Session J11: Boundary Layers: Roughness III

Mean Velocity Profiles and Drag Measurements in Turbulent Boundary Layer Flows over Permeable Substrates

The impact of permeable substrates on skin-friction drag was evaluated experimentally under turbulent flow conditions using a modified flat plate boundary layer setup. Several 3D-printed geometric patterns were examined, taking into account the effective interfacial 'slip-lengths' emerging for each geometry. For a flat plate case with no pressure gradient, these slip lengths are related to the wall-normal shift in the mean streamwise velocity profile for the turbulent boundary layer. Direct force measurements were conducted and compared against a smooth wall test case. Particle image velocimetry (PIV) was used to measure the shift in the mean streamwise velocity profile and estimate changes in turbulence statistics. To obtain a finer representation of the mean velocity profiles, the PIV images were analyzed using a technique that yields single-pixel resolution in the wall-normal direction. These measurements were used to characterize near-wall changes in flow over the permeable substrates relative to smooth wall conditions.   

Role of interfacial roughness in turbulence over porous substrates

A porous substrate affects the overlying turbulence via both its permeability, which allows penetration of flow into the substrate, and its interfacial rough topography, which introduces extra disturbances into turbulence. This work aims to understand the role of interfacial roughness in affecting the properties of overlying turbulence for relatively small grain sizes in wall units, L⁺ ~< 70. We use two methodologies to approach this issue: a systematic exploration of the space of substrate parameters, including grain size L, bulk permeability K, and depth h, using direct numerical simulations; and a linearized framework to access a substrate’s admittance of velocity in response to overlying stress excitations. Our results suggest that flow properties including interfacial slip, drag increase, and structures of near-wall turbulence are essentially governed by a substrate’s emergent macroscopic attributes and largely independent of its microscopic attributes at the pore grain scale. A typical rough surface can be essentially regarded as a shallow porous substrate whose admittance of wall-normal velocity in response to pressure fluctuations is partially suppressed by the proximity of an impermeable floor.   

The effect of porous walls on very large scale motions in turbulent boundary layer

Two aspects of porous materials that substantially influence the turbulent boundary layer convecting over them are surface roughness and permeability, and although their individual effects is well documented, their combined action and non-linear interactions are poorly understood. Therefore, novel experiments were performed, where both permeability and roughness were altered systematically over the wall. Hot-wire anemometry and microphone measurements were performed above these surfaces in the boundary layer wind tunnel. The measurements performed at same permeability but different roughness based Reynolds number reveal the impact of surface roughness on the outer layer. Furthermore, the impact of wall-permeability to outer layer based scale demonstrates the relative importance of roughness and permeability on the outer layer statistics. In particular, for Very-Large-Scale-Motions (VLSMs) a demarcation based on permeability based Reynolds number is achieved, which sets the maximum levels of permeability permissible for the existence of VLSMs over porous walls. Increase in permeability based Reynolds number leads to spectral shrinkage, which reduces the wavelengths associated with the outer-layer peak.   

Measurement of Spanwise motions and secondary flows in a turbulent boundary layer formed over drag-increasing riblets

We experimentally investigate the effect of dispersive stresses and near-wall large-scale motions on the drag increase by non-optimal trapezoidal riblets. This investigation requires measurement in the vicinity of the riblet tips and within the valleys. Performing particle image velocimetry is ideal, however, the strong reflection from the surface prevents accessing any information close to the surface. For this reason, we use a novel technique to minimize the light reflection from the riblet surface and permit measurement in the vicinity of the wall. This technique works by coating the riblets with Rhodamine paint that emits light at a different wavelength when illuminated with a green light laser. We use two stereoscopic particle image velocimetry systems to visualize the flow above the riblet tips and within the valleys.

This work was supported by the Air Force Office of Scientific Research (AFOSR, program manager: Gregg Abate) under award number FA2386-21-1-4018 (program managers: David Newell and Ryan Carr, AOARD) and the Australian Research Council under Discovery Project DP170102595.   

Topographically-driven secondary flows and very-large-scale motions interactions at high Reynolds number boundary layers

Spanwise heterogeneous roughness-wall flows have now been fairly well explored due to the emergence of the secondary flows, known to result in highly three-dimensional flows in the cross-plane, causing considerable changes in the heat and momentum transfer properties. Most studies, however, are presented at moderate Reynolds numbers, hence limiting the range of scale separation. In this study, a turbulent boundary layer flow developing over a spanwise heterogeneous ridge-type topography is examined at high Reynolds numbers. The skin-friction coefficient and turbulent flow properties are measured using a floating-element drag balance together with two simultaneous single-point hot-wire anemometry at the two symmetry planes (ridge and valley), respectively. To investigate the impact of large-to-small scale separation and strength of the secondary motions on the flow behavior, three spanwise spacings are considered at increasing Reynolds numbers. The results revealed that the skin friction varies depending on the surface spanwise characteristic wavelength, owing to changes in the strength of the secondary flows while the turbulence properties showed changes in distribution of the spectral density across different wavelengths. Changes in the amplitude and frequency modulations due to differences in the inner-outer interactions will further be discussed.   

Ridge-induced secondary flows at high Reynolds number boundary layers

Significant research has recently been dedicated to the understanding of spanwise heterogenous roughness-driven secondary flows. These are shown to strongly modulate the baseflow, with significant alternations in the mean and turbulence properties. Despite the current vast body of work, most studies have been performed at low to moderate Reynolds numbers, restricting the investigation of secondary flows at high large-to-small scale separation. To address this, a turbulent boundary layer over a spanwise heterogeneous ridge-type topography is investigated. Using a floating-element drag balance and cross-plane stereoscopic particle image velocimetry, the skin friction and turbulence are examined. Three spanwise wave lengths at high Reynolds numbers are considered. The results showed that depending on the spanwise wavelength, the skin friction remains transitionally rough in a large portion of Reynolds numbers, but weakly asymptotes at high enough Reynolds numbers. Using the friction velocity, the inner and outer scaling of the mean velocity are examined to assess their impact on the roughness function and the wake region. Furthermore, the triple decomposition of the total shear stress allowed the quantification of the dispersive and turbulent stresses to the global momentum fluxes.   

The symmetry breaking effects of directional roughness in turbulent channel flow

Directional surfaces form a special class of roughness which breaks the forward/backward or left/right statistical symmetry of rough-wall turbulence, i.e., their fluid dynamic effect depends on the sign of the mean flow direction. In the present study, the directional dependence of the mean flow and turbulence statistics of a surface composed of triangular bars with scalene cross-section is investigated using direct numerical simulations of turbulent channel flow at friction Reynolds number 550. This surface is inspired by ripples and dunes which are typical bedforms found in nature and can experience flows from different directions. When the bars on the surface are normal to the flow direction, the surface corresponds to ratchet-type roughness. In this case, the Hama roughness function significantly differs between the windward and the leeward orientation of the high-slope side of the scalene bars. When the bars are aligned parallel to the flow, the surface corresponds to asymmetric ridge-type roughness. Here, the roughness effect is significantly lower due to the absence of pressure drag. Nevertheless, effects of directionality remain significant, since scalene ridges induce substantial imbalances in secondary currents which can result in net spanwise flow in the channel.    

Effect of the angle of attack on the drag performance of shark-inspired riblet surfaces

Rib-like structures observed on the skin of fast-swimming sharks are believed to help them experience lower drags and swim faster than other animals in the ocean. Drawing inspiration from this, researchers have been employing two-dimensional riblet surfaces with various cross-sectional shapes to modulate the experienced drag/lift forces under different dynamic conditions. However, much of the previous research has been focused on fully/partially textured one-sided plates examined in zero angle of attack boundary layers, with limited experimental investigations on the local shear stress distributions. In this talk, we will consider a fully textured, symmetric, slender finite-size riblet plate in its entirety, to investigate the effect of the asymmetry imposed by the change of the angle of attack on local performance of the textures on either side of the plate. We present the results of experiments performed in a water tunnel where simultaneous velocity and load measurements are conducted via high-resolution particle image velocimetry (PIV) and load cell, respectively. Through the PIV measurements, we assess local velocity profiles and shear stress distributions in the boundary layers along the length of the samples, as well as the pressure distribution as a function of both geometry and Reynolds number. We calculate both viscous and form drags and compare the local performance of textures on each side. We employ this analysis to discuss the effectiveness of riblets under various angles of attack to identify and propose localized modifications to texture placement for enhanced/similar drag responses at reduced surface coverage. Our findings can contribute to a more in-depth understanding of the relationship between the performance of riblets and different angles of attack and offer insights for optimizing the riblet coverage for a variety of aerial and marine applications.   

Modulation of Turbulent Channel Flow by Bristling Shark Denticles

Shark denticles possess a flexible base anchored in the dermis, enabling pivoting and scale bristling. Direct numerical simulations were conducted to investigate the behavior of channel flows over an array of shark denticle replicas. The simulations were performed at $Re_b = U_b H/ν = 8000$. Without the presence of the denticles, the friction Reynolds number was around $Re_τ = 450$. The shark denticle replicas that had a height of 0.098$H$ are represented by an immersed boundary method. The array was assigned a sinusoidal variation in the angle of inclination, ranging from 0 to 15 and 30 degrees, mimicking the bristling motion. The bristling period was 4.0$H/U_b$, comparable to the bursting time scale of the boundary layer at this Reynolds number. Comparisons were made with the non-bristled denticles to explore the effects of the motion. The bristling significantly increases the total drag. The equivalent sand grain roughness height of the denticle increases from $k_s^+$ = 158 (no-bristling) to $466$ (30-degree bristling). The mean shear over the crown of the denticle is increased (decreased) when the denticle is bristling up (back down). Streamwise-elongated vortices formed between the ridges on the crown are periodically shed off by the bristling and alter the turbulence in the near-wall region.