Rotating Discs for Turbulent Drag Reduction: An Experimental Study

Abstract

The reduction of turbulent skin-friction drag has the potential to generate numerous societal benefits. In large-scale industrial applications, the largest share of energy consumption goes into overcoming the friction drag produced by turbulent boundary layers formed on the outside wall. Estimates assess that over 50% of the total fuel burn is due to overcoming friction drag, highlighting how reducing it would result in large environment and economic savings. In this work, an active technique for friction drag reduction in a turbulent boundary layer is studied in TU Delfts windtunnel laboratory. The flow over a stationary wall is modified by the steady rotation of flush-mounted discs - aiming to mimic a motion of wall-oscillation perpendicular to the flow, a well-documented turbulent drag reduction technique. The effect of the disc motion on turbulent drag is investigated at friction Reynolds numbers of 912 and 1483. The chosen rotational velocity of the discs, the disc diameter and disc spacing represent the optima defined in literature, and a further parameter exploration was conducted around these optima. To assess the performance of the disc array, high resolution planar PIV is employed to measure skin-friction drag reductions at the wall, as well as to quantify second order statistics and other phenomena within the turbulent boundary layer. The results reveal drag reductions in the order of 50%, with Reynolds stresses and peaks of turbulent kinetic energy production decreasing by similar amounts. Analysis of turbulent velocity fluctuations reveals that the prevalence of turbulence producing events is significantly decreased close to the wall, and two-point correlations of vorticity contours show that vortical events are remarkably hindered by the disc rotation. Further analysis of a wall-parallel plane reveals that the mean velocity increases over the discs and decreases in between. This suggest that the detrimental effects of such a disc array occur at spanwise locations between each disc row, in accordance with literature. By using an assumption of outer flow similarity, the performance of the array is then quantified as being 10% less than the local optimum in the disc center.