Extended synthetic eddy method to generate inflow data for turbulent thermal boundary layer

Oh, Geunwoo; Noh, Kyung Min; Park, Hyunwook; Choi, Jung Il

doi:10.1016/j.ijheatmasstransfer.2019.02.061

Cited by 12 publications

(3 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…On the other hand, the synthetic inflow turbulence generation method involves generating turbulence quickly under a certain constraint to provide realistic turbulence fields in a short time. Several models, such as synthetic random Fourier method [8], synthetic digital filtering method [9], and synthetic coherent eddy method [10,11], have been proposed to generate flow with a proper spatiotemporal correlation to prevent rapid dissipation of turbulence [12]. Although these approaches produce reasonable inflow fields, the generation methods depend on ad hoc assumptions.…”

Section: Introductionmentioning

confidence: 99%

Deep unsupervised learning of turbulence for inflow generation at various Reynolds numbers

Kim

Lee

2020

Journal of Computational Physics

View full text Add to dashboard Cite

A B S T R A C TA realistic inflow boundary condition is essential for successful simulation of the developing turbulent boundary layer or channel flows. Recent advances in artificial intelligence (AI) have enabled the development of an inflow generator that performs better than the synthetic methods based on intuitions. In the present work, we applied generative adversarial networks (GANs), a representative of unsupervised learning, to generate an inlet boundary condition of turbulent channel flow. Upon learning the two-dimensional spatial structure of turbulence using data obtained from direct numerical simulation (DNS) of turbulent channel flow, the GAN could generate instantaneous flow fields that are statistically similar to those of DNS. Surprisingly, the GAN could produce fields at various Reynolds numbers without any additional simulation based on the trained data of only three Reynolds numbers. This indicates that the GAN could learn the universal nature of Reynolds number effect and might reflect other simulation conditions. Eventually, through a combination of the GAN and a recurrent neural network (RNN), we developed a novel model (RNN-GAN) that could generate time-varying fully developed flow for a long time. The spatiotemporal correlations of the generated flow are in good agreement with those of the DNS. This proves the usefulness of unsupervised learning in the generation of synthetic turbulence fields.

show abstract

Section: Introductionmentioning

confidence: 99%

Deep unsupervised learning of turbulence for inflow generation at various Reynolds numbers

Kim

Lee

2020

Journal of Computational Physics

View full text Add to dashboard Cite

show abstract

“…Different software packages have been developed to implement this method (for example, using Fortran, Oh et al (2019)). However, these packages are often limited in their applicability and can be challenging for non-experts to use.…”

Section: Statement Of Needmentioning

confidence: 99%

SyntheticEddyMethod.jl: A Julia package for the creation of inlet flow conditions for LES

Brunelli¹

2023

JOSS

View full text Add to dashboard Cite

The Synthetic Eddy Method (SEM) is a numerical simulation technique used to create a turbulent flow field with desired features. It is used in computational fluid dynamics for imposing realistic inlet boundary conditions, improving the fidelity of the results obtained by simulations. Its ability to create fluctuations with prescribed physical features makes it a valuable tool for researchers and engineers seeking to improve the reliability of simulations and also for re-creating as much as possible an environment close to an experimental one. The package allows users to easily generate synthetic turbulence fields that can be used in CFD simulations, and to control the level of turbulence and length-scales of the eddies in the generated fields.

show abstract

“…An easterly wind of 1.8 m/s magnitude is assumed and imposed as a fully developed turbulent flow at the inflow boundary. The inflow velocity field is perturbed by the vortex-generated method (Oh et al [66]) aiming to generate three-dimensional, unsteady inflow condition. The zero-gradient boundary condition is applied at the lateral sides and top, while pressure boundary condition is imposed at the outflow.…”

Section: Computational Domain and Boundary Conditionsmentioning

confidence: 99%

A rational hybrid RANS-LES model for CFD predictions of microclimate and environmental quality in real urban structures

Hadžiabdić

Hafizović

Ničeno

et al. 2022

Building and Environment

View full text Add to dashboard Cite

Extended synthetic eddy method to generate inflow data for turbulent thermal boundary layer

Cited by 12 publications

References 22 publications

Deep unsupervised learning of turbulence for inflow generation at various Reynolds numbers

Deep unsupervised learning of turbulence for inflow generation at various Reynolds numbers

SyntheticEddyMethod.jl: A Julia package for the creation of inlet flow conditions for LES

A rational hybrid RANS-LES model for CFD predictions of microclimate and environmental quality in real urban structures

Contact Info

Product

Resources

About