Title:A modal analysis of segregation of inertial particles in turbulence

Abstract:An asymptotic solution is derived for prediction of segregation of heavy inertial particles in spatially and temporally varying flows. The general solution, which classifies as the Fredholm integral equation of the second kind, is a function of the spectrum of the second invariant of the velocity gradient tensor of the underlying flow. We introduce a one-dimensional canonical flow oscillating at a single frequency to investigate the behavior of the Lyapunov exponent under a wide range of flow conditions. We show that in a straining flow the Lyapunov exponent can be positive or negative (either dispersion and segregation can occur), whereas it is always positive in a rotating regime. The minimum requirement for trajectory crossing is also predicted. The trajectory crossing never occurs in a rotating regime, whereas it occurs in a straining regime if the Lyapunov exponent is positive. Our analysis shows the Lyapunov exponent, normalized by the particle relaxation time, is minus one half at the minimum and grows linearly at small and nonlinearly with a power of one half at large oscillation amplitudes. The direct numerical simulations confirm these predictions. The extension of our analysis to multimodal excitation is discussed and applied to a two-dimensional synthetic straining and a three-dimensional isotropic forced turbulent flow. We show the results are analogous to that of the one-dimensional case, underscoring the relevance of our canonical setting and the fact that particle segregation arises from the general rather than particular solution of the Stokes drag equation. In comparison to two preexisting models, our model has a more general form and reproduces them under further simplifying assumptions. In contrast to these models that are valid only at small Stokes numbers, our model is valid at small and large Stokes numbers.