efficient Deep Learning Technique for the Navier-Stokes Equations: Application to Unsteady Wake Flow Dynamics [article]

Tharindu P. Miyanawala, Rajeev K. Jaiman
2018 arXiv   pre-print
We present an efficient deep learning technique for the model reduction of the Navier-Stokes equations for unsteady flow problems.  ...  A systematic convergence and sensitivity study is performed to identify the effective dimensions of the deep-learned CNN process such as the convolution kernel size, the number of kernels and the convolution  ...  Acknowledgements The first author thanks the National Research Foundation and Keppel Corporation, Singapore for supporting the work done in the Keppel-NUS Corporate Laboratory. References  ... 
arXiv:1710.09099v3 fatcat:yzrdqmsvf5bzzfzkv6dhnpnpee

The emerging role of large eddy simulation in industrial practice: challenges and opportunities

A.G. Hutton
2009 Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences  
There is a growing body of evidence that such methods offer a significant stretch in industrial capability over solely Reynolds-averaged Navier-Stokes (RANS)-based modelling.  ...  That class of methods for treating turbulence gathered under the banner of large eddy simulation is poised to enter mainstream engineering practice.  ...  Particular thanks are due to Dr Richard Ashworth of QinetiQ and Dr Jamil Appa of BAESYSTEMS for their advice, inspiration and deep knowledge.  ... 
doi:10.1098/rsta.2009.0077 pmid:19531504 fatcat:q3ndhgfy4rcy5ogb4outpxj54i

Multi-fidelity information fusion with concatenated neural networks [article]

Suraj Pawar, Omer San, Prakash Vedula, Adil Rasheed, Trond Kvamsdal
2021 arXiv   pre-print
We illustrate how the knowledge from these simplified models results in reducing uncertainties associated with deep learning models.  ...  Recently, computational modeling has shifted towards the use of deep learning, and other data-driven modeling frameworks.  ...  Acknowledgements This material is based upon work supported by the U.S. Department of Energy, Office of Science, Office of Advanced Scientific Computing Research under Award Number DE-SC0019290.  ... 
arXiv:2110.04170v1 fatcat:igabq3c735bjfdq5tmnjk3eq5a
« Previous Showing results 1 — 15 out of 1,020 results