Flow alteration and drag reduction by riblets in a turbulent boundary layer

Detailed measurements with a miniature single-sensor hot-wire probe of the streamwise velocity field within a riblet groove of dimensionless size h+ ~ 14 and s+ ~ 28 (based on the riblet surface average friction velocity) have been made. The wall shear stress, when integrated over the riblet surface, yields about 4% drag reduction com- pared with a smooth surface with the same projected area. This is largely due to the greatly diminished wall shear stress near the bottom of the riblet valley. Four-sensor hot-wire probe measurements reveal that riblets sig- nificantly reduce the vertical flux of streamwise momentum within the riblet valley. HAT, under certain conditions, configuring the boundary be- neath a turbulent flow into small grooves aligned with the mean flow direction, called riblets, will reduce viscous drag has been well documented for over a decade.1'2 However, very few de- tailed measurements of the flowfield within and just above these riblet grooves have been made for the obvious reason that they are very difficult. In most laboratory air boundary layers the physical dimensions of such riblets are considerably less than 1 mm, and the velocities to be measured are extremely small. Vukoslavcevic et al.3 have, however, measured properties of the streamwise velocity field along the vertical planes of symmetry above the valley and above the peak of a triangular riblet surface in an air boundary layer at RQ ~ 1.0 X 10 3. These measurements showed that the mean ve- locity gradient at the valley surface is much lower than at a compar- ison smooth flat plate surface for flows with the same freestream velocity. The mean velocity gradient over the peak was somewhat greater than that over the smooth surface. Vukoslavcevic et al. 3 also determined that the turbulence intensity of the streamwise ve- locity fluctuations was greatly reduced deep in the riblet valley. The skewness and flatness factors indicated that the turbulence only rarely penetrates deep into the valley from the layers above. These measurements, however, were not made normal to or at a sufficient number of stations along the riblet wall surface to esti- mate whether and, if so, how drag reduction was achieved. Further- more, no measurements of the vertical velocity fluctuations v were made, and so information about the vertical flux of streamwise mo- mentum could only be inferred indirectly. Chu et al. 4 and Choi et al. 5 have carried out direct numerical simulations of flow over riblet walls in channel flows. Most of the conclusions made from the laboratory experiment reported herein are qualitatively consistent with their simulation results, although there are some quantitative differences in detail, as will be dis- cussed later when comparisons are made.