High Reynolds number thick axisymmetric turbulent boundary layer measurements

Experimental measurements of the wall shear stress and momentum thickness for thick axisymmetric turbulent boundary layers are presented. The use of a full-scale towing tank allowed zero pressure gradient turbulent boundary layers to be developed on cylinders with diameters of 0.61, 0.89, and 2.5 mm and lengths ranging from 30 m to 150 m. Moderate to high Reynolds numbers (10⁴<Re <10⁵, 10⁸<Re _L<10⁹) are considered. The relationship between the mean wall shear stress, cylinder diameter, cylinder length, and speed was investigated, and the spatial growth of the momentum thickness was determined. The wall shear stress is significantly higher, and the spatial growth of the boundary layers is shown to be lower than for a comparable flat-plate case. The mean wall shear stress exhibits variations with length that are not seen in zero pressure gradient flat plate turbulent boundary layers. The ratio of outer to inner boundary layer length scales is found to vary linearly with Re , which is qualitatively similar to a flat plate turbulent boundary layer. The quantitative effect of a riblet cylindrical cross-sectional geometry scaled for drag reduction based on flat plate criteria was also measured. The flat plate criteria do not lead to drag reduction for this class of boundary layer shear flows.List of symbols a cylinder radius, mm - A _s total cylindrical surface area, m² - C _d tangential drag coefficient - D drag force, Newtons - boundary layer thickness, mm - * displacement thickness, mm - h riblet height, mm - L cylinder length, m - kinematic viscosity, m²/s - momentum thickness, mm - fluid density, kg/m³ - r radial coordinate, mm - Re _L Reynolds number based on length= - Re Reynolds number based on momentum thickness= - s riblet spacing, mm - _w mean wall shear stress, N/m² - u(r) mean streamwise velocity, m/s - u friction velocity= - U _o tow speed, m/s - x streamwise coordinate, m     

Three-component LDA measurements in a turbulent boundary layer

An experimental study of a three-dimensional, pressuredriven, attached turbulent boundary-layer flow was made at Mach 0.4. Both the mean velocities and the full Reynolds stress tensor were measured simultaneously by a three-component LDA system. Value of the resultant shear stress to turbulent kinetic energy ratio varied between 0.1 and 0.2 and did not remain constant across the boundary-layer. Slopes of the streamwise and azimuthal mixing length distributions in the wall region were around 0.4 and 1.2, respectively. Skew angle of the turbulent shear stress was larger than skew angle of the velocity gradient.     

Large eddy simulation of a developing turbulent boundary layer at a low Reynolds number

A spectral code has been used to simulate a developing turbulent boundary layer at low Reynolds number Re_θ (based on free stream velocity and momentum thickness) between 353 and 576. The starting field was generated by allowing a step change of temperature to diffuse outwards from one wall in a fully developed channel flow. The thermal boundary layer so created was conditionally sampled to convert it into a momentum boundary layer with an irrotational free stream region, a process which is justified by appeal to experiments. This initial field was allowed to develop until the momentum boundary layer thickness δ₉₉₅ had grown to about 1·5 times its original thickness. The results of the simulation have been compared with a wide range of experimental data. The outcome of this comparison is generally very satisfactory; the main trends of the experiments are well reproduced and our simulation supplements and extends the existing sets of experimental data. The simulation also gives pressure statistics which cannot be obtained experimentally. In particular, it gives the contribution of pressure diffusion to the balance equations for the Reynolds stress and indicates the error produced by omitting this term.     

Low Reynolds number axisymmetric turbulent boundary layer on a cylinder

In the present study, an axisymmetric turbulent boundary layer growing on a cylinder is investigated experimentally using hot wire anemometry. The combined effects of transverse curvature as well as low Reynolds number on the mean and turbulent flow quantities are studied. The measurements include the mean velocity, turbulence intensity, skewness and flatness factors in addition to wall shear stress. The results are presented separately for the near wall region and the outer region using dimensionless parameters suitable for each case. They are also compared with the results available in the open literature.The present investigation revealed that the mean velocity in near wall region is similar to other simple turbulent flows (flat plate boundary layer, pipe and channel flows); but it differs in the logarithmic and outer regions. Further, for dimensionless moments of higher orders, such as skewness and flatness factors, the main effects of the low Reynolds number and the transverse curvature are present in the near wall region as well as the outer region.     

Double large field stereoscopic PIV in a high Reynolds number turbulent boundary layer

An experiment on a flat plate turbulent boundary layer at high Reynolds number has been carried out in the Laboratoire de Mecanique de Lille (LML, UMR CNRS 8107) wind tunnel. This experiment was performed jointly with LEA (UMR CNRS 6609) in Poitiers (France) and Chalmers University of Technology (Sweden), in the frame of the WALLTURB European project. The simultaneous recording of 143 hot wires in one transverse plane and of two perpendicular stereoscopic PIV fields was performed successfully. The first SPIV plane is 1?cm upstream of the hot wire rake and the second is both orthogonal to the first one and to the wall. The first PIV results show a blockage effect which based on both statistical results (i.e. mean, RMS and spatial correlation) and a potential model does not seem to affect the turbulence organization.     

Low Reynolds number effects in open-channel turbulent boundary layers

The present study reports measurements of a turbulent boundary layer in an open-channel flow using fiber-optic laser Doppler anemometry. The Reynolds numbers based on momentum thickness and depth of flow are in the range 750≤Re _θ ≤2,400 and 15,300≤Re _h ≤54,200, respectively. It is shown that an accurate estimate of the wall shear stress can be made by fitting a fifth-order polynomial to the near-wall data. The effect of Reynolds number on the mean turbulence intensity and triple correlation is examined using both conventional scaling laws and the recent scaling laws proposed by George and Castillo. The present results show that different scaling laws lead to different conclusions on low Reynolds number effects.     

Evaluation of LDA temporal and spatial velocity structure functions in a low Reynolds number turbulent channel flow

 Measurements have been made using Laser Doppler Anemometry (LDA) in a fully developed turbulent channel flow with the aim of determining second-order and third-order temporal and spatial structure functions of the longitudinal velocity fluctuation. A reliable determination of these moments requires the data to be corrected for the effect of noise. Correction procedures are outlined, based on the behaviour of temporal or spatial correlation functions in the limit of small time intervals or small separations. No a priori assumptions about the nature of the noise are made so that the procedure should be quite general. The corrected LDA data indicate that, especially for spatial separations, the effect of noise can be felt even within the inertial range. The corrected structure functions should allow an unambiguous assessment to be made of Taylor’s hypothesis and of the extended self-similarity (ESS) method; examples are given in each case. Temporal structure functions obtained by hot wire anemometry (HWA) are much less affected by noise than the LDA data. Received: 8 April 1998/Accepted: 26 October 1998     

On the logarithmic-law constants and the turbulent boundary layer at low Reynolds numbers

It is shown that a family of formally derived similarity solutions describe to leading order the outer region of a turbulent boundary layer for all Reynolds numbers for which the layer satisfies the logarithmic law-of-the-wall. The family includes Coles' [1] hypothesis. For consistency with this hypothesis and the logarithmic law-of-the-wall, it is further shown that the constants in the latter form the product κC=2+O(ε), suggesting the logarithmic law of the wall be written $${U \mathord{\left/ {\vphantom {U {U_\tau = \kappa ^{ - 1} }}} \right. \kern-\nulldelimiterspace} {U_\tau = \kappa ^{ - 1} }}\ln \left( {e^2 U_\tau {y \mathord{\left/ {\vphantom {y \nu }} \right. \kern-\nulldelimiterspace} \nu }} \right) + O\left( \in \right).$$ A range of data are reprocessed to determine the skin friction coefficientC _f using κC = 2 and these collapse well when plotted against momentum thickness Reynolds number, Re _θ . It is also shown that the form parameter, Π, in Coles hypothesis is not unique but is determined by history effects peculiar to the boundary layer. Expressions are derived forC _f (Re _θ ) and the shape factorH (Re _θ ); both agree closely with the data and are valid over all Reynolds numbers for which the logarithmic law of the wall is satisfied.     

Enhancement of finite Reynolds number effects: Inner-outer sublayer interaction in the turbulent boundary layer

Residual Reynolds number effects in the established data for the velocity profile in turbulent boundary layers (and in pipe or channel flows) are found to be remarkably large. We combine two eddy-viscosity models (with overlapping validity in the inertial sublayer) and show (both analytically and numerically) that this enhancement (which involves a viscous correlation length) arises from inner-outer sublayer interaction.     

Reynolds number effects on a turbulent boundary layer with separation,reattachment, and recovery

The present paper addresses experimental studies of Reynolds number effects on a turbulent boundary layer with separation, reattachment, and recovery. A momentum thickness Reynolds number varies from 1,100 to 20,100 with a wind tunnel enclosed in a pressure vessel by varying the air density and wind tunnel speed. A custom-built, high-resolution laser Doppler anemometer provides fully resolved turbulence measurements over the full Reynolds number range. The experiments show that the mean flow is at most a very weak function of Reynolds number while turbulence quantities strongly depend on Reynolds number. Roller vortices are generated in the separated shear layer caused by the Kelvin–Helmholtz instability. Empirical Reynolds number scalings for the mean velocity and Reynolds stresses are proposed for the upstream boundary layer, the separated region, and the recovery region. The inflectional instability plays a critical role in the scaling in the separated region. The near-wall flow recovers quickly downstream of reattachment even if the outer layer is far from an equilibrium state. As a result, a stress equilibrium layer where a flat-plate boundary layer scaling is valid develops in the recovery region and grows outward moving downstream.     

Combined influence of the Reynolds number and localised wall suction on a turbulent boundary layer

This paper considers the effect the Reynolds number has on a turbulent boundary layer that is subjected to concentrated suction, applied through a short porous wall strip. The response of the skin friction coefficient to suction is strongly modulated by the momentum thickness Reynolds number R ₀. The magnitude and wavelength of the variation of the skin friction decrease as R ₀ increases. Measurements clearly show that it is the combination of R ₀ and the suction rate that controls the boundary layer response. Relaminarisation of the near-wall flow, which occurs just downstream of the suction strip when =5.5 and R ₀=660, could not be achieved at higher Reynolds numbers. It is suggested that the ratio R ₀/ should not exceed a (as yet undetermined) critical value if pseudo-relaminarisation is to occur. The mean velocity profiles differ from the undisturbed profile for all the values of R ₀ and considered here. Although the redistribution of the turbulent kinetic energy among the normal Reynolds stresses is affected by , a variation in R ₀ does not appear to alter the effect of on this redistribution.     

Smooth and rough turbulent boundary layers at high Reynolds number

A 2-D turbulent boundary layer experiment with zero pressure gradient (ZPG) has been carried out over a rough and a smooth surface using two cross hot-wire probes. Wind tunnel speeds of 10 m/s and 20 m/s were set up in order to investigate the effects of the upstream conditions and the Reynolds number on the downstream flow. For a given set of upstream conditions, such as the wind tunnel speed, trip wire size and location, the three components of the velocity field were measured from about 14 m from the inlet of the wind tunnel to 30 m downstream. This experiment is unique because it achieves Reynolds numbers as high as R120,000, for which measurements of the mean velocity are reported. It is shown that by fixing the upstream conditions, the mean deficit profiles collapse with the freestream velocity, , but to different curves depending on the upstream conditions and surface roughness. Moreover, the effects of the upstream conditions, the Reynolds number, and roughness are completely removed from the outer flow when the mean deficit profiles are normalized by the Zagarola/Smits scaling, . Consequently, the true asymptotic profile in the turbulent boundary layer is found in ZPG flow regardless of the range of Reynolds number, surface conditions and initial conditions.     

Simultaneous velocity and passive scalar concentration measurements in low Reynolds number neutrally buoyant turbulent round jets

Velocity and scalar concentration characteristics of low Reynolds number (Re) neutrally buoyant turbulent round jets were studied using coupled particle image velocimetry and laser induced fluorescence. Experiments were conducted on a jet with a fully developed pipe exit profile at Re = 1,500 and Re = 4,000. Measurements were made in the far field (60 < x/D < 80). Results show that the centerline velocity decay constant increases with Re and the virtual origin is located far from the jet exit. The Re does not have such an influence on the scalar concentration decay constant and scalar field virtual origin. Centerline turbulent intensities show evidence of becoming self-similar past x/D ≈ 75 but the same is not observed in the scalar fluctuations. The latter are strongly influenced by the Re, increasing as entrainment of ambient flow on the jet axis increases as the flow approaches laminar conditions. Scalar fluxes showed a marked Re dependence as well. Results were used to estimate the turbulent Schmidt number which decreases as the Re decreases.     

The effect of Reynolds number on boundary layer turbulence

A new facility for studying high Reynolds number incompressible turbulent boundary layer flows has been constructed. It consists of a moderately sized wind tunnel, completely enclosed by a pressure vessel, which can raise the ambient air pressure in and around the wind tunnel to 8 atmospheres. This results in a Reynolds number range of about 20:1, while maintaining incompressible flow. Results are presented for the zero pressure gradient flat plate boundary layer over a momentum thickness Reynolds number range 1500–15?000. Scaling issues for high Reynolds number non-equilibrium boundary layers are discussed, with data comparing the three-dimensional turbulent boundary layer flow over a swept bump at Reynolds numbers of 3800 and 8600. It is found that successful prediction of these types of flows must include length scales which do not scale on Reynolds number, but are inherent to the geometry of the flow.     

Two-phase measurements of saltating turbulent boundary layer flow

Mean wind and particle speed measurements as functions of heigh were made for a saltating turbulent boundary layer flow. An exponential dependence of particle flux with height above the surface was found independent of windspeed and particle size. Particle-speed distributions as functions of height above the surface were measured for experiments conducted at two ambient pressures: atmospheric pressure and approximately 1% atmospheric pressure. The wind was shown to be a more efficient mover of particles at atmospheric pressure.High-speed motion pictures of saltating ground walnut shells (of diameter 500–1500 μm and density 1.1 g/cm³) were taken in an environmental wind tunnel to simulate the planetary boundary layer. These experiments verify the existence and magnitude of particle spin rates proposed by White & Schulz (1977). There was remarkable agreement between numerical trajectory solutions, including the spin effect, and the filmed trajectories. An observation was made that not all particles spin exclusively in the vertical longitudinal plane (in the direction of flow). At low pressures (0.6 kPa) the effect of spin forces on the particle's trajectory was shown to have little influence and was verified by a theoretical force ration balance of spin to drag force.     

RMS spanwise vorticity measurements in a turbulent boundary layer

Turbulent boundary layer measurements of the rms spanwise vorticity t_z with a four hot-wire probe are in reasonable agreement with direct numerical simulations and other published measurements at comparable Reynolds numbers. It is shown that a reasonable accurate approximation for _z can be obtained with only two parallel hot wires.     

Influence of Reynolds number on characteristics of turbulent wall boundary layers

Hot-wire anemometer measurements, using two types of probes, are reported for wall boundary layer flows with particular attention being given to the near-wall region and to measurements at high Reynolds numbers up to R 15,000. To obtain accurate near-wall measurements, the influence of wall proximity on hot-wire readings was eliminated by using a highly insulating wall material. Measurements were carried out with a single hot-wire boundary layer probe to obtain the longitudinal velocity informatemperature-wake sensor for the cross flow tion and a hot-wire, information.The results provided in the paper include measurements of averaged properties like mean velocity, rms-quantities of velocity fluctuations, probability density distributions etc. Conditional averages are also provided in order to yield information related to coherent flow structures present in boundary layer flows. It is shown that these structure remain present up to the highest Reynolds number investigated in the present study. The conditionally averaged data provide quantitative information on the mechanisms that are involved in the production of turbulence in boundary-layer flows.     

Some Reynolds number effects on two- and three-dimensional turbulent boundary layers

Experimental data for a two-dimensional (2-D) turbulent boundary layer (TBL) flow and a three-dimensional (3-D) pressure-driven TBL flow outside of a wing/body junction were obtained for an approach Reynolds number based on momentum thickness of Re _θ =23,200. The wing shape had a 3:2 elliptical nose, NACA 0020 profiled tail, and was mounted on a flat wall. Some Reynolds number effects are examined using fine spatial resolution (Δy ⁺=1.8) three-velocity-component laser-Doppler velocimeter measurements of mean velocities and Reynolds stresses at nine stations for Re _θ =23,200 and previously reported data for a much thinner boundary layer at Re _θ =5,940 for the same wing shape. In the 3-D boundary layers, while the stress profiles vary considerably along the flow due to deceleration, acceleration, and skewing, profiles of the parameter correlate well and over available Reynolds numbers. The measured static pressure variations on the flat wall are similar for the two Reynolds numbers, so the vorticity flux and the measured mean velocities scaled on wall variables agree closely near the wall. The stresses vary similarly for both cases, but with higher values in the outer region of the higher Re _θ case. The outer layer turbulence in the thicker high Reynolds number case behaves similarly to a rapid distortion of the flow, since stream-wise vortical effects from the wall have not diffused completely through the boundary layer at all measurement stations. Received: 9 June 2000/Accepted: 26 January 2001     

Full 3D correlation tensor computed from double field stereoscopic PIV in a high Reynolds number turbulent boundary layer

The turbulence structure near a wall is a very active subject of research and a key to the understanding and modeling of this flow. Many researchers have worked on this subject since the fifties Hama et al. (J Appl Phys 28:388–394, 1957). One way to study this organization consists of computing the spatial two-point correlations. Stanislas et al. (C R Acad Sci Paris 327(2b):55–61, 1999) and Kahler (Exp Fluids 36:114–130, 2004) showed that double spatial correlations can be computed from stereoscopic particle image velocimetry (SPIV) fields and can lead to a better understanding of the turbulent flow organization. The limitation is that the correlation is only computed in the PIV plane. The idea of the present paper is to propose a new method based on a specific stereoscopic PIV experiment that allows the computation of the full 3D spatial correlation tensor. The results obtained are validated by comparison with 2D computation from SPIV. They are in very good agreement with the results of Ganapthisubramani et al. (J Fluid Mech 524:57–80, 2005a).