Wake vortex mechanisms behind cylinders with and without porous coating

Abstract

An experimental analysis of cylinders with different porous coating configurations was performed to understand how these affect near-wake vortex formation mechanics. Five coating configurations were produced using metallic foam to evaluate different vortex formation parameters, namely the vortex shedding frequency, base pressure, formation length, wake thickness, vortex strength, and vorticity ratio. Existing scaling rules proved unable to describe all the vortex shedding patterns observed with less than six parameters, thus a new scaling was produced. By considering that immediately after detaching from a surface vortical structures will behave similarly unless disturbed, it was possible to reduce the scaling to wake velocity components alone. Results showed the new scaling to collapse the description of the near-wake from the initial six parameters to only four: three vortex relative speeds and vortex shedding frequency. Vorticity losses were shown to correlate with the ratio between vortex rotational speed and detaching shear layer speed with good agreement across all tested configurations. This validates the description developed of the vorticity losses in the near-wake, based on previous work describing the existence of two fluctuating wake length scales: formation and diffusion lengths. In order to study the effect of porous coatings on the coupling between drag and vortex shedding frequency, these were measured when sandpaper coatings and splitter plates were applied instead of metallic foams. Results showed the vortex shedding to be decoupled from drag unlike what had been observed for the metal foam coatings. Similar tests were performed on coating configurations with porous coatings removed, with results illustrating the combined effect of porous coatings over base pressure and formation length. The well-established capability of porous coatings to decrease the unstable wake oscillations largely responsible for aerodynamic noise is likely to be related to the suppression of the diffusion length fluctuations.