Filters








663 Hits in 2.4 sec

High-fidelity reconstruction of turbulent flow from spatially limited data using enhanced super-resolution generative adversarial network [article]

Mustafa Z. Yousif, Linqi Yu, HeeChang Lim
2021 arXiv   pre-print
A multi-scale enhanced super-resolution generative adversarial network with a physics-based loss function is introduced as a model to reconstruct the high-resolution flow fields.  ...  The model capability to reconstruct high-resolution laminar flows is examined using data of laminar flow around a square cylinder.  ...  In terms of experimental studies, Deng et al. 20 applied a super-resolution GAN (SRGAN) 21 and enhanced SRGAN (ESRGAN) 22 to reconstruct high-resolution flow fields using PIV measurements of flow  ... 
arXiv:2109.04250v2 fatcat:v6ombdpu65cq7gbezrzjd2dg6e

Application of Video-to-Video Translation Networks to Computational Fluid Dynamics

Hiromitsu Kigure
2021 Frontiers in Artificial Intelligence  
In particular, the time evolution of density distributions in the cases of a high-resolution grid is reproduced from that in the cases of a low-resolution grid through GANs, and the density inhomogeneity  ...  Qualitative and quantitative comparisons of the results of the proposed method with those of several super-resolution algorithms are also presented.  ...  Xie et al. (2018) proposed a method for super-resolution fluid flow by a temporally coherent generative model (tempoGAN).  ... 
doi:10.3389/frai.2021.670208 pmid:34568812 pmcid:PMC8461073 fatcat:3thstqntizdzhiwzju243nmuk4

GANs-based PIV resolution enhancement without the need of high-resolution input

Alejandro Güemes, Carlos Sanmiguel Vila, Stefano Discetti
2021 14th International Symposium on Particle Image Velocimetry  
A data-driven approach to reconstruct high-resolution flow fields is presented.  ...  The results prove that this data-driven method is able to enhance the resolution of PIV measurements even in complex flows without the need of a separate high-resolution experiment for training.  ...  This document reflects only the author's view and the Agency and the Commission are not responsible for any use that may be made of the information it contains.  ... 
doi:10.18409/ispiv.v1i1.160 fatcat:kjogib3ta5c7lnhkaao3ezccgu

tempoGAN: A Temporally Coherent, Volumetric GAN for Super-resolution Fluid Flow [article]

You Xie, Erik Franz, Mengyu Chu, Nils Thuerey
2018 arXiv   pre-print
We propose a temporally coherent generative model addressing the super-resolution problem for fluid flows.  ...  Our method works instantaneously, using only a single time-step of low-resolution fluid data.  ...  We would like to thank Wei He for helping with making the videos, and all members of the graphics labs of TUM, IST Austria and ETH Zurich for the thorough discussions.  ... 
arXiv:1801.09710v2 fatcat:eihainku2vh4zdsfghmz4gbbxm

2021 Index IEEE Transactions on Image Processing Vol. 30

2021 IEEE Transactions on Image Processing  
The Author Index contains the primary entry for each item, listed under the first author's name.  ...  Badoual, A., +, TIP 2021 5739-5753 Adversarial Multi-Path Residual Network for Image Super-Resolution.  ...  Direct Unsupervised Super-Resolution Using Generative Adversarial Network (DUS-GAN) for Real-World Data.  ... 
doi:10.1109/tip.2022.3142569 fatcat:z26yhwuecbgrnb2czhwjlf73qu

Super-Resolution of Near-Surface Temperature Utilizing Physical Quantities for Real-Time Prediction of Urban Micrometeorology [article]

Yuki Yasuda and Ryo Onishi and Yuichi Hirokawa and Dmitry Kolomenskiy and Daisuke Sugiyama
2021 arXiv   pre-print
The present paper proposes a super-resolution (SR) model based on a convolutional neural network and applies it to the near-surface temperature in urban areas.  ...  We train the SR model with sets of low-resolution (LR) and high-resolution (HR) images from building-resolving large-eddy simulations (LESs) in a city, where the horizontal resolutions of LR and HR are  ...  Thuerey, A multi-pass gan for fluid flow super-resolution, Proc. ACM Comput. Graph. Interact. Tech. 2 (7 2019). doi:10.1145/3340251.  ... 
arXiv:2108.00806v2 fatcat:mkfsrpyiwfhpzd3sq4sunkpj3u

Machine Learning for Physics-Informed Generation of Dispersed Multiphase Flow Using Generative Adversarial Networks [article]

B. Siddani, S. Balachandar, W. C. Moore, Y. Yang, R. Fang
2020 arXiv   pre-print
Fluid flow around a random distribution of stationary spherical particles is a problem of substantial importance in the study of dispersed multiphase flows.  ...  In this paper we present a machine learning methodology using Generative Adversarial Network framework and Convolutional Neural Network architecture to recreate particle-resolved fluid flow around a random  ...  [21] proposed their tempoGAN ML architecture, which is a temporally coherent generative adversarial network model, for obtaining super-resolution of fluid flows from an input of coarse-grained flow  ... 
arXiv:2005.05363v1 fatcat:a64v2iimbzg7npti5aor4up2lm

Deep learning of multi-resolution X-Ray micro-CT images for multi-scale modelling [article]

Samuel J. Jackson and Yufu Niu and Sojwal Manoorkar and Peyman Mostaghimi and Ryan T. Armstrong
2022 arXiv   pre-print
We develop a 3D Enhanced Deep Super Resolution (EDSR) convolutional neural network to create super resolution (SR) images from low resolution images, which alleviates common micro-CT hardware/reconstruction  ...  flow pressures and 3D fluid volume fractions).  ...  Although [43] did use optically generated HR data for training, their GAN approach was validated on unpaired data. [44] were the first to demonstrate a super-resolution method using paired LR and HR  ... 
arXiv:2111.01270v2 fatcat:7tkhayyazrazhawukflt6zdhxe

Neural Networks-Based Aerodynamic Data Modeling: A Comprehensive Review

Liwei Hu, Jun Zhang, Yu Xiang, Wenyong Wang
2020 IEEE Access  
Finally, we identify three important trends for future studies in aerodynamic data modeling: a) the transformation method and physics informed models will be combined to solve high-dimensional partial  ...  In this paper, we analyze the shortcomings of computational fluid dynamics (CFD) and traditional reduced-order models (ROMs).  ...  [24] proposed a novel generative model named tempoGAN to generate super-resolution flow field images. TempoGAN is a conditional GAN with double discriminators.  ... 
doi:10.1109/access.2020.2993562 fatcat:hsensfhkfvbafa2gbljmvt6csi

Stacked Generative Machine Learning Models for Fast Approximations of Steady-State Navier-Stokes Equations [article]

Shen Wang, Mehdi Nikfar, Joshua C. Agar, Yaling Liu
2021 arXiv   pre-print
A standard description of fluid dynamics requires solving the Navier-Stokes (N-S) equations in different flow regimes.  ...  To improve the resolution and predictability, we train stacked models of increasing complexity generating the numerical solutions for N-S equations.  ...  A Multi-Pass GAN for Fluid Flow Super-Resolution. Proc. ACM Comput. Graph. Interact. Tech. 2, 1–21 (July 2019). 35. Feng, J. et al.  ... 
arXiv:2112.06419v1 fatcat:hd4caoonive5dokqz7x2cqqq7u

Neural UpFlow: A Scene Flow Learning Approach to Increase the Apparent Resolution of Particle-Based Liquids [article]

Bruno Roy, Pierre Poulin, Eric Paquette
2021 arXiv   pre-print
We present a novel up-resing technique for generating high-resolution liquids based on scene flow estimation using deep neural networks.  ...  We also propose a particle-based approach to interpolate between liquids generated from varying simulation discretizations using a state-of-the-art bidirectional optical flow solver method for fluids in  ...  We would also like to give a special thanks to Nils Thuerey for his insightful advice, comments, and discussions surrounding this project during the research internship at the Technical University of Munich  ... 
arXiv:2106.05143v1 fatcat:yzb76xpbrrbstnyi6c7nwqnvby

Enforcing Statistical Constraints in Generative Adversarial Networks for Modeling Chaotic Dynamical Systems [article]

Jin-Long Wu, Karthik Kashinath, Adrian Albert, Dragos Chirila, Prabhat, Heng Xiao
2019 arXiv   pre-print
We exemplify this approach on the Rayleigh-Benard convection, a turbulent flow system that is an idealized model of the Earth's atmosphere.  ...  physics, which are known to be a major source of uncertainty in simulating multi-scale physical systems, e.g., turbulence or Earth's climate.  ...  To improve GANs performances for physical problems, Xie et al. [27] incorporated temporal coherence to GANs to generate super-resolution realizations of turbulent flows. Yang et al.  ... 
arXiv:1905.06841v1 fatcat:vwocwflaxnbqfld77r7fwt4dlu

Generative Modeling of Turbulence [article]

Claudia Drygala, Benjamin Winhart, Francesca di Mare, Hanno Gottschalk
2021 arXiv   pre-print
We present a mathematically well founded approach for the synthetic modeling of turbulent flows using generative adversarial networks (GAN).  ...  We furthermore simulate the flow around a low pressure turbine stator using the pix2pixHD architecture for a conditional DCGAN being conditioned on the position of a rotating wake in front of the stator  ...  Another application of GAN is the field of super-resolution reconstruction of turbulent flows.  ... 
arXiv:2112.02548v1 fatcat:h5k7a7g445dfpg5zqweunjfque

A novel topology design approach using an integrated deep learning network architecture [article]

Sharad Rawat, M.H. Herman Shen
2019 arXiv   pre-print
There is, therefore, a need for a different approach that will be able to optimize the initial design topology effectively and rapidly.  ...  Therefore, this work presents a new topology design procedure to generate optimal structures using an integrated Generative Adversarial Networks (GANs) and convolutional neural network architecture.  ...  GANs have many variations like a super-resolution GAN which enhances the resolution of the images improving the details in an image, changing the environment of the images (Cycle GAN) and many more.  ... 
arXiv:1808.02334v2 fatcat:trhsdhbupngqzgdkv4rfmzyuci

Multi-fidelity Generative Deep Learning Turbulent Flows [article]

Nicholas Geneva, Nicholas Zabaras
2020 arXiv   pre-print
In this work, a novel multi-fidelity deep generative model is introduced for the surrogate modeling of high-fidelity turbulent flow fields given the solution of a computationally inexpensive but inaccurate  ...  For both of these examples, the model is able to generate unique yet physically accurate turbulent fluid flows conditioned on an inexpensive low-fidelity solution.  ...  Computing resources were provided by the AFOSR Office of Scientific Research through the DURIP program and by the University of Notre Dame's Center for Research Computing (CRC).  ... 
arXiv:2006.04731v1 fatcat:alnvxmuyezh2tbtjn2zih6jaym
« Previous Showing results 1 — 15 out of 663 results