Microbubble drag reduction phenomenon study in a channel flow

Abstract

An experimental study on drag reduction by injection of microbubbles was performed in the upper wall of a rectangular channel at Re = 5128. Particle Image Velocimetry measurement technique (PIV) was used to obtain instantaneous velocity fields in the x-y plane. Microbubbles, with an average diameter of 30??m, were produced by electrolysis using platinum wires with a diameter of 76 ??m. They were injected in the buffer layer producing several different values of local void fraction. A maximum drag reduction of 38.45% was attained with a local void fraction of 4.8 %. The pressure drop in the test station was measured by a reluctance pressure transducer. Several parameters such as velocity profile, turbulent intensities, skewness, flatness, joint probability density function (JPDF), enstrophy, one and two-dimensional energy spectra were evaluated. The results indicate that microbubbles reduced the intermittency of the streamwise fluctuating component in the region near the wall. At the same time they destroy or reduce the vortical structures regions (high shear zones) close to the wall. They also redistribute the energy among different eddy sizes. An energy shift from larger wavenumbers to lower wavenumbers is observed in the near wall region (buffer layer). However, outside this region, the opposite trend takes place. The JPDF results indicate that there is a decrease in the correlation between the streamwise and the normal fluctuating velocities, resulting in a reduction of the Reynolds stresses. The results of this study indicate that pursuing drag reduction by injection of microbubbles in the buffer layer could result in great saving of energy and money. The high wavenumber region of the one dimensional wavenumber spectra was evaluated from PIV spatial information, where the maximum wavenumber depends on the streamwise length (for streamwise wavenumber) of the recorded image and the minimum wavenumber depends on the distance between vectors. On the other hand, the low wavenumber region was calculated from the PIV temporal information by assuming Taylor??s frozen hypothesis. This new approach allows obtaining the energy distribution of a wider wavenumber region.