Inertial effects on two-particle relative dispersion in turbulent flows

TL;DR

This study experimentally investigates how inertial effects influence the relative dispersion of particle pairs in turbulent flows, revealing faster separation for inertial particles compared to tracers and highlighting limitations of existing models.

Contribution

The paper provides experimental data on inertial particle dispersion in turbulence and introduces a model that captures increased relative velocities but fails quantitatively due to flow sampling effects.

Findings

Heavier particles separate faster than tracers in turbulence.

Relative velocity between inertial particles is larger than between tracers.

Existing models do not fully capture the scale dependence due to flow sampling.

Abstract

We report experimental results on the relative motion of pairs of solid spheric particles with initial separations in the inertial range of fully developed turbulence in water. The particle densities were in the range of $1 \lessapprox \rho_{p}/\rho_{f} \lessapprox 8$, \textit{i.e.}, from neutrally buoyant to highly inertial; and their sizes were of the Kolmogorov scale. For all particles, we observed a Batchelor like regime, in which particles separated ballistically. Similar to the Batchelor regime for tracers, this regime was observed in the early stages of the relative separation for times $t \lessapprox 0.1 t_0$ with $t_0$ determined by the turbulence energy dissipation rate and the initial separation between particle pairs. In this time interval heavier particles separated faster than fluid tracers. The second order Eulerian velocity structure functions was found to increase with density. In other words, both observations show that the relative velocity between inertial particles was larger than that between tracers. Based on the widely used, simplified equation of motion for inertial point-particles, we derived a model that shows an increase in relative velocity between inertial particles. In its scale dependence, however, it disagrees quantitatively with the experimental results. This we attribute to the preferential sampling of the flow field by inertial particles, which is not captured by the model.