0
Research Papers: Flows in Complex Systems

Freeman Scholar Review: Passive and Active Skin-Friction Drag Reduction in Turbulent Boundary Layers

[+] Author and Article Information
Marc Perlin

Naval Architecture and Marine Engineering,
University of Michigan,
Ann Arbor, MI 48109

David R. Dowling

Mechanical Engineering,
University of Michigan,
Ann Arbor, MI 48109

Steven L. Ceccio

Naval Architecture and Marine Engineering,
Mechanical Engineering,
University of Michigan,
Ann Arbor, MI 48109

Contributed by the Fluids Engineering Division of ASME for publication in the JOURNAL OF FLUIDS ENGINEERING. Manuscript received April 1, 2015; final manuscript received March 5, 2016; published online June 2, 2016. Editor: Malcolm J. Andrews.

J. Fluids Eng 138(9), 091104 (Jun 02, 2016) (16 pages) Paper No: FE-15-1228; doi: 10.1115/1.4033295 History: Received April 01, 2015; Revised March 05, 2016

A variety of skin-friction drag reduction (FDR) methods for turbulent boundary layer (TBL) flows are reviewed. Both passive and active methods of drag reduction are discussed, along with a review of the fundamental processes responsible for friction drag and FDR. Particular emphasis is given to methods that are applicable to external hydrodynamic flows where additives are diluted by boundary layer entrainment. The methods reviewed include those based on engineered surfaces (riblets, large eddy breakup devices (LEBUs), and superhydrophobic surfaces (SHS)), those based on additives (polymer injection and gas injection), and those based on morphological alterations in the boundary layer flow (air layers and partial cavity formation). A common theme for all methods is their disruption of one or more of the underlying physical processes responsible for the production of skin-friction drag in a TBL. Opportunities and challenges for practical implementation of FDR techniques are also discussed.

Copyright © 2016 by ASME
Your Session has timed out. Please sign back in to continue.

References

Figures

Grahic Jump Location
Fig. 1

TBL geometry. Boundary layer development begins at x = 0, and the surface on which the boundary layer forms coincides with y = 0. The undisturbed flow at speed Ue is parallel to the surface on which the boundary layer forms, and δ(x) is the overall thickness of the boundary layer where the average velocity U(x,y) is less than Ue.

Grahic Jump Location
Fig. 2

TBL skin-friction coefficient versus Rex. The symbols with error bars are measurements from Ref. [12]. The solid curve is Eq. (2.3). The dashed lines are two classic skin-friction correlations for TBLs.

Grahic Jump Location
Fig. 3

Mean velocity profile of a smooth-flat-plate TBL plotted in log-linear coordinates with law-of-the-wall normalizations. The data are replotted from Ref. [12] and represent three Reynolds numbers. The extent of the various layers within thisTBL flow is indicated by vertical dashed lines. The log-layer-to-wake-region boundary is usually assumed to begin at y/δ ≈ 0.15–0.20 in TBLs. Overall, the data collapse well for the inner layer region, as expected, and the logarithmic layer extends for approximately two decades. The wake region shows differences between the Reynolds numbers because its similarity variable is y/δ, and δ/lν differs between the various Reynolds numbers.

Grahic Jump Location
Fig. 4

Mean velocity profile of a smooth-flat-plate TBL plotted using outer region scaling for the velocity defect Ue − U. The plotted data represent 12 different velocity profiles from the experiments reported in Refs. [12] and [19] covering the Reynolds number range 15,000 ≤ Reτ≤ 61,000. Here, the log law diverges from the measurements at y/δ ∼ 0.20. The difference seen between the log law and the data for y/δ > 0.2 is the wake component of the mean velocity profile.

Grahic Jump Location
Fig. 5

Schematic cartoon of flow structures in the buffer layer of near-wall turbulence. The low- and high-speed velocity streaks alternate and occur near the wall between the quasi-streamwise vortices with a nominal spanwise cycle distance of z+ ∼ 100. The nominal diameter of the quasi-streamwise vortices is 40 lν. The peak turbulence intensity and peak turbulence production in a TBL occur near y+ ∼ 12 and are both associated with the dynamic evolution of the depicted near-wall features.

Grahic Jump Location
Fig. 6

Roughness function, ΔB, as a function of inner region scaled equivalent sand-grain roughness height, ks+=ksu*/v. The solid curve is the correlation of Colebrook [27] for surfaces typical of commercial pipes. The long-dashed curve follows the sand-grain roughness results of Nikuradse [26]. The short-dashed curves provide approximate upper and lower bounds for experimental results from a variety of rough surfaces. Although the chosen normalizations produce consistent results below ks+ of unity and above ks+ of ∼20, this figure shows that ks alone is insufficient to describe the effects of wall roughness between these nominal limiting values.

Grahic Jump Location
Fig. 7

The cross-sectional geometry of V-, U-, and L-shaped riblets. The riblet channels are parallel to the streamwise direction of the flow, and the spanwise spacing is s, the height is h, and the area between the riblets is AG.

Grahic Jump Location
Fig. 8

% FDR as a function of s+ for a typical riblet. Adapted from Ref. [34].

Grahic Jump Location
Fig. 9

Schematic diagram of LEBUs employed in the study of Park et al. [47]

Grahic Jump Location
Fig. 10

A droplet on a smooth surface (a), on a textured surface in the Wenzel state where the liquid fill the texture on the surface (b), and in the Cassie–Baxter state (c), where gas pockets separate the liquid drop from the pores on the textured surface

Grahic Jump Location
Fig. 11

This slip length, β = H + l, for the flow over a surface with a gas film (a) or SHS in the Cassie–Baxter state (b). Adapted from Ref. [51].

Grahic Jump Location
Fig. 12

Plan and cross-sectional images of an unstructured SHS examined by Bidkar et al. [59], denoted as “sample 12”

Grahic Jump Location
Fig. 13

Friction drag coefficient versus Reynolds number from Ref. [59], where L is the length of the TBL from the virtual origin; the results from the SHS shown in Fig. 12 are denoted as sample 12; the “historical” is the average smooth-surface friction curve over many previous testing campaigns, and the “baseline” is the friction curve of the smooth sample from the tests of Bidkar et al. [59]. Sample 11 was not as effective and had a different surface coating on a similar roughness.

Grahic Jump Location
Fig. 14

Results from Ref. [73] showing the highest levels of DR seen in these experiments. %DR is presented as a function of downstream distance from the injection location. The different symbols refer to the three speeds with the open circles 6.65 ms−1, the filled circles 13.2 ms−1, and the open squares 19.9 ms−1. The MW was 8 M, the concentration of PEO was 4000 wppm, and the flux, q/qs, was 10.

Grahic Jump Location
Fig. 15

cM as a function of K for WSR-301 polymer. The filled dots, squares, and triangles are for freestream speeds of 6.65, 13.2, and 19.9 ms−1 from Ref. [73]. The open set of the same symbols are for concentrations of 100, 500, and 1000 wppm, and are from Ref. [77]. The plus and minus symbols are for an assumed value of 1000 wppm and represent the data of Vdovin and Smol'yakov [74,78], respectively. The results of Fontaine et al. [79] are given by the x symbols. The three superposed lines are for exponents of 0.20, 2.7, and 0.857 as shown, with the figure from Ref. [73].

Grahic Jump Location
Fig. 16

These figures (see Ref. [80]) demonstrate that the flow is fully rough (a), and that the roughness is of sand-grain type (b)

Grahic Jump Location
Fig. 17

Reproduced from Elbing et al. [80], their Fig. 9 shows three speeds (6.8 (), 13.5 (), and 20.1 () ms−1) with a direct comparison of the DR on the rough (solid symbols) and smooth (open symbols) surfaces as a function of distance downstream

Grahic Jump Location
Fig. 18

Graph of the ratio of skin-friction coefficients for the upstream injection case as a function of distance downstream on the plate. Each of the three nominal speeds, 6, 12, and 18 ms−1, is presented with four injection volumetric flow rates.

Grahic Jump Location
Fig. 19

%DR as a function of volumetric gas injection rate per unit span, q, at four downstream locations. The horizontal line at 80% represents the (arbitrary) threshold used to define ALDR.

Grahic Jump Location
Fig. 20

Air fluxes required for ALDR over rough and smooth surfaces [104], where q is the volume flux of injected gas per unit span beneath the bottom of a horizontal surface, and Ue is the freestream speed. Data for smooth and fully rough surfaces are shown. These data are compared to the approximate air fluxes used in the sea-trials reported by Hoang et al. [106] and Mizokami et al. [107].

Grahic Jump Location
Fig. 21

Profile of the gate and model in the LCC's test section. The origin of the coordinate system is at the base of the backward facing step (BFS). For clarity, the axes are shown shifted to the side and not at the true location of the origin. Inset: the cavity-terminating beach colored gray shown in detail. The upstream height of the cavity (i.e., the backward-facing step) is 0.18 m. Notice that the flap is articulated and was oscillated during the experiments.

Grahic Jump Location
Fig. 22

Nondimensional critical gas flux to establish and maintain cavities as a function of the Froude number for the two scales: the LCC and the mLCC

Tables

Errata

Discussions

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related eBook Content
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In