Bulletin of the American Physical Society
66th Annual Meeting of the APS Division of Fluid Dynamics
Volume 58, Number 18
Sunday–Tuesday, November 24–26, 2013; Pittsburgh, Pennsylvania
Session M32: Particle-Laden Flows VI: Direct Simulation and Turbulence Modulation |
Hide Abstracts |
Chair: Said Elghobashi, University of California, Irvine Room: 403 |
Tuesday, November 26, 2013 8:00AM - 8:13AM |
M32.00001: Evaluating multiphase turbulence statistics using mesoscale DNS of gravity-driven particle-laden flows Rodney Fox, Jesse Capecelatro, Ollivier Desjardins Flow instabilities encountered in fluid-particle flows subjected to a body force (i.e., gravity) can lead to mesoscale structures that control the underlying fluid turbulence. The wide range of length and time scales associated with such flows pose great challenges in understanding and predicting key features. In our recent work, the exact Reynolds-average (RA) equations for monodisperse collisional particles in a constant-density fluid were derived. The turbulence model solves for the RA particle volume fraction, the phase-average (PA) particle velocity, the PA granular temperature, and the PA particle turbulent kinetic energy. Unlike most previous derivations, a clear distinction is made between the PA granular temperature, which appears in the kinetic theory constitutive relations, and the particle-phase turbulent kinetic energy, which appears in the turbulent transport coefficients. Fully coupled Eulerian-Lagrangian simulations are used to evaluate the unclosed terms that arise in the RA equations. A two-step filter is employed during the interphase exchange process, providing a separation of length scales between the microscale and mesoscale structures. The same filtering process is used to evaluate turbulence statistics and modeling constants appearing in the RA model. [Preview Abstract] |
Tuesday, November 26, 2013 8:13AM - 8:26AM |
M32.00002: A study of turbulence modulation by particle clusters in dilute and moderately-dilute channel flows using mesoscale DNS Jesse Capecelatro, Olivier Desjardins This work investigates fluid-particle interactions in turbulent channel flows using highly-resolved Euler-Lagrange simulations. In the dilute regime, particle dynamics are mostly controlled by vortical structures in the flow, and wakes past individual particles can modify the underlying fluid turbulence at the particle scale. At moderate concentrations and mass loadings, flow instabilities may lead to mesoscale structures (i.e., clusters) that control underlying fluid turbulence. A Re=13,500 channel flow is studied in both regimes. It is shown in this work that the fluid turbulence departs significantly from the initially fully-developed turbulent flow when subjected to mean particle volume fractions of 1\%, where all unsteady features are generated by the cluster dynamics. To study the effect of gravity on clustering dynamics, simulations are conducted with gravity aligned in the mean flow direction, as well as gravity opposing the mean flow direction (i.e., a riser configuration). Velocity fluctuations and energy spectra are computed for each case, along with higher order Lagrangian statistics including collision frequency, radial distribution function, and particle number density. [Preview Abstract] |
Tuesday, November 26, 2013 8:26AM - 8:39AM |
M32.00003: Modeling low-order structure functions for inertial particles in isotropic turbulence Andrew Bragg, Lance Collins In this talk we will consider three models for the second order structure function of inertial particle pairs in isotropic turbulence, one by Zaichik \emph{et al.} (New. J. Phys. 11:103018, 2009), the second by Pan \emph{et al.} (J. Fluid. Mech. 661:73, 2010) and the third by Gustavsson \emph{et al.} (Phys. Rev. E. 84:045304, 2011). We find that in general they describe the structure functions in qualitatively similar ways, capturing the influence of the nonlocal dynamics on the formation of caustics and non-smooth scaling behavior in the dissipation range. We then compare the predictions with DNS data and find that although they capture the qualitative behavior of the data consistently, they differ with each other quantitatively, with the theory by Pan \emph{et al.} yielding the closest agreement with the DNS. Finally, we show that a new backward in time dispersion theory we have derived makes improvements to the predictions from the Pan \emph{et al.} theory by improving upon a key closure approximation made in its construction. [Preview Abstract] |
Tuesday, November 26, 2013 8:39AM - 8:52AM |
M32.00004: Near-wall, particle-laden turbulent transport David Richter, Peter Sullivan We use direct numerical simulation coupled with a Lagrangian point-particle formulation to study turbulent planar Couette flow at friction Reynolds numbers ranging between 125 and 900. Modifications to wall-normal scalar and momentum transport are investigated as a function of the size and concentration of the dispersed phase. Furthermore, the dispersed phase effects are examined as the Reynolds number of the flow is increased. In all cases particle phase is observed to weaken the structures responsible for near-wall transport (e.g. hairpins, quasi-streamwise vortices, etc.); an effect which becomes increasingly pronounced as the Reynolds number increases. Physical explanations for this behavior will be presented. [Preview Abstract] |
Tuesday, November 26, 2013 8:52AM - 9:05AM |
M32.00005: Particle-Resolved Direct Numerical Simulation of a Particle-Laden Mixing Layer Mohammad Mehrabadi, Sudheer Tenneti, Shankar Subramaniam The stability of a homogeneous gas-solid suspension has been investigated in the context of kinetic theory (Koch, Phys. Fluids, 1990) and the averaged two-fluid equations (Glasser et al, PRL, 1998) by considering perturbation of the number density. Koch's analysis points to the dependence of average drag on average volume fraction as the mechanism for the development of instabilities in the number density. However, the physical origins of instabilities in the number density have not been firmly established through microscale simulations at the scale of individual particles. In this study, particle-resolved direct numerical simulation (PR-DNS) is used to ascertain the exact physical origins of the growth of number density instabilities in a particle-laden mixing layer. Self-similarity of the temporally evolving number density profile, and the diffusive/convective nature of the instability is examined to ascertain the role of granular temperature in instability growth. The growth of streamwise and cross-stream structures in the particle field are analyzed to draw analogies with the classical Rayleigh-Taylor and Kelvin-Helmholtz instability mechanisms. [Preview Abstract] |
Tuesday, November 26, 2013 9:05AM - 9:18AM |
M32.00006: Direct numerical simulation of forward- and backward-in-time relative dispersion of inertial particles in high-Reynolds-number ($R_\lambda \approx 580$) turbulence Peter Ireland, Andrew Bragg, Lance Collins Turbulence-induced water droplet coalescence is considered to be an important factor in the onset of precipitation in warm cumulus clouds. Theory (Bragg and Collins 2013) shows that the collision kernel for suspended droplets in turbulence is fundamentally related to their backward-in-time relative dispersion, which has yet to be systematically investigated. Using direct numerical simulations on a $2048^3$ lattice with $R_\lambda \approx 580$, we find that inertial particles, like fluid particles, separate more quickly backward than forward in time. However, the degree of asymmetry in the dispersion is significantly greater for inertial particles than for fluid particles. We present new parameterizations for both short and long time relative dispersion and discuss the physical mechanisms leading to the strong asymmetry in the dispersion processes. The results from this work will be used to the improve the theoretical model for particle relative velocities developed by Pan and Padoan (2010), enabling more accurate predictions of collisional droplet growth rates. [Preview Abstract] |
Tuesday, November 26, 2013 9:18AM - 9:31AM |
M32.00007: DNS of fully-resolved droplet-laden decaying isotropic turbulence A. Ferrante, M. Dodd We investigated the effects of finite-size droplets on decaying isotropic turbulence by performing direct numerical simulation (DNS). We performed DNS using our new pressure-correction/volume-of-fluid method that is mass-conservative and second-order accurate. The simulations were performed at Re$_{\lambda 0}=75$ on a $1024^3$ grid such to resolve each droplet with 32 grid points per diameter. We fully resolve all the relevant scales of turbulence around thousands of freely-moving droplets of Taylor length-scale size as well as the fluid motion inside the droplets. We will discuss the effects of the droplets on the temporal development of turbulence kinetic energy and its dissipation rate. Also, we will present the effects on turbulence of the droplet Weber number and of the density ratio between the droplet and the surrounding fluid. [Preview Abstract] |
Tuesday, November 26, 2013 9:31AM - 9:44AM |
M32.00008: DNS of droplet motion in a turbulent flow Michele Rosso, S. Elghobashi The objective of our research is to study the multi-way interactions between turbulence and vaporizing liquid droplets by performing direct numerical simulations (DNS). The freely-moving droplets are fully resolved in 3D space and time and all the relevant scales of the turbulent motion are simultaneously resolved down to the smallest length- and time-scales. Our DNS solve the unsteady three-dimensional Navier-Stokes and continuity equations throughout the whole computational domain, including the interior of the liquid droplets. The droplet surface motion and deformation are captured accurately by using the Level Set method. The pressure jump condition, density and viscosity discontinuities across the interface as well as surface tension are accounted for. Here, we present only the results of the first stage of our research which considers the effects of turbulence on the shape change of an initially spherical liquid droplet, at density ratio (of liquid to carrier fluid) of 1000, moving in isotropic turbulent flow. We validate our results via comparison with available expe [Preview Abstract] |
Tuesday, November 26, 2013 9:44AM - 9:57AM |
M32.00009: Effect of flow straining on particle accelerations and distribution Armann Gylfason, Chung-min Lee, Lahcen Bouhlali, Federico Toschi We explore effects of large scale straining on Lagrangian properties of particles in turbulence. We perform direct numerical simulations of strained turbulence laden with passive and inertial particles of varied inertia, as well as perform particle tracking velocimetry measurement in the same geometry. From both of these studies we analyze particle acceleration statistics to investigate the effect of large scale flow distortion due to straining, resulting in anisotropy that ranges from the large scale down to the inertial range and the dissipative range. A secondary objective is to understand the effects of weak straining on the distribution of particles in the fluid, by examining the evolution of spatial distribution statistics and investigating particle dispersion from simple sources. Particular attention is given to the dependence of these statistics on Reynolds number and rate of strain, in combination with particle inertia. [Preview Abstract] |
Tuesday, November 26, 2013 9:57AM - 10:10AM |
M32.00010: Segregation of heavy particles by gravitational force Yongnam Park, Changhoon Lee The effects of gravitational force on the segregation of heavy particles are investigated in forced isotropic turbulence using direct numerical simulation. The mechanism of preferential concentration of heavy particle has been known to be strongly related with the vortical structures of background turbulence when gravity is not considered. The degree of preferential concentration is maximized when the characteristic time scale of a particle is comparable with the Kolmogorov time scale. In this study, we discover that strong gravity causes a different kind of preferential concentration for high Stokes number particles. Such phenomenon does not seem to be related with the vortical structures. In order to provide a plausible explanation, we investigate the statistics of horizontal divergence at the location of heavy particles. Particles tend to be segregated more as the particles experience longer time of negative divergence, meaning converging motion in the horizontal plane. The ratio of mean duration time of negative divergence to that of positive divergence increases with gravitational force for the high Stokes number particles. More detailed statistics and relevant explanation will be presented in the meeting. [Preview Abstract] |
Follow Us |
Engage
Become an APS Member |
My APS
Renew Membership |
Information for |
About APSThe American Physical Society (APS) is a non-profit membership organization working to advance the knowledge of physics. |
© 2022 American Physical Society
| All rights reserved | Terms of Use
| Contact Us
Headquarters
1 Physics Ellipse, College Park, MD 20740-3844
(301) 209-3200
Editorial Office
1 Research Road, Ridge, NY 11961-2701
(631) 591-4000
Office of Public Affairs
529 14th St NW, Suite 1050, Washington, D.C. 20045-2001
(202) 662-8700