High Order Lagrangian Velocity Statistics in a Turbulent Channel Flow with Re τ = 80

2012 ◽  
Vol 24 (2) ◽  
pp. 287-291 ◽  
Author(s):  
Jian-ping Luo ◽  
Xiang Qiu ◽  
Dong-mei Li ◽  
Yu-lu Liu
2010 ◽  
Vol 22 (S1) ◽  
pp. 215-218 ◽  
Author(s):  
Jian-ping Luo ◽  
Zhi-ming Lu ◽  
TatsLo Ushijima ◽  
Osami Kitoh ◽  
Xiang Qiu ◽  
...  

2006 ◽  
Vol 96 (2) ◽  
Author(s):  
Haitao Xu ◽  
Mickaël Bourgoin ◽  
Nicholas T. Ouellette ◽  
Eberhard Bodenschatz

Author(s):  
Nojan Bagheri-Sadeghi ◽  
Brian T. Helenbrook ◽  
Kenneth D. Visser

Abstract One-dimensional fully developed channel flow was solved using a modified k–ω turbulence model that was recently proposed for use with high-order finite element schemes. In order to study this new turbulence model’s behavior, determine its dependence on boundary conditions and model constants, and find efficient methods for obtaining solutions, the model was first examined using a linear finite element discretization in 1D. The results showed that an accurate estimate of the parameter εk which is used to define k in terms of the working variable k~ is essential to get an accurate solution. Also, the turbulence model depended sensitively on an accurate estimate of the distance of the first grid point from the wall, which can be difficult to estimate in unstructured grids. This is used for the boundary condition of specific dissipation rate on the wall. This model was then implemented in a high-order finite element code that uses an unstructured mesh of triangles to verify that the 1D results were predictive of the behavior of the full 2D discretization. High-order 2D results were obtained on triangular meshes with element aspect ratios up to 250000.


2012 ◽  
Vol 11 (4) ◽  
pp. 1311-1322 ◽  
Author(s):  
Lihao Zhao ◽  
Helge I. Andersson

AbstractParticle dynamics in a turbulent channel flow is considered. The effects of particle concentration and Reynolds number on the particle velocity statistics are investigated. Four different particle response times, τ+=1, 5, 30 and 100, are examined for three different Reynolds numbers, Re*=200, 360 and 790 (based on channel height and friction velocity). The particle concentration evolves with time and statistics obtained during three different sampling periods might be distinctly different. The mean and fluctuating particle velocities are substantially affected both by the particle response time and by the Reynolds number of the flow.


Author(s):  
Daniel A. Khalitov ◽  
Ellen K. Longmire

The response of glass beads was examined in a fully developed turbulent channel flow of air with Reynolds number Reh = 4500 where h is the channel half width. Five sets of monodisperse particles were tested with integral Stokes numbers in the range 0.2–10 at the channel centerplane. Particle-to-air mass loadings were of order 10%. Laser sheets were aligned in streamwise-spanwise flow planes, and seeded, particle-laden flow was imaged with a dual-frame camera. A separation algorithm was used to obtain simultaneous gas and particle velocity measurements of homogeneous flow planes over a range of wall-normal distances. Profiles of single-point velocity statistics including mean and rms slip velocity, particle drift velocity, and gas-particle correlation were computed. In addition, two-point gas-particle and particle-particle velocity correlations were obtained. Near the wall, the gas-gas correlations yielded long, narrow patterns related to near-wall low-speed streaks. At the centerplane, the correlations were broader in the spanwise direction. Gas-particle and particle-particle correlations resembled the gas correlations in shape and size, but the values were smaller. The gas-particle covariance tended toward zero for St = 10. Particle-particle correlations were still significant for St = 4.


Sign in / Sign up

Export Citation Format

Share Document