Bulletin of the American Physical Society
71st Annual Meeting of the APS Division of Fluid Dynamics
Volume 63, Number 13
Sunday–Tuesday, November 18–20, 2018; Atlanta, Georgia
Session F10: Multiphase Flows: Turbulence I |
Hide Abstracts |
Chair: Krishnan Mahesh, University of Minnesota, Twin Cities Room: Georgia World Congress Center B215 |
Monday, November 19, 2018 8:00AM - 8:13AM |
F10.00001: Physics-informed machine learning approach for sub-grid scale modelling in LES Aditya Karnik, Indranil Pan, Lachlan R Mason, Omar K Matar Industrial multiphase flows typically involve turbulent structures on either side of a fluid interface. Modelling efforts are computationally expensive, as large-eddy simulations (LES) require non-trivial near-interface treatments. In this study, we use a physics-informed machine learning (ML) approach for reconstructing LES sub-grid viscosity models. The ML model, trained on a DNS database of low Reynolds number wavy stratified flows, is then evaluated in out-of-sample Reynolds number settings. The ML algorithm can advantageously capture generic small-scale features from computationally expensive flows and can be coupled within each iteration of the physics based flow solver to reduce computational time. We quantify the saving in computational time versus accuracy vis-à-vis traditional dynamic LES models. The study demonstrates how coupling transparent physics based models with data-driven black box machine learning models leverages the best of both methodologies and gives computationally tractable and reliable solutions. |
Monday, November 19, 2018 8:13AM - 8:26AM |
F10.00002: Decomposing the wavelet spectrum of droplet-laden isotropic turbulence Andreas Freund, Antonio Ferrante The classical spectrum of turbulence kinetic energy poses a challenge when used to analyze finite-size droplet-laden flows because the Fourier transform used in its definition does not lend itself to a natural decomposition of the domain into droplet and carrier parts. We recall an alternative definition of the energy spectrum that uses the wavelet transform and apply it to the DNS data of droplet-laden decaying isotropic turbulence of Dodd & Ferrante (J. Fluid Mech. 806 (2016), 356–412). We also compute each term of the wavelet-spectrum evolution equation for two-phase flow. We then introduce a new method for decomposing the spectrum that takes advantage of the wavelet transform's preservation of spatial information, which allows us to isolate the effect of the droplets on the carrier fluid. Lastly, we discuss the results of our decomposition analysis. |
Monday, November 19, 2018 8:26AM - 8:39AM |
F10.00003: Local flow topology in droplet-laden homogeneous isotropic turbulence Michael Dodd, Lluis Jofre A recent DNS study [J. Fluid Mech. 806 (2016) 356-412] showed that the introduction of droplets into homogeneous isotropic turbulence (HIT) increased the decay rate of turbulence kinetic energy due to enhanced dissipation near the droplet interface. Further analysis of this DNS dataset has shown that, over a range of droplet Weber numbers, density ratios, and viscosity ratios, the dissipation rate is about four times larger in a sublayer containing the droplet interface than it is in the carrier flow. To better understand the physical mechanisms leading to this increase, we characterize the local flow topology near the interface by computing the invariants of the velocity-gradient, rate-of-strain, and rate-of-rotation tensors in the carrier and droplet fluid. By conditioning their joint probability density functions on distance from the interface, the DNS results show that away from the interface the local flow topology resembles canonical HIT, but as the droplet interface is approached, the flow features increasingly deviate from HIT. |
Monday, November 19, 2018 8:39AM - 8:52AM |
F10.00004: Droplets and bubbles in homogeneous shear turbulence Marco Edoardo Rosti, Zhouyang Ge, Suhas S Jain, Michael Dodd, Luca Brandt The understanding of turbulent two-phase flows with bubbles and/or droplets is important in many natural and industrial processes, e.g. rain formation, liquid-liquid emulsion, spray cooling and spray atomization in combustors. In these flows the turbulence is altered by the droplet feedback on the surrounding fluid and by droplet-droplet interactions. We perform direct numerical simulations of a homogeneous shear turbulent flow with finite size droplets simulated by a volume of fluid method. Using a number of post-processing techniques, we will discuss the turbulence modulation in terms of statistics and flow structures. The turbulent flow is affected by the dispersed phase which damps its fluctuations; the multiphase flow reaches a statistically steady state where the turbulent production is balanced by the dissipation, and the number of droplets is on average constant. By studying the turbulent kinetic energy balance we show that the droplets are a sink of turbulent kinetic energy with increased dissipation rate density and reduced production. Finally, the merging and breakup processes are critically evaluated and compared with the theory by Hinze. |
Monday, November 19, 2018 8:52AM - 9:05AM |
F10.00005: Turbulence suppression in emulsions Siddhartha Mukherjee, Orest Shardt, Arman Safdari, Mohammad Pourtousi, Harry E.A. Van den Akker Emulsification involves the mixing of immiscible liquids under flow conditions that are often turbulent, which generates a dense droplet suspension. An accurate description of this system comprises the dynamics of deforming interfaces, allowing for coalescence and breakup of droplets in the possible presence of surfactants that can alter interfacial dynamics. Further, a range of length and time scales of turbulence should be resolved. Simulating an emulsion in a periodic box by means of the lattice-Boltzmann method, we consider different volume fractions (φ) of the dispersed fluid while incorporating a low wavenumber forcing to generate homogeneous, isotropic turbulence in the continuous phase. We find that increasing φ significantly suppresses turbulence at high wavenumbers. The continuous phase becomes turbulent at low φ values (below 20%), and droplets undergo coalescence and breakup depending on the local flow dynamics. At higher φ (around 50%), the two liquids form a complex, entangled structure of essentially connected regions along with many small satellite droplets. Here fine scale flow features are not generated as the turbulence cascade is hindered by the connected interfaces. |
Monday, November 19, 2018 9:05AM - 9:18AM |
F10.00006: Catastrophic phase inversion in mayonnaise Taylor-Couette turbulence Dennis Bakhuis, Rodrigo Ezeta Aparicio, Pim Adriaan Bullee, Raymond H. J. Kip, Sander Huisman, Alvaro G. Marin, Detlef Lohse, Chao Sun Emulsions are commonly found in nature and heavily used in industry. While the properties of the individual fluids are known, the mixture of the two immiscibles can show very non-intuitive properties, especially at high Reynolds numbers. In this experimental study, we vary the oil volume fraction of a oil-water emulsion in a Taylor-Couette geometry at typical Reynolds numbers of 10^{6} and measure the torque required to keep the inner cylinder rotating at a constant velocity. When an oil is used with a viscosity larger than water, the apparent viscosity of the emulsions can be described by a laminar model. For an oil which has a viscosity similar to that of water, we witness a catastrophic phase inversion: when the fluid changes from oil droplets in water to water droplets in oil (or vice- versa), the flow morphology changes dramatically, resulting in an almost instantaneous jump in drag. |
Monday, November 19, 2018 9:18AM - 9:31AM |
F10.00007: Simulations of Planar Turbulent Jet Impacting Liquid Coatings Wojciech Aniszewski, Stephane Popinet, Stephane Zaleski In this work, we present a detailed example of numerical study of a planar jet impact onto the liquid film in context of metal coating. Liquid metal is drawn from a reservoir onto a retracting sheet, forming a coat. Planar air jets are then used to control the coat thickness. Real industrial configurations reach Reynolds numbers from tens of thousands to millions (depending on reference scale), the flow is also characterized by significant density ratios (air/liquid metal). The simulations presented in this work have been performed using Basilisk, a grid-adapting, strongly optimized code created by S. Popinet. Restricted variant of local adaptive mesh adaptation allows Basilisk to obtain arbitrary precision in relevant regions such as impact zone , while coarse grid is applied elsewhere to either save computational power or dampen turbulence far from regions of interest. Momentum-conserving code variants are used ensuring code stability. With this, we are able to present both two- and three-dimensional instantaneous results including interface geometry or film thickness even for realistic, industrial configurations. |
Monday, November 19, 2018 9:31AM - 9:44AM |
F10.00008: DNS of flow over realistically rough superhydrophobic surfaces Karim Alame, Krishnan Mahesh Direct numerical simulations are performed for two wall-bounded flow configurations: laminar Couette flow at Re= 740 and turbulent channel flow at Reτ= 180, where τ is the shear stress at the wall. The top wall is smooth and the bottom wall is a realistically rough superhydrophobic surface (SHS), generated from a three-dimensional surface profile measurement. The air-water interface, which is assumed to be flat, is simulated using the volume-of-fluid (VOF) approach. The laminar Couette flow is studied with varying interface heights h to understand the effect on slip and drag reduction (DR). A power law linear regression fit is used to obtain an effective slip as a function of gas fraction For the turbulent channel flow, statistics of the flow field are compared to that of a smooth wall. The fully wetted roughness increases the peak value in turbulent intensities, whereas the presence of the interface suppresses them, as evident from vorticity and turbulent kinetic energy. Overall, there exists a competing effect between the interface and the asperities, where the interface suppresses turbulence whereas the asperities enhance them. |
Monday, November 19, 2018 9:44AM - 9:57AM |
F10.00009: Settling dynamics of inertial disks in turbulence Luis Blay Esteban, John Shrimpton, Bharathram Ganapathisubramani Sedimentation of solid particles is encountered in engineering and environmental flows, yet the dynamics of finite-size particles in homogeneous anisotropic turbulence is still not fully understood. We perform experiments of disks falling in quiescent flow under different regimes, and compare their dynamics with the results from sedimentation under turbulent environments. We use two facing random jet arrays (RJA) of water pumps to generate turbulence with negligible mean flow and shear over a volume that is much larger than the initial characteristic turbulent large scale of the flow. The Reynolds number for forced turbulence is Re_{λ}≈580 and the axial-to-radial ratio of the rms velocity fluctuations is 1.22. The turbulence decay is investigated by monitoring velocity fluctuations, dissipation and turbulent length scales over time and disks are released at different decay times. The solid to fluid density ratio range is 2.7 to 7.5 and the particle diameter is of the same order of magnitude as the Taylor microscale. Turbulence has a severe influence on the settling dynamics of disks, modifying the mean velocity and the particle lateral dispersion; and introducing fast and slow events on the particle descent that are captured in the frequency content of the velocity fluctuations. |
Monday, November 19, 2018 9:57AM - 10:10AM |
F10.00010: Experimental Data for Solid-Liquid Flows at Intermediate and High Stokes Numbers Sarah E Mena, Jennifer Sinclair Curtis The present investigation details experimental data for solid-liquid turbulent flows in the intermediate and large Stokes numbers. The experiments varied the particle size from 0.5 mm to 5 mm, at concentrations up to 2% V/V and Reynolds numbers from 200,000 to 350,000. The data include direct measurements for the three velocity components for each of the phases using Laser Doppler Velocimetry and Phase Doppler Anemometry (LDV/PDA), pressure drops, and solid volume fractions measured by direct sampling. Reynolds stresses, granular temperature, and turbulence modulation calculations are also presented, along with qualitative observations of the radial distribution of the particles. In general, the mean liquid velocity profiles in the presence of the particles follow the single-phase profiles, while the mean solid profiles exhibit slip with a maximum relative velocity at the center. Turbulence augmentation was observed for all the experiments. In addition, the velocity fluctuations for the solid and the liquid are reduced with increasing Reynolds numbers for all particle sizes and concentrations. |
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. |
© 2023 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