Improving separated-flow predictions using an anisotropy-capturing subgrid-scale model

The major conclusion of this paper is that resolution requirements for large-eddy simulation (LES) of flow separation and reattachment can be significantly reduced using the anisotropy-capturing explicit algebraic subgrid-scale (SGS) stress model (EASSM) of Marstorp et al. (J. Fluid Mech., vol. 639, 2009, pp. 403–432), instead of the conventional isotropic dynamic eddy-viscosity model (DEVM). LES of flow separation in a channel with streamwise periodic hill-shaped constrictions and spanwise homogeneity is performed at coarse resolutions for which it is observed that flow separation cannot be predicted without a SGS model and cannot be correctly predicted by the DEVM, while reasonable predictions are obtained with the EASSM. It is shown that the lower resolution requirements by the EASSM, compared to the DEVM, is not only due its nonlinear formulation, but also due to the better formulation of its eddy-viscosity part. The improvements obtained with the EASSM have previously been demonstrated using higher-order numerical solvers for channel flows. In this study, it is observed that these improvements still remain using a low-order code with significant inherent numerical dissipation.     

Turbulent boundary layer flow with a step change from smooth to rough surface

A direct numerical simulation (DNS) dataset of a turbulent boundary layer (TBL) with a step change from a smooth to a rough surface is analyzed to examine the characteristics of a spatially developing flow. The roughness elements are periodically arranged two-dimensional (2-D) spanwise rods, with the first rod placed 80θ_in downstream from the inlet, where θ_in denotes the inlet momentum thickness. Based on an accurate estimation of relevant parameters, clear evidence for mean flow universality is provided when scaled properly, even for the present roughness configuration, which is believed to have one of the strongest impacts on the flow. Compared to previous studies, it is shown that overshooting behavior is present in the first- and second-order statistics and is locally created either within the cavity or at the leading edge of the roughness depending on the type of statistics and the wall-normal measurement location. Inspection of spatial two-point correlations of the streamwise velocity fluctuations shows a continuous increase of spanwise length scales of structures over the rough wall after the step change at a greater growth rate than that over smooth wall TBL flow. This is expected because spanwise energy spectrum shows presence of much energetic wider structures over the rough wall. Full images of the DNS data are presented to describe not only predominance of hairpin vortices but also a possible spanwise scale growth mechanism via merging over the rough wall.     

Multi-scale analysis of subgrid stress and energy dissipation in turbulent channel flow

In present study, the subgrid scale （SGS） stress and dissipation for multiscale formulation of large eddy simulation are analyzed using the data of turbulent channel flow at Ret = 180 obtained by direct numerical simulation. It is found that the small scale SGS stress is much smaller than the large scale SGS stress for all the stress components. The dominant contributor to large scale SGS stress is the cross stress between small scale and subgrid scale motions, while the cross stress between large scale and subgrid scale motions make major contributions to small scale SGS stress. The energy transfer from resolved large scales to subgrid scales is mainly caused by SGS Reynolds stress, while that between resolved small scales and subgrid scales are mainly due to the cross stress. The multiscale formulation of SGS models are evaluated a priori, and it is found that the small- small model is superior to other variants in terms of SGS dissipation.     

Effects of two-dimensional V-shaped grooves on turbulent channel flow

Detailed Laser Doppler velocimeter (LDV) measurements have been carried out in a turbulent rectangular channel flow with one rough wall. The roughness elements of two-dimensional spanwise 120° V-shaped grooves are periodically arranged with different depths and pitches. The Reynolds number based on the centerline velocity, and the channel half height ranges from 2,740 to 20,000. The comparisons of turbulence statistics over smooth and rough walls indicate that the present roughness leads to a significant change in the turbulence both in the inner and in the outer flow. Particularly, the distribution density of the grooves is a key parameter to evaluate the effect of roughness. The low-Reynolds-number dependence of turbulence statistics is also observed. The rough walls with the same pitch-to-depth ratio exhibit the equivalent roughness function under the corresponding Reynolds numbers. The disagreement of velocity defect profiles between smooth and rough walls challenges the defect universal law. The variations of the turbulence stresses and Reynolds shear stress decomposition in the outer layer suggest that the turbulent motions may be modified by the present grooves. The importance of sweep events for the present groove-roughened walls is reflected by the differences in relative contribution to Reynolds shear stress from each quadrant and the higher-order moments over smooth and rough walls.     

基于全局动态Vreman模型的复杂非定常湍流的大涡模拟

在湍流数值模拟方法中,大涡模拟方法可以提供丰富的大涡旋信息,已逐渐成为复杂湍流问题数值研究的重要方法。而大涡模拟中,最重要的一环是尽量准确地构建能反映流场物理本质特征的亚格子应力模型。基于该思想,将一种新型的大涡模拟亚格子应力模型-Vreman亚格子应力模型用于高雷诺数三维后台阶流动的求解,计算结果与实验结果进行对比分析结果较吻合,验证了该模型的可靠性。这是对该模型用于无任何均匀流动方向的高雷诺数复杂湍流非定常流动的首次检验,计算结果优于基于传统的Smagorinsky涡粘性的动态亚格子模型。     

RESEARCH ON IMPROVEMENTS OF LRN TURBULENCE MODEL BASED ON FLOW AROUND AUTOMOBILE BODY 1)

本文将汽车绕流模块化为各典型局部流动,通过常用湍流模型对各典型局部流动进行数值模拟,结果验证了湍流模型对转捩的捕捉能力是准确模拟汽车绕流的关键. 在分析汽车绕流分离及转捩机理的基础上,优化了稳态和瞬态求解方法,改进了湍流模型对转捩的预测能力,进而提高了湍流模型在汽车流场模拟上的精度. 针对汽车绕流的稳态问题,将流线曲率因子及 响应阈值引入 LRN $k$-$\varepsilon $ (low Reynolds number $k$-$\varepsilon $) 模型,获得了一种能够更准确预 测转捩的改进低雷诺数湍流模型 (modified LRN $k$-$\varepsilon $),改善了原模型对湍流耗散率的过强依赖性及全应力发展预测不足等问题;针对汽车绕流瞬态求解,通过分析 RANS/LES 混合湍流模型的构造思想及特点,引入约束大涡模拟方法,结合本文提出的改进的 LRN $k$-$\varepsilon $ 湍流模型,提出了一种能准确捕捉转捩现象 的转捩 LRN CLES 模型. 分别将改进的模型用于某实车外流场和风振噪声仿真中,通过 Ansys Fluent 求解器计算,并将计算结果与常用湍流模型的仿真结果、HD-2 风洞试验结果和实车道路实验结果进行对比,表明改进后的湍流模型能够更准确模拟复杂实车的稳态和瞬态特性,为汽车气动特性的研究提供了可靠理论依据及有效数值解决方法.     

基于车身绕流的低雷诺数湍流模型改进研究

本文将汽车绕流模块化为各典型局部流动,通过常用湍流模型对各典型局部流动进行数值模拟,结果验证了湍流模型对转捩的捕捉能力是准确模拟汽车绕流的关键. 在分析汽车绕流分离及转捩机理的基础上,优化了稳态和瞬态求解方法,改进了湍流模型对转捩的预测能力,进而提高了湍流模型在汽车流场模拟上的精度. 针对汽车绕流的稳态问题,将流线曲率因子及 响应阈值引入 LRN $k$-$\varepsilon $ (low Reynolds number $k$-$\varepsilon $) 模型,获得了一种能够更准确预 测转捩的改进低雷诺数湍流模型 (modified LRN $k$-$\varepsilon $),改善了原模型对湍流耗散率的过强依赖性及全应力发展预测不足等问题;针对汽车绕流瞬态求解,通过分析 RANS/LES 混合湍流模型的构造思想及特点,引入约束大涡模拟方法,结合本文提出的改进的 LRN $k$-$\varepsilon $ 湍流模型,提出了一种能准确捕捉转捩现象 的转捩 LRN CLES 模型. 分别将改进的模型用于某实车外流场和风振噪声仿真中,通过 Ansys Fluent 求解器计算,并将计算结果与常用湍流模型的仿真结果、HD-2 风洞试验结果和实车道路实验结果进行对比,表明改进后的湍流模型能够更准确模拟复杂实车的稳态和瞬态特性,为汽车气动特性的研究提供了可靠理论依据及有效数值解决方法.      

Large eddy simulations on the effect of the irregular roughness shape on turbulent channel flows

Large Eddy Simulations (LES) are carried out to investigate on the mean flow in turbulent channel flows over irregular rough surfaces. Here the attention is focused to selectively investigate on the effect induced by crests or cavities of the roughness. The irregular shape is generated through the super-imposition of sinusoidal functions having random amplitude and four different wave-lengths. The irregular roughness profile is reproduced along the spanwise direction in order to obtain a 2D rough shape. The analysis of the mean velocity profiles shows that roughness crests induce higher effect in the outer-region whereas roughness cavities cause the highest effects in the inner-region with a reduced effect in the external region. The numerical simulations have been carried out at friction Reynolds number Re_τ=395. Similar results have been found for the higher order statistics: turbulence intensities or shear stresses. The analysis of the Reynolds stress anisotropy tensor confirms the existence of specific roles of cavities and crests in the turbulence modulation.     

A new dynamic subgrid-scale model using artificial neural network for compressible flow

The subgrid-scale (SGS) kinetic energy has been used to predict the SGS stress in compressible flow and it was resolved through the SGS kinetic energy transport equation in past studies. In this paper, a new SGS eddy-viscosity model is proposed using artificial neural network to obtain the SGS kinetic energy precisely, instead of using the SGS kinetic energy equation. Using the infinite series expansion and reserving the first term of the expanded term, we obtain an approximated SGS kinetic energy, which has a high correlation with the real SGS kinetic energy. Then, the coefficient of the modelled SGS kinetic energy is resolved by the artificial neural network and the modelled SGS kinetic energy is more accurate through this method compared to the SGS kinetic energy obtained from the SGS kinetic energy equation. The coefficients of the SGS stress and SGS heat flux terms are determined by the dynamic procedure. The new model is tested in the compressible turbulent channel flow. From the a posterior tests, we know that the new model can precisely predict the mean velocity, the Reynolds stress, the mean temperature and turbulence intensities, etc.     

Turbulence statistics in rotating channel flows with rough walls

We performed direct simulations of channel flow subjected to rotation about a spanwise axis, comparing cases with smooth and rough walls. The destabilizing effect of roughness counteracts the stabilizing effect of rotation on the cyclonic (stable) side. When the surface is rough the Reynolds stresses remain significant at all rotation rates considered, even those that results in a quasi-laminar state when the wall is smooth. The wake fluctuations result in significant dispersive stresses, which give an important contribution to the generation of turbulence on the stable side, mainly through added production of shear stresses. The dispersive stresses are mostly associated with the channeling of the flow between roughness elements.     

基于人工神经网络的亚格子应力建模

亚格子(SGS)应力建模在湍流大涡模拟(LES)中有着极为重要的作用. 传统亚格子应力模型存在相对误差较大、耗散过强等问题. 近年来, 计算机技术的发展使得人工神经网络(ANN)等机器学习方法逐渐成为亚格子应力建模型的新研究范式. 本文着重考虑滤波宽度及雷诺数影响, 在不可压缩槽道湍流中建立了亚格子应力的ANN模型. 该模型以滤波后的直接数值模拟(fDNS)流场物理量及滤波尺度为输入信息, 相应滤波尺度下的亚格子应力为输出量. 通过对不同滤波尺度及不同雷诺数数据的训练, ANN模型能够给出与直接数值模拟(DNS)高度吻合的亚格子应力. 此外, 模型在亚格子耗散等非ANN建模量上也有着优异的预测性能, 与基于DNS获得的对应物理量的相关系数大都在0.9以上, 较梯度模型及Smagorinsky模型有明显提升. 在后验测试中, ANN模型对流向平均速度剖面的预测同样优于梯度模型、Smagorinsky模型及隐式大涡模拟(ILES)等传统LES模型. 在脉动速度均方根预测方面, 除了某些法向位置外ANN模型的性能整体上相对其他3个模型有所提升. 然而, 随着网格尺度的增大ANN模型预测的结果与fDNS结果的偏差逐渐增大. 总之, ANN方法在发展高精度亚格子应力模型上具有很大的潜力.      

Turbulent flow over a rough backward-facing step

This work characterizes the impacts of the realistic roughness due to deposition of foreign materials on the turbulent flows at surface transition from elevated rough-wall to smooth-wall. High resolution PIV measurements were performed in the streamwise-wall-normal (x–y) planes at two different spanwise positions in both smooth and rough backward-facing step flows. The experiment conditions were set at a Reynolds number of 3450 based on the free stream velocity U_∞ and the mean step height h, expansion ratio of 1.01, and the ratio of incoming boundary layer thickness to the step height, δ/h, of 8. The mean flow structures are observed to be modified by the roughness and they illustrate three-dimensional features in rough backward-facing step flows. The mean reattachment length X_r is significantly reduced by the roughness at one PIV measurement position while is slightly increased by the different roughness topography at the other measurement position. The mean velocity profiles at the reattachment point indicate that the studied roughness weakens the perturbation of the step to the incoming turbulent flow. Comparisons of Reynolds normal and shear stresses, productions of normal stresses, quadrant analysis of the instantaneous shear-stress contributing events, and mean spanwise vorticity reveal that the turbulence in the separated shear layer is reduced by the studied roughness. The results also indicate an earlier separation of the turbulent boundary layer over the current rough step, probably due to the adverse pressure gradient produced by the roughness topography even before the step.     

Swirling-strength based large eddy simulation of turbulent flow around single square cylinder at low Reynolds numbers

In view of the fact that large scale vortices play the substantial role of momentum transport in turbulent flows, large eddy simulation(LES) is considered as a better simulation model. However, the sub-grid scale(SGS) models reported so far have not ascertained under what flow conditions the LES can lapse into the direct numerical simulation. To overcome this discrepancy, this paper develops a swirling strength based the SGS model to properly model the turbulence intermittency, with the primary characteristics that when the local swirling strength is zero, the local sub-grid viscosity will be vanished. In this paper, the model is used to investigate the flow characteristics of zero-incident incompressible turbulent flows around a single square cylinder(SC)at a low Reynolds number range Re ∈ [103, 104]. The flow characteristics investigated include the Reynolds number dependence of lift and drag coefficients, the distributions of time-spanwise averaged variables such as the sub-grid viscosity and the logarithm of Kolmogorov micro-scale to the base of 10 at Re = 2 500 and 104, the contours of spanwise and streamwise vorticity components at t = 170. It is revealed that the peak value of sub-grid viscosity ratio and its root mean square(RMS) values grow with the Reynolds number. The dissipation rate of turbulent kinetic energy is larger near the SC solid walls.The instantaneous factor of swirling strength intermittency(FSI) exhibits some laminated structure involved with vortex shedding.     

Görtler vortices promoted by wall roughness

Numerical experiments are conducted to investigate spatially developing Görtler vortices and the way in which wall roughness promotes their formation and growth. Several different types of walls are examined and their relative merits as vortex promoters assessed. The only disturbances of the flow are due to the rough wall; hence, at each downstream station the local field feels (1) the upstream flow distribution (produced by the upstream wall conditions) and (2) the local forcing at the wall. Rapid vortex formation and growth, like in the case of ribleted walls, can be qualitatively explained by the positive combination of these two effects; when the two influences on the local flow field compete, e.g. for randomly distributed wall roughness, the equations with the boundary conditions filter the disturbances over some streamwise length, function of the roughness amplitude, to create coherent patches of vorticity out of the random noise. These patches can then be amplified by the instability mechanism. If a thin rough strip is aligned along the span of an otherwise smooth wall to trip the boundary layer, the filtering region is shorter and growth of the vortices starts earlier. Also for the case of an isolated three-dimensional hump a rapid disturbance amplification is produced, but in this case the vortices remain confined and a very slow spanwise spreading of the perturbation occurs. In all naturally developing cases, where no specific wavelengths are explicity favored, the average spanwise wavelengths computed are very close to those of largest growth from the linear stability theory.     

SPH modelling of depth‐limited turbulent open channel flows over rough boundaries

A numerical model based on the smoothed particle hydrodynamics method is developed to simulate depth‐limited turbulent open channel flows over hydraulically rough beds. The 2D Lagrangian form of the Navier–Stokes equations is solved, in which a drag‐based formulation is used based on an effective roughness zone near the bed to account for the roughness effect of bed spheres and an improved sub‐particle‐scale model is applied to account for the effect of turbulence. The sub‐particle‐scale model is constructed based on the mixing‐length assumption rather than the standard Smagorinsky approach to compute the eddy‐viscosity. A robust in/out‐flow boundary technique is also proposed to achieve stable uniform flow conditions at the inlet and outlet boundaries where the flow characteristics are unknown. The model is applied to simulate uniform open channel flows over a rough bed composed of regular spheres and validated by experimental velocity data. To investigate the influence of the bed roughness on different flow conditions, data from 12 experimental tests with different bed slopes and uniform water depths are simulated, and a good agreement has been observed between the model and experimental results of the streamwise velocity and turbulent shear stress. This shows that both the roughness effect and flow turbulence should be addressed in order to simulate the correct mechanisms of turbulent flow over a rough bed boundary and that the presented smoothed particle hydrodynamics model accomplishes this successfully. © 2016 The Authors International Journal for Numerical Methods in Fluids Published by John Wiley & Sons Ltd     

岩石结构面直剪力学特征的颗粒流宏细观分析

目前，由于颗粒尺寸和位置随机分布而产生的粗糙度，导致在模拟过程中出现应力集中现象，从而无法利用PFC2D准确模拟岩石结构面的剪切力学行为。为了避免以上弊端，采用结构面两侧颗粒组接触判别法分别对倾角为20°和30°的三角形单锯齿结构面进行不同法向应力下的直剪模拟。模拟结果与物理试验结果基本吻合，表明此模拟方法可以真实模拟岩石结构面剪切力学行为。为了进一步验证此模拟方法在天然结构面模拟试验中的可行性，利用PFC2D对Barton标准十条节理轮廓线模型进行不同法向应力下的直剪试验，模拟结果与理论值吻合，表明模拟方法具有可行性，为剪切破坏过程和机理分析提供可靠的理论依据，为研究粗糙结构面强度特征提供了一种思路。     

Temporal and spatial transients in turbulent boundary layer flow over an oscillating wall

Direct numerical simulations have been performed to study the effect of an oscillating segment of the wall on a turbulent boundary layer flow. Two different oscillation amplitudes with equal oscillation period have been used, which allows a direct comparison between a relatively weak and strong forcing of the flow. The weaker forcing results in 18% drag reduction while the stronger forcing, with twice the amplitude, yields 29% drag reduction. The downstream development of the drag reduction is compared with earlier simulations and experiments. In addition, a simulation with identical oscillation parameters as in previous numerical and experimental investigations allows for an estimation of the effect of the Reynolds number on the drag reduction.Reductions in the Reynolds stresses and the important role that the edge of the Stokes layer has is explained.An estimation of the idealized power consumption shows that a positive energy budget is only possible for the weaker wall velocity case.Spatial and temporal transients are investigated and a transformation between spatial and temporal coordinates via a convection velocity is shown to facilitate a comparison between the two transients in a consistent manner. The streamwise shear exhibits a similar monotonic behavior in the spatial and temporal transients, while the non-monotinic temporal transient of the longitudinal Reynolds stress has no counterpart in the spatial development. Furthermore, the evolution in time of the spanwise Reynolds stress is very similar to previously reported channel flow data.The instantaneous spanwise velocity profile (only averaged in the homogeneous spanwise direction) will for the first time be presented from a boundary layer over an oscillating wall, and comparisons with the analytical solution to the laminar Navier–Stokes equations show very good agreement.     

Receptivity mechanisms for Görtler vortex modes

The receptivity problem for Görtler vortices induced by wall roughness or freestream disturbances is reviewed. To date, receptivity studies for this problem have been exclusively linear in character and here we show how the roughness and freestream disturbance mechanisms can each play dominant or inconsequential roles as possible routes to transition. The importance of each process is dependent on the exact situation at hand. For example, distributed wall-roughness elements tend to be more important in the generation of O(1) wave-number vortices than are isolated roughness patches whilst variations in the freestream velocity can easily provoke high wave-number disturbances on which roughness distributions typically have negligible effect.It has only been in recent times that the influence a spanwise component of the underlying basic boundary-layer flow may have on the Görtler mechanism has come to be appreciated. In some new computations we show that the imposition of such a spanwise component can lead to an increase in the coupling coefficient (a measure of the efficiency of a generating process) for modes provoked by wall roughness. However, such crossflow tends to reduce the global amplification rate of the most unstable mode so has the overall effect of restricting vortex growth downstream of any roughness element. We also demonstrate how the nonparallelism of Görtler vortices implies that conclusions concerning vortex receptivity properties can only be drawn after taking full account of upstream conditions and the precise form of the generating mechanism but it appears that for a large class of flows distributed wall forcing is more important in the provocation of modes than are either isolated roughness or freestream disturbances.This work was completed whilst APB was on study leave at the School of Mathematics, University of New South Wales, Sydney. He is indebted to the Royal Society of London and the Australian Research Council without whose grants (the latter to Dr. Peter Blennerhassett) his visit would not have been possible. In addition he is grateful to the staff and students of New College, UNSW for their provision of a Visiting Fellowship and to the School of Mathematics for their hospitality.     

Numerical investigation of turbulent channel flow controlled by spatially oscillating spanwise Lorentz force

A formulation of the skin-friction drag related to the Reynolds shear stress in a turbulent channel flow is derived. A direct numerical simulation (DNS) of the turbulent control is performed by imposing the spatially oscillating spanwise Lorentz force. Under the action of the Lorentz force with several proper control parameters, only the periodically well-organized streamwise vortices are finally observed in the near-wall region. The Reynolds shear stress decreases dramatically, especially in the near-wall area, resulting in a drag reduction.     

The effects of surface roughness geometry of flow undergoing Poiseuille flow by molecular dynamics simulation

Numerical simulation of Poiseuille flow of liquid Argon in a rough nano-channel using the non-equilibrium molecular dynamics simulation is performed. Density and velocity profiles across the channel are investigated in which roughness is implemented only on the lower wall. The Lennard–Jones potential is used to model the interactions between all particles. The effects of surface roughness geometry, gap between roughness elements (or roughness periodicity), surface roughness height and surface attraction energy on the behavior of the flow undergoing Poiseuille flow are presented. Results show that surface shape and roughness height have a decisive role on the flow behaviors. In fact, by increasing the roughness ratio (height to base ratio), the slip velocity and the maximum velocity in the channel cross section are reduced, and the density fluctuations near the wall increases. Results also show that the maximum density near the wall for a rough surface is less than a smooth wall. Moreover, the simulation results show that the effect of triangle roughness surface on the flow behavior is more than the cylindrical ones.