Session GA: Turbulent Boundary Layers: Experimental Studies I

PIV Measurements of Turbulence in a Hypersonic Boundary Layer

Previous experiments on hypersonic turbulent boundary layers have documented the general features of the mean flow behavior, but virtually no high quality data exist describing the turbulence behavior for Mach numbers greater than about 5. To help improve our understanding of high Mach number wall-bounded turbulence, we perform PIV measurements of two components of velocity fluctuations in a flat plate, turbulent boundary layer at Mach 8 in a perfect gas, at a Reynolds number based on momentum thickness of about 4000. The results are compared with DNS under identical flow conditions. Supported under NASA Grant NNX08AB46A, Program Manager Catherine McGinley.   

Air Layer Drag Reduction

A set of experiments have been conducted at the US Navy's Large Cavitation Channel to investigate skin-friction drag reduction with the injection of air into a high Reynolds number turbulent boundary layer. Testing was performed on a 12.9 m long flat-plate test model with the surface hydraulically smooth and fully rough at downstream-distance-based Reynolds numbers to 220 million and at speeds to 20 m/s. Local skin-friction, near-wall bulk void fraction, and near-wall bubble imaging were monitored along the length of the model. The instrument suite was used to access the requirements necessary to achieve air layer drag reduction (ALDR). Injection of air over a wide range of air fluxes showed that three drag reduction regimes exist when injecting air; (1) bubble drag reduction that has poor downstream persistence, (2) a transitional regime with a steep rise in drag reduction, and (3) ALDR regime where the drag reduction plateaus at 90{\%} $\pm $ 10{\%} over the entire model length with large void fractions in the near-wall region. These investigations revealed several requirements for ALDR including; sufficient volumetric air fluxes that increase approximately with the square of the free-stream speed, slightly higher air fluxes are needed when the surface tension is reduced, higher air fluxes are required for rough surfaces, and the formation of ALDR is sensitive to the inlet condition.   

Turbulent Drag Reduction Using Superhydrophobic Surfaces

Superhydrophobic surfaces have received considerable attention for their ability to reduce drag in laminar flows. In this talk we demonstrate that engineered, micropatterned, superhydrophobic surfaces produce the same effect, with similar geometric scaling, in turbulent flows. Direct velocity measurements were used to measure slip velocities up to 40{\%} of the mean flow. Shear stress reductions up to 60{\%} were noted in comparison between smooth and superhydrophobic walls, with slip lengths up to 230\textit{$\mu $}m for several microridge geometries. Drag reduction was noted to increase with microfeature spacing and with reduced laminar sublayer thickness for a fixed shear free area ratio.   

PIV Measurements of Laminar and Turbulent Channel Flow with Superhydrophobic Walls

We report PIV measurements characterizing laminar and turbulent flow in a channel with a superhydrophobic bottom wall. The superhydrophobic wall is fabricated with micro-ribs and cavities that are oriented parallel and transverse to the flow direction and are subsequently coated with a hydrophobic coating. As a basis for comparison measurements were also made in a channel with smooth walls. The measurements span the Reynolds number range from 1000 to 7500. The channel hydraulic diameter is 8.3 mm with an aspect ratio of 10. The widths of the micro-ribs and cavities were 7.2 and 33 \textit{$\mu $}m, respectively, with a depth of 15 \textit{$\mu $}m. The PIV data captured aggregate velocities over multiple rib/cavity modules, such that a spanwise averaged velocity profile was obtained at the channel centerline. These measurements permitted characterization of the effective slip-velocity on the superhydrophobic wall in the laminar and turbulent flow regimes. The results for the laminar flow regime reveal no discernible slip velocity. In the turbulent flow regime, however, measurable slip was observed. Results presented include detailed time-averaged velocity, rms velocity, and the turbulent stress distributions.   

A Large Scale Wind Tunnel for the Study of High Reynolds Number Turbulent Boundary Layer Physics

Progress and the basic features of the University of New Hampshire's very large multi-disciplinary wind tunnel are reported. The refinement of the overall design has been greatly aided through consultations with an external advisory group. The facility test section is $73~m$ long, $6~m$ wide, and $2.5~m$ nominally high, and the maximum free stream velocity is 30 m/s. A very large tunnel with relatively low velocities makes the small scale turbulent motions resolvable by existing measurement systems. The maximum Reynolds number is estimated at $\delta^+ = \delta u_{\tau}/\nu \simeq 50000$, where $\delta$ is the boundary layer thickness and $u_{\tau}$ is the friction velocity. The effects of scale separation on the generation of the Reynolds stress gradient appearing in the mean momentum equation are briefly discussed to justify the need to attain $\delta^+$ in excess of about $40000$. Lastly, plans for future utilization of the facility as a community-wide resource are outlined. This project is supported through the NSF-EPSCoR RII Program, grant number EPS0701730.   

Experimental Investigation of Near-Wall Flow Structures in a Rough-Wall Turbulent Channel Flow

This study focuses on the near-wall flow structure in a turbulent channel flow with rough walls. To suppress strong reflections at the fluid-roughness interface that hamper typical optical measurements, we utilize a facility containing a fluid with the same refractive index as the rough acrylic wall, making the interface almost invisible. Presently, stereo (SPIV) is used to measure the flow field in the vicinity of uniformly distributed, closely-packed 0.45mm high pyramids with slope angle of 22$^{\circ}$. The flow is fully developed in the sample area, and has a friction velocity Reynolds number of 12,400 and wall units of 5.7-- 6.4 $\mu $m. Measurements are performed in several planes and magnifications. In y$>$3-5k, i.e. in the ``external flow,'' the mean flow and Reynolds Stress profiles agree with results of contemporary studies. However, near the wall, at y$<$3k, there is a remarkable upsurge in the normal Reynolds stress components, but not in the shear stress. On going effort involves high magnification SPIV measurements in the near wall region.   

Investigation of turbulent boundary layer structures using Tomographic PIV

Tomographic particle image velocimetry (TPIV) data were acquired in the logarithmic region of a zero pressure gradient turbulent boundary layer flow at friction Reynolds number \textit{Re}$_{\tau }$ = 1160. Experiments were conducted in a suction type wind tunnel seeded with olive oil particles of diameter $\sim $ 1\textit{$\mu $}m. The volume of interest was illuminated by two Nd:YAG laser beams expanded with appropriate optics into sheets of 8mm thickness in the wall-normal direction ($z)$. Images were acquired by four 2$k$ x 2$k$ pixel cameras, and correlation of reconstructed fields provided the full velocity gradient tensor in a volume of 0.7\textit{$\delta $} x 0.7\textit{$\delta $} x 0.07\textit{$\delta $}, which resolved the region $z^{+}$ = 70-150 in the log layer. Various vortex identification techniques, such as Galilean decomposition and iso-surfaces of two- and three-dimensional swirl, were utilized to visualize and analyze the eddy structures present in instantaneous fields. The results of the present study will be compared to results from earlier experimental studies that relied on planar PIV data only to identify vortices and vortex packets as well as from a direct numerical simulation of fully developed channel flow at comparable \textit{Re}$_{\tau }$.   

Flow of He II in a channel with ends blocked by superfleaks

We report experiments of He II flow thermally induced by a fountain pump in channels of square cross-section with ends blocked by sintered silver superleaks and its decay (PRL 100 (2008) 215302). We confirm a critical velocity $v_{\rm{cr}}^{\rm I}$ of order $1$~cm/s, which does not scale with the channel size and is therefore an intrinsic property of the self-sustained vortex tangle of vortex line density $L$, measured by the second sound attenuation. In addition to the previously reported turbulent A-state characterized by $L^{1/2}=\gamma(T) (v-v_{\rm{cr}}^{\rm I})$ we have discovered a new B-state characterized by $L = \beta (v-v_{\rm{cr}}^{\rm II}) $. We offer a phenomenological model assuming that in the B-state the superflow matches the classical parabolic profile, with a slip velocity $v_{\rm{cr}}^{\rm II}$ of order few cm/s and that a confined viscous normal fluid flow of toroidal form is induced inside the channel due to the mutual friction force. When the fountain pump is switched off, after an initial decay, a confined quasi-viscous flow of a single-component fluid with effective kinematic viscosity $\nu_ {eff}(T)$ establishes, giving rise to the observed exponential decay. The calculated values of $\nu_{eff}(T)$ are in agreement with those deduced from other experiments on decaying He II turbulence.   

Experimental Investigation of Zero Pressure-Gradient Turbulent Boundary Layers Using Particle Image Velocimetry

Due to the practical importance of turbulent boundary layers in fluid dynamic engineering, there is a need to predict or compute their behaviour. Numerous experimental and numerical studies have been conducted to examine the characteristics of turbulent boundary layers. Low-speed PIV was employed to measure the stream-wise velocities of zero pressure-gradient boundary layers in the turbulence water tunnel research facility at Cambridge University Engineering Department. The measurement position was 4 m downstream of a tripping rod with freestream velocities of 0.53 m/s and 0.64 m/s respectively. The use of low-speed PIV in this experiment enabled measurements of the mean flow fields and the flow statistics. Data is also reported for normalized mean velocity based on friction velocity (u$_{\tau })$ calculated using Clauser chart method.   

Investigation of coherent structures in rough and smooth wall boundary layers using tomographic PIV

While mean-flow universality can be seen in fully-rough walled turbulent boundary layers, spectral data of turbulence does not collapse in the same way. It is believed that spectral and mean flow properties can be explained by reference to coherent structures existing within boundary layers. The condition (i.e. roughness) of the wall may affect the formation mechanism(s) by which coherent structures occur, and therefore affect the turbulence spectra. Progress towards fully 3D, fully time resolved data revealing the formation dynamics is presented, with comparison between smooth and rough-walled cases at high Reynolds numbers. The limitations and advantages of the tomographic PIV technique are discussed, with especial reference to turbulent boundary layers.