粗糙壁面湍流边界层流动和拟序结构的大涡模拟研究

采用LES方法和虚拟粗糙元模型,,模拟了Rer=180有粗糙壁面的充分发展槽道湍流流动.从流向平均速度、脉动速度均方根、雷诺剪切应力、脉动速度相关,以及瞬时近壁准流向涡结构等多个方面,对粗糙壁面湍流边界层流动和拟序结构进行了分析.结果表明,粗糙壁面处雷诺剪切应力、脉动速度均方根以及近壁条带平均距离和准流向涡尺度都大于光滑壁面,且随着粗糙元高度的增加这种趋势更加明显.本工作为进一步研究颗粒在近壁区域的弥散规律奠定了基础.     

横向粗糙元壁面槽道湍流的大涡模拟研究

采用大涡模拟和浸没边界法相结合对不同高度和不同间距横向粗糙元壁面槽道湍流进行了模拟,得到了光滑壁面和粗糙壁面湍流的流向平均速度分布,雷诺剪切应力,脉动速度均方根和近壁区拟序结构。结果发现横向粗糙元降低了流向平均速度,增大了流动阻力,粗糙壁面湍流的雷诺剪切应力大于光滑壁面。粗糙元降低了流向脉动速度,增强了展向和法向脉动速度。粗糙元高度越高,对湍流流动影响越大,而粗糙元间距对湍流统计特性的影响不大。粗糙壁面仍然存在着和光滑壁面类似的条带结构。     

扰流板作用下湍流/非湍流界面特性

在零压梯度平板湍流边界层流动中安装垂直流向高度为h的扰流板,诱导流场产生横向涡,研究横向涡影响下湍流/非湍流界面特性沿流向的发展。结果表明,在本实验条件下,整个流动经历了从湍流边界层到流动分离和再附,再向湍流边界层恢复的过程。在扰流板下游约18h距离后,扰流板尾迹的影响逐渐衰退,壁面剪切对湍流强度的贡献开始逐渐恢复,在扰流板下游约55h距离后,湍流边界层再次充分发展。与此同时,由于扰流板后流场流动结构拟序性的增强,湍流/非湍流界面的分形维度受扰流板影响而减小,表明脱落涡有使界面多尺度特性、三维性减弱的趋势。此外,界面高度的概率密度分布受扰流板影响呈现显著的右偏,主要与扰流板增强喷射运动强度,使得界面更容易抬升相关。流动结构及界面特性受扰流板影响后的流向演化有同步变化的模式,扰流板对界面特性影响主要集中于（-5~18）h的流向范围。      

确定分布的展向Lorentz力调制下的槽道湍流涡结构

采用直接数值模拟方法,对槽道湍流中确定分布的Lorentz力的流动控制与减阻问题进行研究.讨论了Lorentz力作用于槽道湍流后,流场的特性和涡结构的特性,并对此类Lorentz力对槽道湍流的控制与减阻机理进行了讨论.研究发现:1)Lorentz力诱导的层流流场壁面附近存在梯度极大的展向速度剪切层,该剪切层容易形成流向涡结构;2)在给定合适参数的确定分布的Lorentz力作用下,湍流流场仅剩周期分布的准流向涡;3)与未控制流场相比,控制后的流场中,准流向涡的抬升高度大大降低,从而减小猝发强度,使壁面阻力下降.     

扩压叶栅端壁角区流向涡对湍流特性的影响

本文用三维激光多普勒测速技术测量了平面扩压叶栅的叶片吸力面角区流动湍流特性，如紊流动能分布、雷诺应力分布等．分析了角区内流向涡和通道涡对湍流特性分布的影响，结果表明：流向涡的核心区域及旋涡与附面层相互作用区域的湍流动能增加．同时，旋涡和旋涡运动强烈影响着雷诺正应力和切应力的分布规律．     

平板湍流边界层瞬时摩擦阻力的光学测量和统计分析

采用近壁粒子图像测速（particle image velocimetry，PIV）技术实现了中低Reynolds数光滑平板湍流边界层瞬时摩擦阻力的测量.发展了单行互相关算法和迭代拟合技术估算瞬时摩擦阻力，其法向空间分辨率可达到2~4 pixel.作为对比，同时分析了相似Reynolds数范围的直接数值模拟数据.统计结果表明:摩擦阻力的脉动强度随Reynolds数升高呈现对数增长趋势，证实了外区大/超大尺度结构对壁面的影响.谱分析结果表明近壁区的低速条带结构主导着摩擦阻力的多尺度脉动特性，外区大/超大尺度结构对摩擦阻力的印迹作用较小，其主要通过调制作用实现对摩擦阻力的影响，这一结论可以从瞬时摩擦阻力的偏度以及概率密度函数曲线具有Reynolds数不变性得到进一步证实.      

粗糙壁面上湍流边界层的直接数值模拟

本文采用多重直接力内嵌边界法计算流体与粗糙壁面之间的相互作用力,直接数值模拟了粗糙壁面上湍流边界层的流动状况。粗糙壁面则通过流向和展向分别叠加正弦函数的方法产生。研究发现,峰值较大的位置优先诱导产生出涡结构,并且加大了涡结构的倾斜角度。此外,较大峰值处的粗糙度所产生出的"喷出"和"扫入"事件更加强烈,其对外层的作用范围也就更大。同时粗糙壁面的峰值大小对于雷诺应力的输运也起着至关重要的作用。     

两相可压缩边界层的颗粒运动特性分析

本文采用高精度直接数值模拟方法,研究了马赫数为2的可压缩湍流边界层中,两种不同粒径颗粒的运动特性.研究发现近壁面的颗粒流向平均速度要小于流体;颗粒具有法向指向壁面的平均速度;近壁面颗粒的流向雷诺正应力要略大于流体:其中大颗粒具有更加显著的上述趋势.小颗粒的法向和展向雷诺正应力基本等于流体,而大颗粒则都要远小于流体.条件平均统计显示,大颗粒在低速区域的富集导致了这一现象.     

平板通道内充分发展湍流直接数值模拟

采用有限差分法对平板通道内充分发展湍流进行了直接模拟(DNS),Reτ=150,网格数64×64×64。得到了湍流脉动速度沿y 分布的等值线图,分析了近壁面脉动特性研究的重要性,为理解湍流物理机制提供数据。获得了雷诺应力的变化、雷诺应力输运方程中各项的平衡关系以及湍动能方程中各项的变化曲线,为进一步评价和改进湍流模式提供依据。     

正弦波沟槽对湍流边界层相干结构影响的TR-PIV实验研究

利用高时间分辨率粒子图像测速(time-resolved particle image velocimetry, TR-PIV)技术,在不同雷诺数下对光滑壁面和二维顺流向、三维正弦波(two/three dimensional, 2D/3D)沟槽壁面湍流边界层流场进行了实验测量,从不同沟槽对湍流边界层相干结构影响的角度分析了其减阻的机理.对比不同壁面的各阶统计量结果发现:沟槽降低了壁面摩擦阻力,存在减阻效果,正弦波沟槽的减阻率增大.运用相关函数、λ_(ci)检测准则等方法提取了不同壁面湍流边界层发卡涡和低速条带等典型相干结构的空间拓扑形态,结果表明:两种沟槽壁面的相干结构在流向和法向上的空间尺度均有不同程度的减小,且相干结构与主流之间的倾角趋于更小,流体在法向上的运动及结构的抬升受到明显抑制,发卡涡诱导喷射和扫掠的能力降低,从而影响了湍流中能量与动量的输运过程及湍流的自维持机制,且相比于2D沟槽, 3D正弦波沟槽作用效果更为明显.在同一雷诺数下,随着距离壁面法向位置的增加,不同壁面湍流边界层低速条带的展向间距都变宽;但同一法向位置处2D/3D沟槽壁面湍流边界层低速条带的间距与光滑壁面的相比更宽,沟槽的存在有效抑制了低速条带在展向上的运动,使得低速条带更稳定.     

不同光线入射角度下超声速湍流边界层气动光学效应的实验研究

利用基于纳米粒子的平面激光散射技术获取超声速(Ma=3.0)湍流边界层的密度分布,采用光线追迹方法计算其对应的光程差分布,并结合边界层气动光学相似律验证实验结果的可靠性.着重研究了光线入射角度对超声速湍流边界层气动光学效应的影响,并对其内在机理进行了分析.研究表明,气动光学效应对光线入射角度的依赖性源于光线在流场中的传输路径,传输路径的不同导致了光线在流场中的传输距离以及对应密度脉动互相关结果的差异.光线倾斜入射导致其在流场中传输距离增长,进而气动光学效应出现恶化.光线入射方向与壁面垂直方向之间的夹角越大,气动光学效应越显著,而且不同时刻的差异性增加,气动光学效应校正的难度增加.超声速湍流边界层中大量具有特定方向的涡结构导致了湍流边界层气动光学效应的各向异性.当光线倾斜向下游入射时,光线传播方向与流场中的涡结构具有较好的一致性,体现为此方向上密度脉动互相关系数较大,故气动光学效应比较严重.而当光线倾斜向上游入射时,相关系数较小,故气动光学效应较弱.     

高聚物减阻溶液对壁湍流输运过程的影响

基于相同雷诺数下清水和高分子聚合物溶液壁湍流的高时间分辨率粒子图像测速技术(time-resolved particle image velocimetry, TRPIV)的对比实验, 从高聚物溶液对湍流边界层动量能量输运影响的角度分析其减阻的机理. 对比两者的雷诺应力发现高聚物的存在抑制了湍流输运过程. 这一影响与高聚物对壁湍流中占主导地位的涡旋运动和低速条带等相干结构的作用密切相关. 运用条件相位平均、相关函数和线性随机估计(linear stochastic estimation, LSE)等方法, 分析提取了高聚物溶液流场中的发卡涡和发卡涡包等典型相干结构的空间拓扑形态. 相比于清水, 高聚物溶液中相干结构的流向尺度增大, 涡旋运动的发展及低速流体喷射的强度受到削弱, 表明了添加的高聚物阻碍了湍流原有的能量传递和自维持的机理. 正是通过影响相干结构, 高聚物抑制了湍流边界层中近壁区与外区之间的动量和能量输运, 使得湍流的无序性降低, 从而减小了湍流流动的阻力.     

Turbulent plane wall jets over smooth and rough surfaces

Large-eddy simulations were carried out to study the effects of surface roughness on a plane wall-jet using the Lagrangian dynamic eddy-viscosity subgrid-scale model, at Re = 7500 (based on the jet bulk velocity and height). Results over both smooth and rough surfaces were validated by experimental data at the same Reynolds number. As the jet is injected into the still environment, large-scale rollers are generated in the shear layer between the high-momentum fluid of the jet and the surrounding and are convected downstream with the flow. To understand the extent to which the outer-layer structures modify the flow in the inner layer and the extent to which the effect of roughness spreads away from the wall, both instantaneous and mean flow fields were investigated. The results revealed that, for the Reynolds number and roughness height considered in this study, the effect of roughness is mostly confined to the near-wall region of the wall jet. There is no structural difference between the outer layer of the wall jet over the smooth and rough surfaces. Roughness does not affect the size of the outer-layer structures or the scaling of the profiles of Reynolds stresses in the outer layer. However, in the inner layer, roughness redistributes stresses from streamwise to wall-normal and spanwise directions toward isotropy. Contours of joint probability-density function of the streamwise and wall-normal velocity fluctuations at the bottom of the logarithmic region match those of the turbulent boundary layer at the same height; while the traces of the outer-layer structure were detected at the top of the logarithmic region, indicating that they do not affect the flow very close to the wall, but still modify a major portion of the inner layer. This modification must be taken into consideration when the inner layer of a wall jet is compared with the conventional turbulent boundary layer.     

A wave field synthesis approach to reproduction of spatially correlated sound fields

This article discusses an open-loop wave field synthesis (WFS) approach for the reproduction of spatially correlated sound fields. The main application concerns laboratory reproduction of turbulent boundary layer wall pressure on aircraft fuselages and measurement of their sound transmission loss. The problem configuration involves reconstruction of random sound pressure distributions on a planar reproduction surface using a planar array of reproduction monopoles parallel to the reproduction plane. In this paper, the WFS formulation is extended to sound fields with imposed time and spatial correlation properties (or equivalently imposed cross-spectral density in the frequency and wave number domains). Numerical examples are presented for the reproduction of a propagating plane wave, diffuse acoustic field and wall pressure in subsonic or supersonic turbulent boundary layers. The reproduction accuracy is examined in terms of the size of the source plane and reproduction plane, their separation, and the number of reproduction sources required per acoustic wavelength. While the reproduction approach cannot reconstruct sub-wavelength correlation scales of subsonic turbulent boundary layers, it effectively reconstructs correlation scales larger than the acoustic wavelength, making it appropriate for diffuse acoustic field and supersonic turbulent layers.     

Reynolds number invariance of the structure inclination angle in wall turbulence

Cross correlations of the fluctuating wall-shear stress and the streamwise velocity in the logarithmic region of turbulent boundary layers are reported over 3 orders of magnitude change in Reynolds number. These results are obtained using hot-film and hot-wire anemometry in a wind tunnel facility, and sonic anemometers and a purpose-built wall-shear stress sensor in the near-neutral atmospheric surface layer on the salt flats of Utah's western desert. The direct measurement of fluctuating wall-shear stress in the atmospheric surface layer has not been available before. Structure inclination angles are inferred from the cross correlation results and are found to be invariant over the large range of Reynolds number. The findings justify the prior use of low Reynolds number experiments for obtaining structure angles for near-wall models in the large-eddy simulation of atmospheric surface layer flows.     

Surface roughness effects on the turbulence statistics in a low Reynolds number channel flow

A high-resolution particle image velocimetry was used to characterize a low Reynolds number turbulent flow in a channel. Experiments were conducted over a sand grain-coated surface of large relative roughness, and the results were compared with measurements over a smooth surface. The roughness perturbation significantly modified the outer layer. Even though the streamwise Reynolds stress shows less sensitivity in the outer layer to the boundary condition, significant enhancements were observed in the wall-normal Reynolds stress and the Reynolds shear stress. These modifications were considered as footprints of the larger-scale eddies transporting intense wall-normal motions away from the rough wall. A quadrant decomposition shows that strong and more frequent ejections are responsible for the larger values of the mean Reynolds shear stress over the rough wall. The results also indicate that spanwise vortex cores with mean vorticity of the same sign as the mean shear are the dominant smaller-scale vortical structures over the smooth and rough walls. A linear stochastic estimation-based analysis shows that the average larger-scale structure associated with these vortices is a shear layer that strongly connects the outer layer flow to the near-wall flow. A proper orthogonal decomposition of the flow suggests that the large-scale eddy is more energetic for the rough wall, and contributes more significantly to the resolved turbulent kinetic energy and the Reynolds shear stress than the smooth wall.     

Level-crossing statistics and production in low Reynolds number wall turbulence

Statistical properties at velocity level crossings are analysed in a low Reynolds number fully developed turbulent channel flow. Emphasis is placed on local production statistics conditioned by fixed amplitudes of the streamwise u and wall-normal v velocity components. Direct numerical simulations performed in large computational domains are used for this purpose. The Reynolds number based on the channel half width and the shear velocity varies from 180 to 1100. Particular attention is paid to correctly determine the conditional quantities at level crossings of u and v velocity fluctuations, to prevent biasing effects. Level crossings along the longitudinal x and spanwise z directions in homogeneous planes are introduced together with different characterisations, such as directional and contour crossings. The frequency of events detected at fixed thresholds follows a Gaussian model acceptably well. There is more level-crossing activity in the spanwise direction than in the streamwise. The mean conditional production distributions are dissymmetrical with respect to the level-crossing thresholds in the low buffer and viscous sublayers. These statistics differ significantly from a statistical model that assumes joint normality between u and v. There is a clear but relatively mild Reynolds number dependence of the conditional expected means of u and v that do not scale with inner, outer or mixed variables. The Reynolds number sensitivity of these statistics increases towards the edge of the wall layer.     

Characteristics of large-scale and superstructure motions in a turbulent boundary layer overlying complex roughness

The structural attributes of turbulent flow over a complex roughness topography are explored using high-frame-rate stereo particle-image velocimetry measurements in the wall-normal–spanwise plane. The roughness under consideration was replicated from a turbine blade damaged by deposition of foreign materials and contains a broad range of topographical scales arranged in a highly irregular manner. Previous results from Barros and Christensen [Observations of turbulent secondary flows in a rough-wall boundary layer. J Fluid Mech. 2014;748] revealed strong spanwise heterogeneity in the flow attributed to the formation of roughness-induced turbulent secondary flows identified by spanwise-alternating low- and high-momentum flow pathways (HMP & LMP, respectively) in the mean flow marked by enhanced Reynolds stresses and turbulent kinetic energy. Frequency spectra of streamwise velocity at fixed wall-normal location presented herein also display strong dependence on spanwise position. In particular, the roughness promotes enhanced energy content of the large-scale and smaller-scale motions (as opposed to very-large-scale ones). Depending on spanwise position, pre-multiplied spectra highlight significant modification of the energy content of the very large-scale motions (superstructures) due to roughness compared to smooth-wall flow. Of note, a shift in both TKE and RSS content to shorter streamwise scales at an LMP was noted, while less of an impact was found coincident with an HMP.     

Analytical model of the time developing turbulent boundary layer

An analytical model for the time-developing turbulent boundary layer (TD TBL) over a flat plate is presented. The model provides explicit formulae for the temporal behavior of the wall-shear stress and both the temporal and spatial distributions of the mean streamwise velocity, the turbulence kinetic energy and Reynolds shear stress. The resulting profiles are in good agreement with the DNS results of spatially-developing turbulent boundary layers at momentum thickness Reynolds numbers equal to 1430 and 2900 [5–7]. Our analytical model is, to the best of our knowledge, the first of its kind for TD TBL. The text was submitted by the authors in English.