Homogeneous dilute turbulent particle-laden water flows.

An experimental and theoretical study of a homogeneous particle-laden turbulent flow was conducted, involving glass particles falling in a stagnant (in the mean) water bath. The experiments included various loadings of nearly monodisperse particles having mean diameters of 0.5, 1.0 and 2.0 mm. Measurements included liquid and particle velocities, and the temporal spectra and spatial correlations of liquid-phase velocity fluctuations. The motion of isolated particles falling in stagnant water was also calibrated using shadowgraph movies. An approximate analysis for liquid-phase properties was developed, based on linear superposition of randomly-arriving particle velocity fields. In addition, stochastic simulations, solving the Lagrangian equations of motion of the particles, were used to predict particle properties. For present test conditions, liquid velocity fluctuations correlated solely as a function of the rate of dissipation of kinetic energy in the liquid. The correlation coefficients and temporal spectra were also independent of particle loading. The flows were highly anisotropic with streamwise velocity fluctuations being roughly twice the cross-stream velocity fluctuations. The temporal spectra indicated a large range of frequencies even when particle Reynolds numbers were low due to the fact that both mean and fluctuating velocities in the particle wakes contribute to the spectra because particle arrivals are random. The analysis predicted many of the features of the flows reasonably well but must be improved to obtain satisfactory predictions of the temporal and spatial scales of the flow. In addition to effects of turbulent dispersion, the particles exhibited self-induced lateral motion due to eddy-shedding and irregularities of shape--particularly the larger particles. Streamwise particle velocity fluctuations were large because particle size variations induced variations of terminal velocities. Cross-stream velocity fluctuations were dominated by effects of turbulent dispersion for the smallest particles with effects of self-induced lateral motion becoming more important with increasing particle size. After accounting for effects of particle size variations and self-induced motion, the stochastic simulation predicted particle properties reasonably well.PhDAerospace engineeringApplied SciencesMechanical engineeringUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/128386/2/9001695.pd