Interactive boundary layers in turbulent flow

An asymptotic analysis of the structure of the flow at high Reynolds number around a streamlined body is presented. The boundary layer is turbulent. This question is studied with the successive complementary expansion method, SCEM. The starting point is to look for a uniformly valid approximation (UVA) of the velocity field, including the boundary layer and the external flow. Thanks to the use of generalized expansions, SCEM leads to the theory of interactive boundary layer, IBL. For many years, IBL model has been used successfully to calculate aerodynamic flows. Here, the IBL model is fully justified with rational mathematical arguments. The construction of a UVA of the velocity profile in the boundary layer is also studied. To cite this article: J. Cousteix, J. Mauss, C. R. Mecanique 335 (2007).     

Outer flow scaling of smooth and rough wall turbulent boundary layers

Mean velocity profiles in a zero pressure gradient turbulent boundary layer were measured on a hydraulically smooth surface and three different rough surfaces created from sand paper, perforated plate, and woven wire mesh. The physical size and geometry of the roughness elements were chosen to encompass both transitionally and fully rough flow regimes. The mean velocity profiles were measured using a Pitot tube in a subsonic wind tunnel, for Reynolds numbers (based on momentum thickness) ranging from 3,730 to 12,260. Three different outer velocity scales were used to analyze the defect profile. The results show that application of a so called mixed outer scale causes the velocity profile in the outer region to collapse onto the same curve for different Reynolds numbers and roughness conditions. Although the mixed scale collapses defect profiles on different surfaces, the effect of surface roughness is still observed in the outer region.     

An experimental study of turbulent boundary layers approaching separation

The present paper deals with the experimental analysis of a strong decelerated turbulent boundary layer developed on a flat plate. The aim of the study was to examine the effects of pressure gradient on a non-equilibrium boundary layer while indicating local areas of equilibrium flow. The effect of the Reynolds number on a turbulent boundary layer developed with matching the external pressure gradient conditions was also analysed. The emphasis was on the analysis of mean flow statistics i.e. mean velocity profiles, streamwise Reynolds stress and the effect of large- and small-scale interactions by analysing the skewness factor and energy isocontours maps. The comparative analysis of the external data indicated that the structure of the turbulent boundary layer depends not only on local effects of pressure gradient but also on the upstream history of the flow. For the same condition of pressure gradient, the increased momentum is observed near the wall with the increase of the Reynolds number at the Incipient Detachment, where increased turbulence production is also observed, leading to the failure of the outer scaling methods. Surprisingly, the effect of the Reynolds number decays at the intermittent transitory detachment where similar profiles were observed. The upper inflection point in the mean profile corresponded well with the outer maximum of the Reynolds stress and zero crossing of skewness factor. Position of this point occurs at different locations, depending on the flow history effects. The last observation demonstrates that the inflection points results from large- and small-scale interactions, which led to the increased convection velocity of small scales near the wall.     

Drag reducing outer-layer devices in rough wall turbulent boundary layers

The ability of outer-layer devices to reduce wall shear stress over a substantial streamwise distance in rough-wall turbulent boundary layers has been studied experimentally. The devices examined are a pair of thin flat ribbons placed in tandem as well as those having symmetric airfoil sections. The wall conditions examined are smooth, d- and k-type transverse-groove and sandgrain roughnesses. The wall drag is found to be reduced from the respective normal levels in all rough walls. All k-type rough walls exhibit a similar level of relative wall drag reduction which is also smaller than that in a smooth-wall. The d-type rough walls exhibit a transitional behaviour — the relative wall drag reduction drops from the smooth wall level to that of the k-type roughness with increasing roughness Reynolds number. However, the absolute reductions in the local wall shear stress are similar in both the rough and smooth walls. On the other hand, the relative reductions are lower in the rough walls because of a higher reference drag which is caused by the unique presence of a pressure component on which the devices are not as effective.     

Scaling of the bursting period in turbulent rough wall boundary layers

Measurements in laboratory rough wall turbulent boundary layers have been made over a sufficient Reynolds number range for independence on the Reynolds number to be established adequately. The data have been analyzed with respect to the average period between ejections and sweeps, using methods previously applied to smooth wall flows. The results clearly show that the period is independent of Reynolds number when it is scaled on outer variables.     

Shock-induced turbulent boundary layers

Experimental data for an incompressible turbulent moving surface boundary layer are reviewed and a theoretical extension of their predictions is suggested for the case of finite free stream velocities. It is argued that such a boundary layer provides an incompressible analogue for shock-induced turbulent boundary layers. Coles's transformation is used to predict the behaviour of the shock-induced case from the incompressible analogue. These predictions are used to attempt to correlate the available experimental shock-induced turbulent boundary layer data. It is felt that the correlations are reasonably successful for some of the data. It is suggested that the remaining data have been affected by the premature arrival of the contact region and reflected rarefaction wave.     

Mechanism of turbulent boundary layer separation from a smooth wall

The mechanism of turbulent boundary layer separation under the influence of a positive pressure gradient is analyzed. The process of turbulent separation from a smooth wall in a plane diffuser channel has been experimentally investigated. It is shown that separation is determined by the nature of the flow in a certain inner part of the boundary layer, where the friction effect is unimportant. This region of the boundary layer is most exposed to the action of the positive pressure gradient and it is there that the stagnant zone primarily appears.Translated from Izvestiya Akademii Nauk SSSR, Mekhanika Zhidkosti i Gaza, No. 6, pp. 69–77, November–December, 1990.     

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.     

Fence probe measurement of flow reversal in separating turbulent boundary layers

A small fence probe was evaluated for measurements in the time-dependent flow reversal region of the transition from boundary layer to separated flow. For moderate and high Reynolds numbers, the fence probe is demonstrated to be a usable tool for the measurement of the reverse flow associated with separation. Although the present probe pressure transducer system was limited to approximately 200?Hz, pulses of positive and negative shear stress were readily detected. At or near the location of zero surface shear stress, the measurements were limited by the signal-to-noise ratio. For the separated flow investigated, a marked reduction in the pressure gradient occurred when the fence probe indicated approximately 20?% reversal for the higher Reynolds numbers. The reversal increased to 24?% for the lower Reynolds numbers. The measurements indicate that flow reversal alone may not be adequate to identify the degree of separation. Upstream of turbulent boundary layer (intermittent) separation, the duration of the reversed shear stress was found to be very short (0.002?C0.007?s), suggesting a local, small-scale, impulse-type separation. At and beyond the location of intermittent separation, the shear stress reversal duration was an order of magnitude longer. Estimates of the maximum and minimum surface shear stress in the separation region were also obtained with the fence probe.     

Direction of streamlines near the wall in a three-dimensional turbulent boundary layer

Expressions are derived that, for a large class of flows, enable one to determine the direction of the streamlines near the wall in a three-dimensional turbulent boundary layer without integrating the system of boundary-layer equations. Comparison with available experimental data reveals a completely satisfactory agreement between the calculations and the experimental data.Translated from Izvestiya Akademii Nauk SSSR, Mekhanika Zhidkosti i Gaza, No. 2, pp. 19–25, March–April, 1980.     

Simulation of particles released near the wall in a turbulent boundary layer

A direct numerical simulation was used along with a Lagrangian particle tracking technique to study particle motion in a horizontal, spatially developing turbulent boundary layer along an upper-wall (with terminal velocity directed away from the wall). The objective of the research was to study particle diffusion, dispersion, reflection, and mean velocity in the context of two parametric studies: one investigated the effect of the drift parameter (the ratio of particle terminal velocity to fluid friction velocity) for a fixed and finite particle inertia, and the second varied the drift parameter and particle inertia by the same amount (i.e. for a constant Froude number). A range of drift parameters from 10⁻⁴ to 10⁰ were considered for both cases. The particles were injected into the simulation at a height of four wall units for several evenly distributed points across the span and a perfectly elastic wall collision was specified at one wall unit.Statistics collected along the particle trajectories demonstrated a transition in particle movement from one that is dominated by diffusion to one that is dominated by gravity. For small and intermediate sized particles (i.e. ones with outer Stokes numbers and drift parameters much less than unity) transverse diffusion away from the wall dominated particle motion. However, preferential concentration is seen near the wall for intermediate-sized particles due to inhomogeneous turbulence effects (turbophoresis), consistent with previous channel flow studies. Particle–wall collision statistics indicated that impact velocities tended to increase with increasing terminal velocity for small and moderate inertias, after which initial conditions become important. Finally, high relative velocity fluctuations (compared to terminal velocity) were found as particle inertia increased, and were well described with a quasi-one-dimensional fluctuation model.     

Bursting frequency in turbulent boundary layers

Employing laser Doppler anemometry and VITA techniques, the bursting frequency in turbulent boundary layers has been measured over the Reynolds-number range 320 to 1470. The result indicates that the mean and non-dimensional bursting frequency scaled with the variables appropriate for the wall region was constant and independent of Reynoids number. When the same data are plotted using the outer variables of boundary layer to normalize the bursting frequency, the non-dimensional frequency increases as the Reynolds number increases. This is in agreement with the results of Blackwelder et al. (1983) who used hot wire anemometry and VITA technique. The project is supported by the National Natural Science Foundation of China     

Some equilibrium turbulent boundary layers

Using the Coles additive law of the wall and law of the wake for the mean velocity profile of a two-dimensional turbulent boundary layer, a differential equation for the friction and wake parameters is derived from the momentum integral equation with a view to finding out the conditions under which the boundary layer can exhibit equilibrium. It is predicted that equilibrium is possible for boundary layers in favorable pressure gradient over smooth as well as k-type rough walls. When the roughness height is allowed to increase linearly with the streamwise distance, equilibrium exists also in zero pressure gradient. For a d-type rough wall, equilibrium is possible for a certain range of pressure gradients, from favorable to adverse. Most of the predictions are verified by evaluating the friction and wake parameters from the available experimental data on mean velocity measurements.     

Vortex reynolds number in turbulent boundary layers

The effects of vortex Reynolds number on the statistics of turbulence in a turbulent boundary layer have been investigated. Vortex Reynolds number is defined as the ratio of circulation around the vortex structure to the fluid viscosity. The vortex structure of the outer region was modeled and a full numerical simulation was then conducted using a high-order spectral method. A unit domain of the outer region of a turbulent boundary layer was assumed to be composed of essentially three elements: a wall, a Blasius mean shear, and an elliptic vortex inclined at 45° to the flow direction. The laminar base-flow Reynolds number is roughly in the same range as that of a turbulent boundary layer based on eddy viscosity, and the vortex-core diameter based on the boundary-layer thickness is nearly the same as the maximum mixing length in a turbulent boundary layer. The computational box size, namely, 500, 150, and 250 wall units in the streamwise, surface-normal, and spanwise directions, respectively, is approximately the same as the measured quasi-periodic spacings of the near-wall turbulence-producing events in a turbulent boundary layer. The effects of vortex Reynolds number and the signs of the circulation on the moments of turbulence were examined. The signs mimic the ejection and sweep types of organized motions of a turbulent boundary layer. A vortex Reynolds number of 200 describes the turbulence moments in the outer layer reasonably well.     

Surface roughness effects in turbulent boundary layers

The effects of surface roughness on a turbulent boundary layer are investigated by comparing measurements over two rough walls with measurements from a smooth wall boundary layer. The two rough surfaces have very different surface geometries although designed to produce the same roughness function, i.e. to have nominally the same effect on the mean velocity profile. Different turbulent transport characteristics are observed for the rough surfaces. Substantial effects on the stresses occur throughout the layer showing that the roughness effects are not confined to the wall region. The turbulent energy production and the turbulent diffusion are significantly different between the two rough surfaces, the diffusion having opposite sign in the region γ/δ < 0.5. Although velocity spectra exhibit differences between the three surfaces, the mean energy dissipation rate does not appear to be significantly affected by the roughness. Received: 19 August 1998/Accepted: 16 February 1999     

A note on velocity-profile similarity of turbulent boundary layers with negligible wall stress

Summary The velocity profiles of a turbulent boundary layer with zero skin friction throughout its region of pressure rise, measured by Stratford in 1959, are analyzed in terms of a law of the wall and a velocity-defect law with a common velocity scale and a logarithmic velocity profile in the region of overlap. The analysis deviates from earlier work by Stratford and Townsend. It is shown that the flow in Stratford's boundary layer, even at the largest value of x ₁ at which measurements were taken, is not yet in a state of equilibrium. The velocity scale for turbulent boundary layers with zero skin friction is proportional to the cube root of the pressure gradient.