Super-resolution reconstruction of turbulent flows with machine learning

We use machine learning to perform super-resolution analysis of grossly under-resolved turbulent flow field data to reconstruct the high-resolution flow field. Two machine learning models are developed, namely, the convolutional neural network (CNN) and the hybrid downsampled skip-connection/multi-s...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Journal of fluid mechanics Jg. 870; S. 106 - 120
Hauptverfasser: Fukami, Kai, Fukagata, Koji, Taira, Kunihiko
Format: Journal Article
Sprache:Englisch
Veröffentlicht: Cambridge, UK Cambridge University Press 10.07.2019
Schlagworte:
Accuracy
Artificial intelligence
Artificial neural networks
Computational fluid dynamics
Computer simulation
Cylinders
Data
Data compression
Datasets
Fluid dynamics
Fluid flow
Fluid mechanics
Homogeneous turbulence
Image reconstruction
JFM Papers
Laminar flow
Learning algorithms
Machine learning
Neural networks
Paper machines
Performance assessment
Physics
Resolution
Reynolds number
Spatial discrimination
Spatial resolution
Training
Turbulence
Turbulence models
Turbulent flow
Two dimensional models
United States > US
computational methods
wakes
homogeneous turbulence
ISSN:0022-1120, 1469-7645
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:We use machine learning to perform super-resolution analysis of grossly under-resolved turbulent flow field data to reconstruct the high-resolution flow field. Two machine learning models are developed, namely, the convolutional neural network (CNN) and the hybrid downsampled skip-connection/multi-scale (DSC/MS) models. These machine learning models are applied to a two-dimensional cylinder wake as a preliminary test and show remarkable ability to reconstruct laminar flow from low-resolution flow field data. We further assess the performance of these models for two-dimensional homogeneous turbulence. The CNN and DSC/MS models are found to reconstruct turbulent flows from extremely coarse flow field images with remarkable accuracy. For the turbulent flow problem, the machine-leaning-based super-resolution analysis can greatly enhance the spatial resolution with as little as 50 training snapshot data, holding great potential to reveal subgrid-scale physics of complex turbulent flows. With the growing availability of flow field data from high-fidelity simulations and experiments, the present approach motivates the development of effective super-resolution models for a variety of fluid flows.
Bibliographie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0022-1120
1469-7645
DOI:10.1017/jfm.2019.238