Gas-particle interaction in turbulent channel flow

A fully developed, turbulent flow of air in a rectangular vertical channel was investigated. The Reynolds number, based on the mean centerline gas velocity and on the channel half-width, was 4540. The flow was laden with particles (solid glass beads) at 10% mass fraction. A small amount of tracers (fine glycerin droplets) was added to the flow as a means for gas velocity measurements. To discriminate signals from solids and tracers, we developed a novel two-phase PIV method. Then, instantaneous gas and particle velocity fields were acquired simultaneously in homogeneous planes parallel to the channel sidewall.; Five sets of sized solid particles were used, so that the average particle response time varied from 50 to 3190 friction time scales. Eight distances from the wall were considered, ranging from 16 to 238 wall units. At each location, single-point velocity statistics were computed for development of Eulerian (two-fluid) models. We also computed gas and particle two-point planar velocity correlations, suitable for testing two-point PDF models.; The RMS particle slip velocity, if scaled as a Reynolds number, appears to fit a similarity law dependent on particle Stokes number based on the Kolmogorov time scale over the range 4 < St_K < 1000. Over this range, RMS particle Reynolds number varied between 1.5 and 32. The drift velocity data show that low-inertia particles tend to accumulate in low-speed regions near the wall. Single-point gas-particle velocity covariances increase with decreasing particle inertia, and for low-inertia particles, they increase with the wall normal distance as well. All covariances are positive up to St_K = 1160, the largest value tested, and streamwise covariances are typically higher than spanwise.; Two-point gas-gas velocity correlations are consistent with the presence of hairpin vortex structures, known from many previous single-phase studies. Gas-particle correlations show that particles of all sizes respond to these structures to some degree, but often with a significant lag. Gas-particle and particle-particle correlation values both increase with decreasing particle inertia. In all cases, the size of correlation patterns matches the dominant fluid structure size. Positive particle-particle correlations imply reduced inter-particle collision rates compared to kinetic theory models.