Search

Our ability to reliably and efficiently predict complex high-Reynolds-number ( R e ) turbulent flows is essential for dealing with a large variety of problems of practical relevance. However, experiments as well as computational methods such as direct numerical simulation (DNS) and large eddy simulation (LES) face serious questions regarding their applicability to high R e turbulent flows. The most promising option to create reliable guidelines for experimental and computational studies is the use of analytical conclusions. An essential criterion for the reliability of such analytical conclusions is the inclusion of a physically plausible explanation of the asymptotic turbulence regime at infinite R e in consistency with observed physical requirements. Corresponding analytical results are reported here for three canonical wall-bounded turbulent flows: channel flow, pipe flow, and the zero-pressure gradient turbulent boundary layer. The asymptotic structure of the mean velocity and characteristic turbulence velocity, length, and time scales is analytically determined. In outer scaling, a stable asymptotic mean velocity distribution is found corresponding to a linear probability density function of mean velocities along the wall-normal direction, which is modified through wake effects. Turbulence tends to decay in this regime. In inner scaling, the mean velocity is governed by a universal log-law. Turbulence does survive in an infinitesimally thin layer very close to the wall. [ABSTRACT FROM AUTHOR]

The vertical distribution of longitudinal velocity and the turbulence structure of open channel flows with natural‐like reedy emergent vegetation were investigated in this study. The natural‐like reedy vegetation consisted of a basal stem and flexible blades with shape and reconfiguration behaviour comparable with those found in actual riparian areas. The 3D velocity field was measured using an acoustic Doppler velocimeter. The vertically non‐uniform distribution of the frontal area could dramatically affect the exchange of mass and momentum between emergent vegetation and their surroundings. The average time velocity profile varied inversely with the frontal area. The high velocity gradient appearing at the middle section of the zone with frontal area varied considerably. The structures of vortices and vertical transport were analysed using spectral analysis and conditional sampling to identify the prominent vortical structures and their contribution to momentum transport. Spectral analysis suggested that the stem‐scale vortices set by stem spacing where the vertical frontal area had no clear change were dominant, and shear‐layer vortices where vertical frontal area dramatically changed cannot be disregarded. Quadrant analysis showed that inward and outward interactions were the dominant contributors to Reynolds stress, excluding the zones approaching the channel boundary and the vertical frontal area that dramatically changed. A four‐layer model was developed to predict the vertical distribution of longitudinal velocity with a newly established profile of local drag coefficient. The predicting results agreed well with the experimental data. [ABSTRACT FROM AUTHOR]

Direct numerical simulations of turbulent boundary layers with roughness elements on a wall were performed to investigate the spatial characteristics of rough-wall turbulence and establish a corresponding prediction method. When the roughness height was larger than the buffer layer, the rough-wall turbulence exhibited different spatial characteristics of the turbulence structures from those pertaining to a smooth wall. A novel spatial scaling method was established to examine the universal spatial characteristics of turbulence structures in the presence and absence of wall roughness. Specifically, the viscous length was determined by modifying the definition of the friction velocity in the region in which the roughness influenced the flow. The rough-wall turbulence could be accurately predicted by performing large eddy simulations using the subgrid scale model with the filter width, which was modified using the proposed spatial scaling method. The proposed model can be used to design more efficient fluid machinery in engineering applications. [ABSTRACT FROM AUTHOR]

Ice cover formed on a river surface is a common natural phenomenon during the winter season in cold high latitude northern regions. For the ice‐covered river with compound cross‐section, the interaction of the turbulence caused by the ice cover and the channel bed bottom affects the transverse mass and momentum exchange between the main channel and floodplains. In this study, laboratory experiments are performed to investigate the turbulent flow of a compound channel with shore ice covered on the floodplains. Results show that the shore ice resistance restricts the development of the water flow and creates a relatively strong shear layer near the edge of the ice‐covered floodplain. The mean streamwise velocity in the main channel and on the ice‐covered floodplains shows an opposite variation pattern along with the longitudinal distance and finally reaches the longitudinal uniformity. The mixing layer bounded by the velocity inflection point consists of two layers that evolve downstream to their respective fully developed states. The velocity inflection point and strong transverse shear near the interface in the fully developed profile generate the Kelvin‐Helmholtz instability and horizontal coherent vortices. These coherent vortices induce quasi‐periodic velocity oscillations, while the dominant frequency of the vortical energy is determined through the power spectral analysis. Subsequently, quadrant analysis is used in ascertaining the mechanism for the lateral momentum exchange, which exhibits the governing contributions of sweeps and ejections within the vortex center. Finally, an eddy viscosity model is presented to investigate the transverse momentum exchange. The presented model is well validated through comparison with measurements, whereas the constants α and β appeared in the model need to be further investigated. Key Points: The longitudinal evolution of the streamwise velocity in a compound channel with shore ice is presented using laboratory experimentsThe shore ice induced resistance restricts the flow development and affects the width and momentum thickness of the mixing layerWhether the floodplain is covered by the shore ice, the transverse momentum exchange mechanism is similar [ABSTRACT FROM AUTHOR]

We performed direct numerical simulations of a turbulent channel flow in viscoelastic fluids to assess various spatial scaling methods and to find a scaling method that can visualise the spatial features regardless of fluid properties. The Giesekus and Oldroyd-B models were used as constitutive equations to model the viscoelastic fluids alongside a Newtonian fluid. The initial scaling method that can consider local viscosity variations demonstrated an increase in the size of the velocity streaks and a decrease in the number of quasi-streamwise vortices, with no change in the observed spatial features of the turbulence structures. After scaling using the ratio of the wall friction coefficients, we observed typical spatial features of the principal turbulence structures that coincided with those of the Newtonian fluid. The results indicate that the characteristics peculiar to the turbulent flow of viscoelastic fluids can be predicted in a manner similar to that for Newtonian fluids. [ABSTRACT FROM AUTHOR]

Turbulence structures in flow over three types of wall roughness: sand-grain, cube roughness and a realistic, multi-scale turbine-blade roughness, are compared to structures observed in flow over a smooth wall in open channel flow at R e τ = 1000 , using direct numerical simulations. Two-point velocity correlations, length scales, inclination angles, and velocity spectra are analysed, and the applicability of Townsend's outer layer similarity hypothesis [Townsend. The structure of turbulent shear flow. Cambridge: Cambridge University Press; 1976] to these parameters was examined. Results from linear stochastic estimation suggest that, near the wall, the quasi-streamwise vortices observed in smooth-wall flow are present in the large-scale recessed regions of multi-scale roughness, whereas they are replaced by a pair of 'head-up, head-down' horseshoe structures in sandgrain and cube roughness, similar to those observed by Talapatra and Katz [Coherent structures in the inner part of a rough-wall channel flow resolved using holographic PIV. J Fluid Mech. 2012;711:161–170]. The configuration of conditional eddies near the wall suggests that the kinematical process of vortices differ for each kind of rough surface. Eddies over multiscale roughness are conjectured to obey a growth mechanism similar to those over smooth walls, while around the cube roughness the head-down horse-shoe vortices of Talapatra and Katz [Coherent structures in the inner part of a rough-wall channel flow resolved using holographic PIV. J Fluid Mech. 2012;711:161–170] may undergo solid-body rotation on top of a cube roughness on account of the strong shear layer, shortening the longitudinal extent of near-wall structure and promoting turbulence production during this process. These results illustrate the sensitivity of turbulence structure to the roughness texture, particularly within the roughness sublayer. [ABSTRACT FROM AUTHOR]

Turbulence structures in flow over three types of wall roughness: sand-grain, cube roughness and a realistic, multi-scale turbine-blade roughness, are compared to structures observed in flow over a smooth wall in open channel flow at R e τ = 1000 , using direct numerical simulations. Two-point velocity correlations, length scales, inclination angles, and velocity spectra are analysed, and the applicability of Townsend's outer layer similarity hypothesis [Townsend. The structure of turbulent shear flow. Cambridge: Cambridge University Press; 1976] to these parameters was examined. Results from linear stochastic estimation suggest that, near the wall, the quasi-streamwise vortices observed in smooth-wall flow are present in the large-scale recessed regions of multi-scale roughness, whereas they are replaced by a pair of 'head-up, head-down' horseshoe structures in sandgrain and cube roughness, similar to those observed by Talapatra and Katz [Coherent structures in the inner part of a rough-wall channel flow resolved using holographic PIV. J Fluid Mech. 2012;711:161–170]. The configuration of conditional eddies near the wall suggests that the kinematical process of vortices differ for each kind of rough surface. Eddies over multiscale roughness are conjectured to obey a growth mechanism similar to those over smooth walls, while around the cube roughness the head-down horse-shoe vortices of Talapatra and Katz [Coherent structures in the inner part of a rough-wall channel flow resolved using holographic PIV. J Fluid Mech. 2012;711:161–170] may undergo solid-body rotation on top of a cube roughness on account of the strong shear layer, shortening the longitudinal extent of near-wall structure and promoting turbulence production during this process. These results illustrate the sensitivity of turbulence structure to the roughness texture, particularly within the roughness sublayer. [ABSTRACT FROM AUTHOR]

A new particle subgrid scale model is proposed to consider the effect of gas flow on particle motions. Multiphase gas-particle turbulent flow is modeled by a second-order moment two-phase turbulence model involving a four-way coupling strategy to describe the interactions among gas-particle, particle-gas and particle-particle collisions. A large eddy simulation algorithm is developed to solve the hydrodynamic parameters of confined swirling and non-swirling particle-laden flow in coaxial chamber. Results show that predictions are well agreed with experimental data. Vortex structures and vortices distributions of gas and particle flow are quietly different. Coherent structures of swirling particle flows are not observed, and length of recirculation region is almost one quarter of non-swirling flow. Compared to developed flow region of non-swirling flow, standard deviation values of granular temperature at near entrance decreased by 5.5 times. Dominant frequencies of particle number density of non-swirling and swirling flows are 18 Hz and 10 Hz. Particle dispersions exhibit anisotropic characteristics, and their distributions are not in accordance with the normal distribution. The 2D turbulence model needs to be further improved due to the failure of describing vortex stretch in this job. • A new particle subgrid scale model is proposed to consider the effect of gas flow on particles. • Coherent structure and vortices of confined particle-laden flow are revealed by large eddy simulation. • Recirculation length of non-swirling flow is almost 4.0 times larger than that of swirling flow. • Standard deviations of granular temperature of non-swirling flow have decreased by 6.5 times. • Dominant frequencies of particle number density are 18 Hz and 10 Hz in non-swirling and swirling flow. [ABSTRACT FROM AUTHOR]

Clustering of coherent fine-scale eddies in a turbulent mixing layer has been analyzed by using direct numerical simulation (DNS) data at Re λ  ≃ 250. The coherent fine-scale eddies are defined based on the second invariant of the velocity gradient tensor and the vorticity vector. The clustering is evaluated by the number density of coherent fine-scale eddies, and the large-scale structures are extracted by low-pass filtered velocity fields. Conditional averaging shows that the large-scale enstrophy increases with the number density, whereas the large-scale strain rate stays around the average in the high number density region. The alignments of the vorticity vector and the eigenvectors of the large-scale strain rate tensor are conditioned by the number density or the strain rate magnitude. The eigenvectors and the vorticity vector indicate strong preferential alignments under the intense large-scale strain rate condition. On the other hand, those alignments become weak in the high number density regions. The inter-scale energy transfer between grid and subgrid scales is significantly correlated with the magnitude of the large-scale strain rate while there is no apparent correlation with the number density. [ABSTRACT FROM AUTHOR]

The results are presented from an experimental study to investigate three-dimensional turbulence structure profiles, including turbulence intensity and Reynolds stress, of different non-uniform open channel flows over smooth bed in subcritical flow regime. In the analysis, the uniform flow profiles have been used to compare with those of the non-uniform flows to investigate their time-averaged spatial flow turbulence structure characteristics. The measured non-uniform velocity profiles are used to verify the von Karman constant κ and to estimate sets of log-law integration constant Br and wake parameter П, where their findings are also compared with values from previous studies. From κ, Br and П findings, it has been found that the log-wake law can sufficiently represent the non-uniform flow in its non-modified form, and all κ, Br and П follow universal rules for different bed roughness conditions. The non-uniform flow experiments also show that both the turbulence intensity and Reynolds stress are governed well by exponential pressure gradient parameter β equations. Their exponential constants are described by quadratic functions in the investigated β range. Through this experimental study, it has been observed that the decelerating flow shows higher empirical constants, in both the turbulence intensity and Reynolds stress compared to the accelerating flow. The decelerating flow also has stronger dominance to determine the flow non-uniformity, because it presents higher Reynolds stress profile than uniform flow, whereas the accelerating flow does not. [ABSTRACT FROM AUTHOR]

Turbulent flow in the axial direction along a cylinder representing a monofilament yarn was reproduced for a relatively wide range of radius Reynolds numbers using direct numerical simulations. In the simulation of the thinnest cylinder, the friction coefficient agreed with a previously published formula. A pair of high- and low-speed streaks was detected around even the thinnest cylinder, whereby it was confirmed that turbulence was sustained according to the analysed turbulence statistics. Even when only a single pair of streaks around the cylinder was detected, the characteristics of the turbulence structures, such as the mean streak spacing based on the viscous length scale, were the same as those in flows over a flat plate. It was found that the friction coefficient changes in a way that maintains the structural characteristics of the flow, consistent with the view that universal characteristics of turbulence structures exist. [ABSTRACT FROM PUBLISHER]

• Lateral distribution of velocity may have two dramatically increasing segments. • Lateral distribution of the Reynolds stress shows a bimodal characteristic. • Large-scale quasi-2-D coherent structures exist in the shear layer and mixing layer. • Ejections and sweeps events are dominant in the shear layer and mixing layer. Aquatic plants can alter flow structure, affect sediment transport and contaminants diffusion, and is noteworthy for channel restoration and navigation. In the present work, experiments are conducted to investigate the velocity and turbulence structures in a compound channel with varying coverage of riparian vegetation. Lateral profile of velocity may change from a traditional single sharply increasing segment to two sharply increasing segments, the shear layer (SL) and the mixing layer (ML), which are influenced by bed morphology and vegetation. Lateral profile of Reynolds stress is bimodal, with a crest value that decreases with increasing ratio of the width of nonvegetated floodplain to the bankfull height (B NV / h). Quasi-two-dimensional (Quasi-2-D) coherent structures can be found in both SL and ML when B NV / h is small according to power spectral density and temporal autocorrelation function, which have a significant influence on Reynolds stress. Brief and intense ejections and sweeps events dominate the contribution to Reynolds stress, and according to quadrant analysis, an ejections event of the SL may occur simultaneously with each event of the ML. [ABSTRACT FROM AUTHOR]

The gas flow and flame structures surrounding an evaporating and combusting ethanol droplet under forced convection with different gas relative velocities, ambient temperatures and droplet sizes are studied by direct numerical simulation. The results show that the flames develop in the coherent-structure region for both the regimes of fully enveloped flames and wake flames. The evaporating and combusting droplet reduces the gas turbulence but not like the cold solid particle enhances the gas turbulence by its wake effect. [ABSTRACT FROM AUTHOR]

Direct numerical simulations (DNS) of turbulent channel flow of viscoelastic fluid were performed with the Oldroyd-B model as a constitutive equation to identify the temporal characteristics of turbulence structures. First, using the results of the DNS in a minimal flow unit (MFU), temporal variations in the energy input and dissipation rates in fully developed wall turbulence of viscoelastic fluid were observed and compared with those of Newtonian fluid. Then, a modified time scale reflecting the local turbulence state was proposed by considering the temporal variation in the viscous stretching stress intensity. It was also found that the idea of the modified time scale could be extended and applied locally to viscoelastic fluid turbulence in a more significant computational domain than MFU to investigate the temporal characteristics of the turbulence structures. Finally, using the local modified time scale, a prediction method employing large eddy simulation for Newtonian fluid turbulence was extended for viscoelastic fluid turbulence. We proved that the drag reduction of viscoelastic fluid turbulence could be accurately predicted with a drag reduction ratio of approximately 40% by using the local modified time scale and the model in the constitutive equation. • The characteristics of viscoelastic fluid turbulence were observed from DNS results. • Active and hibernating turbulence states were compared with those in Newtonian fluid. • A local modified time scale which reflected the local turbulence state was proposed. • LES was extended using the modified time scale for viscoelastic fluid turbulence. • The proposed LES could predict the drag reduction of viscoelastic fluid turbulence. [ABSTRACT FROM AUTHOR]

Highlights: [•] Adverse pressure gradient diminished the size of the hairpin packet over the smooth wall. [•] Adverse pressure gradient produced more intense ejection, sweep and interaction motions over smooth and rough walls. [•] The turbulence structure over the k-type rough wall is larger compared to the smooth wall. [ABSTRACT FROM AUTHOR]

An analysis of the turbulence structure in a perturbed boundary layer and in low-wind regimes is presented. The study is based on 15 months of continuous wind and turbulence measurements gathered, within the framework of the Urban Turbulence Project, at three levels (5, 9 and 25 m) on a mast located in the outskirts of the city of Turin (Italy). The aim of the work is to investigate low-frequency processes in a perturbed boundary-layer. In fact, the urban canopy and the heat island, together with frequent low-wind conditions, interact with and modify the turbulence structure. In order to investigate this modification, the velocity Eulerian autocorrelation functions together with both the Eulerian and Lagrangian time-scales are shown and compared with the classical theory. The comparisons show that in low-wind cases the velocity autocorrelation functions are not simply exponential but present an oscillating behaviour. A method of normalization is proposed together with an analysis on the applicability of this function. The estimated Lagrangian time-scales are compared with two widely used parametrizations. It is found that the presence of the urban fabric influences the turbulence time-scales and suggests the development of new parametrizations. Finally, higher-order statistics are evaluated and the relationship between higher-order and lower-order moments are analysed, pointing out the effects due to the urban environment. Copyright © 2012 Royal Meteorological Society [ABSTRACT FROM AUTHOR]

ABSTRACT Northern rivers experience freeze-up over the winter, creating asymmetric under-ice flows. Field and laboratory measurements of under-ice flows typically exhibit flow asymmetry and its characteristics depend on the presence of roughness elements on the ice cover underside. In this study, flume experiments of flows under a simulated ice cover are presented. Open water conditions and simulated rough ice-covered flows are discussed. Mean flow and turbulent flow statistics were obtained from an Acoustic Doppler Velocimeter (ADV) above a gravel-bed surface. A central region of faster flow develops in the middle portion of the flow with the addition of a rough cover. The turbulent flow characteristics are unambiguously different when simulated ice covered conditions are used. Two distinct boundary layers (near the bed and in the vicinity of the ice cover, near the water surface) are clearly identified, each being characterized by high turbulent intensity levels. Detailed profile measurements of Reynolds stresses and turbulent kinetic energy indicate that the turbulence structure is strongly influenced by the presence of an ice cover and its roughness characteristics. In general, for y/ d > 0·4 (where y is height above bed and d is local flow depth), the addition of cover and its roughening tends to generate higher turbulent kinetic energy values in comparison to open water flows and Reynolds stresses become increasingly negative due to increased turbulence levels in the vicinity of the rough ice cover. The high negative Reynolds stresses not only indicate high turbulence levels created by the rough ice cover but also coherent flow structures where quadrants one and three dominate. Copyright © 2012 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

Abstract: In this paper we will present some recent results concerning the simplest decomposition of a turbulent field. The large scale filtering operator is simply given by the two-point sum in space and the associated fluctuation is given by the two-point difference. In the paper we will present the general properties of this simple decomposition and their particular relations with the subgrid stresses and the generalized central moments associated to a generic filtering operator. [Copyright &y& Elsevier]

Abstract: In order to understand the effect of the wall permeability on the turbulent vortex structure near porous walls, based on PIV experimental data, a probability density analysis of fluctuating velocities, statistical quadrant and quadrant-hole analyses of the Reynolds shear stress are performed. The investigated flow fields are turbulent channel flows whose bottom walls are made of porous media. The porous media used are three kinds of foamed ceramics which have almost the same porosity (∼0.8) but different permeability. From the discussions on those analyses, a conceptual scenario of the development of the vortex structure near a permeable wall is proposed for a moderate permeability Reynolds number case. It explains the reason why the near-wall long streaky structure tends to vanish near a porous wall with increasing wall permeability. [Copyright &y& Elsevier]

When wind-induced water waves appear over the free-surface flows such as natural rivers and artificial channels, large amounts of oxygen gas and heat are transported toward the river bed through the interface between water and wind layers. In contrast, a bed region is a kind of turbulent boundary layer, in which turbulence generation and its transport is promoted by the production of bed shear stress. In particular, coherent hairpin vortices, together with strong ejection events toward the outer part of the layer, promote mass and momentum exchanges between the inner and outer layers. It is inferred that such a near-bed turbulence may be influenced significantly by these air-water interfacial fluctuations accompanied with free-surface velocity shear and wind-induced water waves. However, these wind effects on the wall-turbulence structure are less understood. To address these exciting and challenging topics, we conducted particle imagery velocimetry (PIV) measurements in open-channel flows combined with air flows, and furthermore the present measured data allows us to investigate the effects of air-water interactions on turbulence structure through the whole depth region. [ABSTRACT FROM AUTHOR]

Abstract: Sharply sheared interfaces determine the structure of turbulent motions both on large and small scales, as recent experiments and simulations have demonstrated. Shear layers form along contours defined by very large gradients in turbulence intensity and/or large gradients in the large scale velocity field. Inhomogeneous eddy motions adjacent to these layers stretch and distort the larger scale vorticity so as to counter the tendency of the interface to thicken by small scale eddy diffusion. These concepts are applied to the structure of high Reynolds number turbulence by focusing on the dynamics of the flow within and near the shear layers that are observed to exist between large eddies. The numerical simulations of show that on the edges of the layers and in their interiors thin viscous layers form, with thickness approximately equal to the Taylor microscale. Intermittent small-scale vortices within these viscous layers are stretched and amplified, as their thickness reduces to a limiting value determined by viscous diffusion, which is equal to the Kolmogorov micro length scale. But the root-mean -square values of vorticity and velocity, and also their flatness factors, are given by statistical averaging across these layers, taking into account the degree to which the thin layers are sufficiently convoluted to be ‘space-filling’. Through the blocking of the external scale eddies outside the layers as they impinge onto the interfaces, upscale and downscale energy transfer processes which are of comparable magnitude lead to distorted inertial-range motions with a wide range of length scales and characteristic self-similar power-law spectra. Here the skewness of the velocity derivatives becomes negative. The upscale motions are essential for maintaining the large structures and the thin shear layers between them, as two-dimensional experiments demonstrate. A significant implication of this ‘interface dynamics’ mechanism is that small scale turbulence is produced as much by larger eddy structures, as by interactions between progressively smaller, and more numerous, self-similar eddies – the ‘cascade’ mechanism proposed by L.F. Richardson. The latter mechanism, which is dynamically equivalent to the Kolmogorov–Obukhov statistical-physics models for homogeneously distributed fluctuations, implies that turbulence structure evolves more slowly than is observed. By contrast the model presented here for the statistics and dynamics of high Reynolds number turbulence is based on a quasi-deterministic zonal analysis within and adjacent to intermittently distributed thin shear layers, which occupy a finite fraction of the flow and cause the eddy motions to have different ranges of length scales in separate parts of the local flow fields. The structure does not change qualitatively as the Reynolds number varies, but is significantly changed if the turbulence is forced at small as well as large scales. [Copyright &y& Elsevier]

Three dimensional large eddy simulation (LES) is performed in the investigation of stably stratified turbulence with a sharp thermal interface. Main results are focused on the turbulent characteristic scale, statistical properties, transport properties, and temporal and spatial evolution of the scalar field. Results show that the buoyancy scale increases first, and then goes to a certain constant value. The stronger the mean shear, the larger the buoyancy scale. The overturning scale increases with the flow, and the mean shear improves the overturning scale. The flatness factor of temperature departs from the Gaussian distribution in a fairly large region, and its statistical properties are clearly different from those of the velocity fluctuations in strong stratified cases. Turbulent mixing starts from small scale motions, and then extends to large scale motions. [ABSTRACT FROM AUTHOR]

Abstract: The advanced ‘comprehensive’ laser–Doppler velocimeter is used to acquire spatially and temporally resolved turbulence structural measurements in high Reynolds number two- and three-dimensional turbulent boundary layers. The new instrument directly measures three-dimensional particle trajectories at high repetitions. These trajectories are analyzed in post-processing to obtain fluctuating velocity gradient tensor fields, which lead to direct measurements of turbulent viscous dissipation rates. Such data acquired in two- and three-dimensional boundary layers with an approach flow momentum thickness Reynolds number, Re θ =7500 are presented. Results indicate that anisotropy of the dissipation rate of Reynolds stresses persists to similar heights in viscous wall units as obtained with direct numerical simulations at lower Reynolds numbers. Measurements in a three-dimensional turbulent boundary layer in the vicinity of a wing/body junction also indicate that a reduction in the value of the velocity/pressure-gradient correlations for the Reynolds normal stresses reduces the turbulent energy redistribution and contributes to reduced shear stress magnitudes, as observed previously through DNS (Moin, P., Shih, T.-H., Driver, D., Mansour, N.N., 1990. Direct numerical simulation of a three-dimensional turbulent boundary layer, Phys. Fluids A 2 (10), 1846–1853). [Copyright &y& Elsevier]

Abstract: The paper reports an investigation of the influence of geometric profile of a long slot nozzle on the statistical properties of a plane jet discharging into a large space. The nozzle-exit profile was varied by changing orifice-plates with different exit radii (r) over the range of 0< r/h <3.60, where h is the slot-height. The present measurements were made at a slot-height based Reynolds number (Re h ) of 1.80×104 and a slot aspect ratio (span/height) of 72. The results obtained show that both the initial flow and the downstream flow are dependent upon the ratio r/h. A “top-hat” mean exit velocity profile is closely approximated when r/h approaches 3.60. The decay and spread rates of the jet’s mean velocity decrease asymptotically as r/h is increased, with the differences becoming small as r/h approaches 3.60. A decrease in r/h results in a higher formation rate of the primary vortices in the near-field. The far-field values of the centerline turbulence intensity are higher for smaller r/h, and display asymptotic-like convergence as r/h approaches 3.60. Overall, the effect of r/h on the mean and turbulence fields decreases as r/h increases. [Copyright &y& Elsevier]

Abstract: The applicability of the large eddy simulation (LES) technique to wake flows past a bluff body should be clarified in order to improve numerical accuracy for the estimation of aerodynamic forces and pressures acting on a bluff body. Here, we conduct both LES and particle image velocimetry (PIV) measurement for turbulent flows past rectangular cylinders with a depth/breadth ratio ranging from D/B=2.00 to 3.00 at Re=2000 and 20,000, which sensitively change turbulence structures due to flow separation and reattachment. For LES, we use the dynamic Smagorinsky-type SGS (sub-grid scale) model. First, at Re=2000 where pure turbulent viscosity without numerical viscosity is realized, the accuracy of the SGS modeling for the prediction of not only aerodynamic characteristics but also turbulence statistics in the wake of rectangular cylinders is examined in comparison with PIV data. Furthermore, at Re=20,000 where the numerical dissipation must be incorporated, we clarify the unfavorable effects of the numerical dissipation on the turbulence structures in the wake. [Copyright &y& Elsevier]

Abstract: Turbulence measurements were made in a Taylor–Couette apparatus as a basis for future flame propagation studies. Results of the present study extend that of earlier work by more complete characterization of the featureless turbulence regime generated by the Taylor–Couette apparatus. Laser Doppler Velocimetry was used to measure Reynolds stresses, integral and micro time scales and power spectra over a wide range of turbulence intensities typically encountered by turbulent pre-mixed hydrocarbon–air flames. Measurements of radial velocity intensities are consistent with earlier axial and circumferential velocity measurements that indicated a linear relationship between turbulence intensity and the Reynolds number based on the average cylinder rotation speed and wall separation distance. Measured integral and micro time scales and approximated integral length scales were all found to decrease with the Reynolds number, possibly associated with a confinement of the largest scales (of the order of the cylinder wall separation distance). Regions of transverse isotropy were discovered in axial–radial cross correlations for average cylinder Reynolds numbers less than 6000 and are predicted to exist also for circumferential cross correlations at higher average Reynolds numbers, greater than 6000. Power spectra for the independent directions of velocity fluctuation exhibited −5/3 slopes, suggesting that the flow also has some additional isotropic characteristics and demonstrating the role of the Taylor–Couette apparatus as a novel means for generating turbulence for flame propagation studies. [Copyright &y& Elsevier]

The statistical and structural properties of two wall jets, one developing along a real wall and the other along a shear-free wall, are contrasted with the equivalent properties in the post-reattachment-recovery and reverse-flow regions of a separated backward-facing step flow, respectively. The study was motivated, principally, by the wish to isolate and understand the mechanisms responsible for the poor representation of the post-reattachment recovery returned by most RANS closures. It demonstrates a substantive commonality in the turbulent processes between, on the one hand, the real wall jet and post-reattachment recovery, and on the other hand, between the zero-wall-shear jet and the reattachment zone, as well as the reverse-flowing near-wall layer in the backward-facing-step flow. All regions under consideration are found to be characterised by an influential interaction between an outer shear layer and the wall, but an important distinction arises from the presence or absence of strong near-wall shear which introduces significant changes to the nature of this interaction: in the absence or near-absence of wall shear, the interaction is largely inviscid, associated with wall-blocking, while strong shear tends to shield the wall from the outer flow. The commonality between the two sets of flow conditions is explored by way of mean-flow properties, budgets, anisotropy maps, departures from local equilibrium, non-dimensional strain parameter, length-scale variations and structural properties. The study shows that the post-reattachment-recovery region, like the real wall jet, is far from the state of local equilibrium found in a standard boundary layer, that the recovery of the flow towards a state of equilibrium is slow, and that the near-wall region is strongly affected by turbulent diffusion associated with the migration of large-scale structures towards the wall. In statistical terms, this last process is represented by the importance of turbulent transport by third moments and pressure-velocity correlations. Within the reattachment and reverse-flow layer, near-wall shear is weak, and so is the shielding of the wall from the outer layer. In these regions, the flow shares many of the properties of the zero-wall-shear jet. [ABSTRACT FROM AUTHOR]

Abstract: An advanced laser Doppler velocimetry system is developed to acquire measurements of fluctuating velocity–acceleration statistics in turbulent boundary layers. These correlations are enabled by customized burst signal processing that estimates both the Doppler frequency and the rate-of-change of Doppler frequency, which are related to the particle velocity and acceleration by the interference fringe spacing. The measurements give important insight into the near-wall turbulence structure since the statistical correlations of interest, appear directly in the Reynolds stress transport equations as a sum of the velocity–pressure gradient correlation, , the dissipation rate, , and the viscous diffusion, . The immediate power of such measurements is that combinations of terms in the Reynolds stress transport equation may be characterized by a single statistical measurement at one location in the flow—no gradients need be computed. In the present paper, data are presented for a constant-pressure 2D turbulent boundary layer at Re θ =6800. Near-wall results for the dominant term in the velocity–acceleration tensor, the streamwise correlation , compare favorably with DNS for the same quantity at Re θ =1410 and Re τ =640; furthermore, the quantity exhibits no Reynolds number effects within experimental uncertainties. The balance of the velocity–acceleration equation in the streamwise direction is presented, giving the first measurements for the profile of the velocity–pressure gradient correlation with this technique. This study exhibits the potential of the technique to be applied to more complex flows, particularly those 3D separating flows in which the motions contributing to the velocity–acceleration correlations become dominant. [Copyright &y& Elsevier]

Abstract: This article is a sequel to the recent paper by Ashrafian et al. [Ashrafian, A., Andersson, H.I., Manhart, M., 2004. DNS of turbulent flow in a rod-roughened channel, International Journal of Heat and Fluid Flow 25, 373–383] on direct numerical simulation of a turbulent flow in a rod-roughened channel. Higher-order statistics are compared and the effects of surface roughness on the relatively large scales of the turbulent flow are investigated. It is observed that roughness tends to increase the intensity of the vorticity fluctuations in the roughness sublayer. In the outer layer, however, the r.m.s. of the vorticity fluctuations are unaffected. The Reynolds stress anisotropy invariant maps for the smooth and rough cases clearly showed that the states of near-wall turbulence for the two cases were substantially different whereas in the regions away from the wall the two cases exhibited close similarities. Differences in the turbulent transport process between the two cases were shown through examination of the third-order moments of the velocity fluctuations and the turbulent kinetic energy budgets. It was observed that wall-ward transport of the kinetic energy is substantially increased very close to the wall while the away-from-the-wall transport of kinetic energy is relatively reduced at the edge of the roughness sublayer. The length scales and flow dynamics in the roughness sublayer were also investigated. It was found that turbulence statistics and structure outside the roughness sublayer are hardly affected by the condition at the walls. These observations lent support to the classical wall similarity hypothesis. [Copyright &y& Elsevier]

Two-dimensional experiments investigating sediment transport and turbulence structure in sustained turbidity currents that cross breaks in slope are presented as analogue illustrations for natural flows. The results suggest that in natural flows, turbulence generation at slope breaks may account for increased sand transport into basins and that the formation of a hydraulic jump may not be necessary to explain features such as the occurrence of submarine plunge pools and the deposition of coarser-grained beds in the bottomsets of Gilbert-type fan deltas. Experimental flows were generated on 0°, 3°, 6° and 9° slopes of equal length which terminated abruptly on a horizontal bed. Two-component velocities were measured on the slope, at the slope break and downstream of the slope break. Flows were depositional and non-uniform, visibly slowing and thickening with distance downstream. One-dimensional continuous wavelet transforms of velocity time series were used to produce time-period variance maps. Peaks in variance were tested against a background red-noise spectrum at the 95% level; a significant period banding occurs in the cross-wavelet transform at the slope break, attributed to increased formation of coherent flow structures (Kelvin–Helmholtz billows). Variance becomes distributed at progressively longer periods and the shape of the bed-normal-velocity spectral energy distribution changes with distance downstream. This is attributed to a shift towards larger turbulent structures caused by wake stretching. Mean velocity, Reynolds shear stress and turbulent kinetic energy profiles illustrate the mean distribution of turbulence through the currents. A turbulent kinetic energy transfer balance shows that flow non-uniformity arises through the transfer of mean streamwise slowing to mean bed-normal motion through the action of Reynolds normal stresses. Net turbulence production through the action of normal stresses is achieved on steeper slopes as turbulence dissipation due to mean bed-normal motion is limited. At the slope break, an imbalance between the production and dissipation of turbulence occurs because of the contrasting nature of the wall and free-shear boundaries at the bottom and top of the flows, respectively. A rapid reduction in mean streamwise velocity predominately affects the base of the flows and steeper proximal slope flows have to slow more at the break in slope. The increased turbulent kinetic energy, limited bed-normal motion and strong mixing imposed by steep proximal slopes means rapid slowing enhances turbulence production at the break in slope by focusing energy into coherent flow structures at a characteristic period. Thus, mean streamwise slowing is transferred into turbulence production at the slope break that causes increased transport of sediment and a decrease in deposit mass downstream of the slope break. The internal effects of flow non-uniformity therefore can be separated from the external influence of the slope break. [ABSTRACT FROM AUTHOR]

Abstract: We applied a technique that defines and extracts “structures” from a DNS dataset of a turbulence variable in a way that allows concurrent quantitative and visual analysis. Local topological and statistical measures of enstrophy and strain-rate structures were compared with global statistics to determine the role of mean shear in the dynamical interactions between fluctuating vorticity and strain-rate during transition from isotropic to shear-dominated turbulence. We find that mean shear adjusts the alignment of fluctuating vorticity, fluctuating strain-rate in principal axes, and mean strain-rate in a way that (1) enhances both global and local alignments between vorticity and the second eigenvector of fluctuating strain-rate, (2) two-dimensionalizes fluctuating strain-rate, and (3) aligns the compressional components of fluctuating and mean strain-rate. Shear causes amalgamation of enstrophy and strain-rate structures, and suppresses the existence of strain-rate structures in low-vorticity regions between enstrophy structures. A primary effect of shear is to enhance “passive” strain-rate fluctuations, strain-rate kinematically induced by local vorticity concentrations with negligible enstrophy production, relative to “active,” or vorticity-generating strain-rate fluctuations. Enstrophy structures separate into “active” and “passive” based on the level of the second eigenvalue of fluctuating strain-rate. We embedded the structure-extraction algorithm into an interactive visualization-based analysis system from which the time evolution of a shear-induced hairpin enstrophy structure was visually and quantitatively analyzed. The structure originated in the initial isotropic state as a vortex sheet, evolved into a vortex tube during a transitional period, and developed into a well-defined horseshoe vortex in the shear-dominated asymptotic state. [Copyright &y& Elsevier]

Gravity-driven bubbly flows are a specific class of flows, where all action is provided by gravity. An industrial example is formed by the so-called bubble column: a vertical cylinder filled with liquid through which bubbles flow that are introduced at the bottom of the cylinder. On the bubble scale, gravity gives rise to buoyancy of individual bubbles. On larger scales, gravity acts on nonuniformities in the spatial bubble distribution present in the bubbly mixture. The gravity-induced flow and flow structures can increase the inhomogeneity of the bubble distribution, leading to a turbulent flow. In this flow, specific scales are identified: a large-scale circulation with the liquid flowing upward in the center of the column and downward close to the wall. On the intermediate scale there are vortical structures; eddies of liquid, with a size on the order of the diameter of the column, that stir the liquid and radially transport the bubbles. On the small scale there is the local stirring of the bubbles. We describe the ideas developed over time and identify some open questions. We discuss the experimental findings on the turbulence generated, the stability of the flow, axial dispersion, and the similarities between bubble columns and air lifts. Especially for higher gas fractions, many questions still lack accurate answers. The lateral lift force in bubble swarms and the structure of the turbulence in the bubbly mixture are important examples of inadequately understood physical phenomena, providing many challenges for fundamental and applied research on bubbly flows. [ABSTRACT FROM AUTHOR]

Turbulent momentum and heat transport in idealized Czochralski crystal growth configurations is investigated by means of direct numerical simulation. The analysis of the flow data focuses on the influence of crystal and crucible rotation on the flow structures and the development of temperature fluctuations. A numerical parameter study is performed to investigate how the variation of the numerous flow parameters affects the turbulent transport processes. Finally, a direct numerical simulation is conducted with parameters taken from experiment in order to allow a direct comparison between numerical and experimental results. It is found that counter-rotation of the crystal and crucible leads to a complex flow, which is characterized by three major recirculation zones, if crucible rotation dominates the flow. The dynamics of the flow are controlled by centrifugal forces counteracting buoyancy and surface tension effects. High temperature fluctuations are created within or close to the crystallization zone. Neither a variation of the melt height, nor a reduction of the crystal rotation rate or a change of the Grashof and Marangoni numbers has a major effect on the bulk flow structure and overall heat transfer. Increasing rotation of the crystal changes the bulk flow structure strongly and leads to an increased value of maximum rms temperature fluctuations, the position of which is shifted towards the crucible bottom. A shifted position of maximum rms temperature fluctuations is also observed if heat radiation across the free surface is taken into account. [Copyright &y& Elsevier]

In this paper, the possibility of extending the Kolmogoroff''s model of multi-scale turbulence structure to dense slurry flows in pipelines is explored. The analysis is based on experimental data of slurry behavior as a Newtonian liquid with some equivalent density and viscosity. A connection is made between the turbulence structure and the head losses in the pipeline that can be calculated by the known engineering methods. The latter, the total loss, is considered as comprising of viscous and turbulent components, and the turbulent component is associated with the turbulence structure. In this paper, the lower turbulence scale for the dense slurries is estimated. For the simulation of turbulence in stratified flows, a modified application of the known two-layer model of slurry flow is proposed. [Copyright &y& Elsevier]

A meteorological tower significantly modifies the air flow, the mean wind speed and wind direction as well as the turbulence structure of the air. Such changes can be noticed in particular in the wake region of the tower. Measurements on the 200 m tower of Forschungszentrum Karlsruhe were carried out using Solent sonic anemometers in the lee of the tower and cup anemometers on both sides. In the wake region, spectral energy density is increased in the high-frequency range. Superposition of this disturbance spectrum on the undisturbed spectrum yields a `knee' in the resulting spectrum. In the case of low turbulence intensity with stable stratification, a plateau with a constant energy content is observed in front of the knee. This effect is caused by the new production of turbulence energy from the mean flow as well as by an energy transfer from larger to smaller vortices. Power spectra in strongly stable conditions show a more rapid decrease of intensity in the region where the inertial subrange is expected. The relevant scales of wake turbulence are derived from the maximum of the disturbance spectrum. Locations of the high-frequency peak do not depend on atmospheric stability, but are controlled mainly by mean wind speed. Apart from the reduction of the mean wind speed, the spectra and cospectra exhibit a strong anisotropy for such cases. The results demonstrate the significant influence of a tower on turbulence spectra in the wake region. [ABSTRACT FROM AUTHOR]

Three aircraft-based studies of boundary-layer fronts (BLFs) were carried out during the experiment KABEG in April 1997 near the sea-ice edge over the Davis Strait. The boundary-layer flow was parallel to the ice edge and hence two independent turbulent regimes could develop in an identical synoptic framework, separated by the frontal zone along the ice edge. The zone of strongest crosswind horizontal gradients was typically 20 km wide, while the downstream scale of the BLF was observed to be several hundreds of kilometres. For two of the three cases the investigation of turbulence structures was possible and the results are given herein. Horizontal and vertical profiles of turbulent fluxes and other turbulent quantities are presented. A spectral analysis reveals the coexistence of small-scale turbulence with roll motions. These roll motions can be hidden or can be visible as cloud streets. The associated transport mechanisms are highly relevant for the choice of suitable averaging intervals for turbulent flux calculations and model validation. Parameterizations for the vertical velocity variance, countergradient transport, sea surface roughness and eddy diffusivity are evaluated and compared for this baroclinic strong-wind convective boundary-layer environment. Analogously, drag coefficients and bulk transfer coefficients are derived from measurements. [ABSTRACT FROM AUTHOR]

Wind-tunnel simulations of the atmospheric stable boundary layer (SBL) developed over a rough surface were conducted by using a thermally stratified wind tunnel at the Research Institute for Applied Mechanics (RIAM), Kyushu University. The present experiment is a continuation of the work carried out in a wind tunnel at Colorado State University (CSU), where the SBL flows were developed over a smooth surface. Stably stratified flows were created by heating the wind-tunnel airflow to a temperature of about 40–50°and by cooling the test-section floor to a temperature of about 10°. To simulate the rough surface, a chain roughness was placed over the test-section floor. We have investigated the buoyancy effect on the turbulent boundary layer developed over this rough surface for a wide range of stability, particularly focusing on the turbulence structure and transport process in the very stable boundary layer. The present experimental results broadly confirm the results obtained in the CSU experiment with the smooth surface, and emphasizes the following features: the vertical profiles of turbulence statistics exhibit different behaviour in two distinct stability regimes with weak and strong stability, corresponding to the difference in the vertical profiles of the local Richardson number. The two regimes are separated by the critical Richardson number. The magnitudes in turbulence intensities and turbulent fluxes for the weak stability regime are much greater than those of the CSU experiments because of the greater surface roughness. For the very stable boundary layer, the turbulent fluxes of momentum and heat tend to vanish and wave-like motions due to the Kelvin–Helmholtz instability and the rolling up and breaking of those waves can be observed. Furthermore, the appearance of internal gravity waves is suggested from cross-spectrum analyses. [ABSTRACT FROM AUTHOR]

Turbulence measurements from a 30 m tower in the stably stratified boundary layer over the Greenland ice sheet are analyzed. The observations include profile and eddy-correlation measurements at various levels. At first, the analysis of the turbulence data from the lowest level (2 m above ground) shows that the linear form of the non-dimensional wind profile (ϕm) is in good agreement with the observations for z/L <0.4, where L represents the Obukhov length. A linear regression yields ϕm=1+5.8z/L. The non-dimensional temperature profile (ϕh) at the 2m level shows no tendency to increase with increasing stability. The data from the upper levels of the tower are analyzed in terms of both local scaling and surface-layer scaling. The ϕm and the ϕh values show a tendency to level off at large stability (z/Λ>0.4) where Λ represents the local Obukhov length. Hence, the linear form of the ϕ functions is no longer appropriate under such conditions. The best correspondence to the data can be achieved when using the expression of Beljaars and Holtslag for ϕm and ϕh. The vertical profiles of the turbulent fluxes, the wind velocity variances and temperature variance are also determined. The momentum flux profile and the profiles of wind speed variances are in general agreement with other observations if a well developed low-level wind maximum occurs, and the height of this maximum is used as a height scale. [ABSTRACT FROM AUTHOR]

Turbulence structure in stably stratified boundary layers is experimentally investigated by using a thermally stratified wind tunnel. A stably stratified flow is created by heating the wind tunnel airflow to a temperature of about 50 °C and by cooling the test-section floor to a surface temperature of about 3 °C. In order to study the effect of buoyancy on turbulent boundary layers for a wide range of stability, the velocity and temperature fluctuations are measured simultaneously at a downwind position of 23.5 m from the tunnel entrance, where the boundary layer is fully developed. The Reynolds number, Reδ, ranges from 3.14 × 104 to 1.27 × 105, and the bulk Richardson number, Riδ, ranges from 0 to 1.33. Stable stratification rapidly suppresses the fluctuations of streamwise velocity and temperature as well as the vertical velocity fluctuation. Momentum and heat fluxes are also significantly decreased with increasing stability and become nearly zero in the lowest part of the boundary layer with strong stability. The vertical profiles of turbulence quantities exhibit different behaviour in three distinct stability regimes, the neutral flows, the stratified flows with weak stability (Riδ = 0.12, 0.20) and those with strong stability (Riδ = 0.39, 0.47, 1.33). Of these, the two regimes of stratified flows clearly show different vertical profiles of the local gradient Richardson number Ri, separated by the critical Richardson number Ri cr of about 0.25. Moreover, turbulence quantities in stable conditions are well correlated with Ri. [ABSTRACT FROM AUTHOR]

• Open channel flow with floating-leaved vegetation was experimentally investigated. • Vertical distribution of streamwise velocity was predicted mathematically. • The turbulence structure was examined via spectral and quadrant analyses. • The vortices are mainly affected by stem-scale turbulence. Floating-leaved vegetation alters the flow structure through the combined action of floating leaves, stems, and the channel bed. This study investigated the vertical distribution of streamwise velocity and the turbulence structure of an open channel flow with floating-leaved vegetation through theoretical analysis and laboratory experiments. Bionic lotus leaves fixed on circular wooden cylinders were used to imitate floating-leaved vegetation, and the 3D velocity field was measured by using an acoustic Doppler velocimeter. The measured data showed that with the presence of the channel bed and vegetation cover, strong velocity gradients appear at the bottom and top layers, whereas the velocity is almost uniformly distributed in the internal layer. On the basis of the modified mixing length hypothesis, the vertical distribution of streamwise velocity was numerically predicted, and the prediction results agreed well with the experimental data. The model parameters were then determined, and the effect of mixing length was analyzed. Spectral analysis showed that the peak frequency in the streamwise velocity spectra is within the range of 0.1 Hz to 0.2 Hz and that the vortices are mainly affected by stem-scale turbulence. Quadrant analysis showed that the ejections and sweeps are the major contributors to Reynolds stress in the bottom layer, inward and outward interactions are prominent in the top layer, and the contribution of sweeps and ejections are approximately equivalent to those of outward and inward interactions in the internal layer. Different to open channel flow featured with emergent rigid vegetation, the velocity and wake production decrease while the Reynolds stress and shear production increase towards the floating leaf. The effect of stem diameter, vegetation density, stem flexibility and boundary roughness on flow structure of open channel flow with floating-leaved vegetation need to be examined in future research. [ABSTRACT FROM AUTHOR]

This paper aims to identify the atmospheric boundary layer turbulence structure and its effect on severe foggy haze events frequently occurring in Northern China. We use data collected from a ground eddy covariance system, meteorology tower, and a PM 2.5 collector in Baoding, China during December 2016. The data shows that 73.5% of PM 2.5 concentration is greater than 100 μg m−3 with a maximum of 522 μg m−3. Analyses on vertical turbulence spectrum also reveal that 1) during the pollution period, lower wind can suppress large-scale turbulence eddies, which are more likely inhomogeneous, breaking into small-scale eddies, and 2) the air pollutant scattering effect for radiation could decrease the air temperature near the ground and generate weak vertical turbulence during the daytime. At night, air pollutants suppress the land surface cooling and decrease the air temperature difference as well as the vertical turbulence intensity difference. The vertical turbulence impact analysis reveals that the percentage of large-scale turbulence eddies can also change the atmospheric vertical mixing capacity. During the daytime, the air pollution evolution is controlled by the wind speed and vertical turbulence intensity. While at night, the vertical turbulence is weak and the atmospheric vertical mixing capacity is mainly controlled by the large-scale eddies' percentage. The increased number of large-scale turbulence eddies led by low wind at night could increase the vertical mixing of air pollutants and decrease its concentration near the ground. Image 1 • The lower wind can suppress large-scale turbulence eddies break into small-scale eddies. • At night, the vertical turbulence is weak and the atmospheric vertical mixing capacity is mainly controlled by the large-scale eddies' percentage. • The increased number of large-scale turbulence eddies at night could increase the vertical transport of air pollutants. [ABSTRACT FROM AUTHOR]

Studying the effects of turbulent structure with different scale on aero-optics is of great significance to improve the accuracy of wavefront measurement and simulation, and to guide the suppression of aero-optics through large eddy breakup. Based on the high-spatial resolution (11.4 μm/pixel) density field of the supersonic (M ∞ = 3.0) turbulent boundary layer (SPTBL) obtained by Nano-tracer-based planar laser scattering (NPLS) technique, the curvelet transformation is used to extract the density field at a specific turbulence scale. The statistical results of the propagation of Gaussian plane light wave in the corresponding density field at different turbulence scales are calculated. The smaller the scale of turbulence structure is, the less contribution it makes to aero-optical distortion. The fully developed SPTBL has less effect on the beam center offset (x ¯), but more obvious effect on the beam spread about the center ( x ' 2 ¯). x ¯ is mainly caused by the turbulent structure with relatively large scale (turbulence characteristic length scale L j > δ), while the contribution of large-scale structure to the x ' 2 ¯ is not significant. The x ' 2 ¯ is mainly caused by the turbulent structure with relatively small size (L j ≈ 0.13 δ) in the flow field. The smaller scale (L j < 0.94 θ) turbulence structure has less influence on the x ¯ and x ' 2 ¯. θ can be also used as a validation index of aero-optical effective scale. [ABSTRACT FROM AUTHOR]

To obtain the differences in the turbulence structures between the light and heavy haze pollution process, a light pollution episode (LPE) from November 3 to 72,017, and a heavy pollution episode (HPE) from December 15 to 232,016, were investigated. The weather conditions, mean meteorological fields, and turbulence structures of the two episodes were analysed. The results show that the two cases have similar weather backgrounds, and the differences in mean meteorological field are mainly related to the difference in the degree of solar radiation weakening. The turbulence during the HPE was suppressed to a greater extent in both thermal and dynamic effects. The structure of turbulence is weak and typically characterized by intermittent turbulence in the stable boundary layer. The parameter IS (Intermittent Strength) is introduced to indicate the strength of intermittency. Because of the complexity of atmospheric motion, we use the change in IS, i.e., ∆ IS, to indicate the change in the intensity of turbulent intermittency during each pollution process. In the accumulation stage, the weakening of turbulent exchange caused by the influence of sub-mesoscale motion may result in poor turbulent diffusion conditions, which should be considered because they can lead to heavy pollution events when ∆ IS > 0.1. The same conclusions were verified in an analysis of Tianjin for these two cases. Further analysis of the vertical structure of turbulence indicated that during the pollution stage of the LPE, the turbulent energy comes from the surface and is consistent with the traditional surface layer turbulent transport mechanism. Moreover, the upper layer intermittent strength is stronger than that in the lower layer, and ∆ IS < 0.1 at different heights. During the pollution stage of the HPE, the boundary layer is almost decoupled from the surface, and ∆ IS > 0.1 in all layers, which suggests that the weakening of turbulent exchange affected the vertical turbulence structure and resulted in pollutant accumulation. During the dissipation stage of both the HPE and the LPE, the low-level jets provide a turbulence source in the upper layer. The effect of sub-mesoscale motion is weak, and the turbulence component diffuses pollutants. In this stage, ∆ IS < 0.1 in all layers. The upside-down structure of the atmospheric boundary layer leads to full coupling with the ground and promotes the dissipation of pollutants during the HPE. And, the analysis of the pollution processes, from December 16, 2016 to January 8, 2017, also indicate that ∆IS > 0.1 is a feature of severe pollution. The results for these cases indicate that ∆ IS = 0.1 can be used as an indicator to characterize the turbulent exchange conditions and different levels of contamination. • To describe the differences in the turbulence structures between the light and heavy haze pollution episodes • Introduce the parameter IS to indicate the strength of turbulence intermittency. • ∆ IS = 0.1 can be used as an indicator to characterize the turbulent exchange and pollution conditions. [ABSTRACT FROM AUTHOR]