Bulletin of the American Physical Society
70th Annual Meeting of the APS Division of Fluid Dynamics
Volume 62, Number 14
Sunday–Tuesday, November 19–21, 2017; Denver, Colorado
Session L2: Multiphase Flows: Computational Methods ICFD Multiphase
|
Hide Abstracts |
Chair: Duan Zhang, Los Alamos National Laboratory Room: 402 |
Monday, November 20, 2017 4:05PM - 4:18PM |
L2.00001: Simulating compressible-incompressible two-phase flows Fabian Denner, Berend van Wachem Simulating compressible gas-liquid flows, e.g. air-water flows, presents considerable numerical issues and requires substantial computational resources, particularly because of the stiff equation of state for the liquid and the different Mach number regimes. Treating the liquid phase (low Mach number) as incompressible, yet concurrently considering the gas phase (high Mach number) as compressible, can improve the computational performance of such simulations significantly without sacrificing important physical mechanisms. A pressure-based algorithm for the simulation of two-phase flows is presented, in which a compressible and an incompressible fluid are separated by a sharp interface. The algorithm is based on a coupled finite-volume framework, discretised in conservative form, with a compressive VOF method to represent the interface. The bulk phases are coupled via a novel acoustically-conservative interface discretisation method that retains the acoustic properties of the compressible phase and does not require a Riemann solver. Representative test cases are presented to scrutinize the proposed algorithm, including the reflection of acoustic waves at the compressible-incompressible interface, shock-drop interaction and gas-liquid flows with surface tension. [Preview Abstract] |
Monday, November 20, 2017 4:18PM - 4:31PM |
L2.00002: All-Mach conservative scheme for transonic multiphase flows Michael Kuhn, Olivier Desjardins Transonic fluid flows pose significant challenges in computational simulations since the flow solver must be robust and accurate in subsonic, sonic, and supersonic regimes. Multiphase transonic flows amplify this difficulty even more, since the the Mach numbers in each phase can be dramatically different, like in the case of a liquid jet in supersonic crossflow. Typically, advection schemes for sonic and supersonic flows introduce significant numerical dissipation, which allows the solver to be robust in the presence of shocks. However, this numerical dissipation means that kinetic energy is not conserved as the Mach number goes to zero and that turbulence cannot be represented accurately in the lower Mach regions of the flow. By utilizing a dissipative Semi-Lagrangian advection scheme with a centered kinetic-energy-conserving scheme and introducing a framework for switching between the two, we present work toward an all-Mach, multiphase-ready solver that conserves mass, momentum, and energy at all Mach numbers and conserves kinetic energy in the low Mach limit. [Preview Abstract] |
Monday, November 20, 2017 4:31PM - 4:44PM |
L2.00003: A Parallel Geometry and Mesh Infrastructure for Explicit Phase Tracking in Multiphase Problems Fan Yang, Anirban Chandra, Yu Zhang, Ehsan Shams, Saurabh Tendulkar, Rocco Nastasia, Assad Oberai, Mark Shephard, Onkar Sahni Numerical simulations with explicit phase/interface tracking in a multiphase medium impact many applications. One such example is a combusting solid involving phase change. In these problems explicit tracking is crucial to accurately model and capture the interface physics, for example, discontinuous fields at the interface such as density or normal velocity. A necessary capability in an explicit approach is the evolution of the geometry and mesh during the simulation. In this talk, we will present an explicit approach that employs a combination of mesh motion and mesh modification on distributed/partitioned meshes. At the interface, a Lagrangian frame is employed on a discrete geometric description, while an arbitrary Lagrangian-Eulerian (ALE) frame is used elsewhere with arbitrary mesh motion. Mesh motion is based on the linear elasticity analogy that is applied until mesh deformation leads to undesirable cells, at which point local mesh modification is used to adapt the mesh. In addition, at the interface the structure and normal resolution of the highly anisotropic layered elements is adaptively maintained. We will demonstrate our approach for problems with large interface motions. Topological changes in the geometry (of any phase) will be considered in the future. [Preview Abstract] |
Monday, November 20, 2017 4:44PM - 4:57PM |
L2.00004: A Stabilized Finite Element Method for Compressible Phase Change Problems Yu Zhang, Fan Yang, Anirban Chandra, Ehsan Shams, Mark Shephard, Onkar Sahni, Assad Oberai The numerical modeling of multi-phase interfacial phase change phenomena, such as evaporation of a liquid or combustion of a solid, is essential for several important applications. A mathematically consistent and robust computational approach to address challenges such as large density ratio across phases, discontinuous fields at the interface, rapidly evolving geometries, and compressible phases, is presented in this work. We use the stabilized finite element methods on unstructured grids for solving the compressible Navier-Stokes equations. The rate of phase change rate is predicted from thermodynamic variables on both sides of the interface. We enforce the continuity of temperature and velocity in the tangential direction by using a penalty approach, while appropriate jump conditions derived from conservation laws across the interface are handled by using discontinuous interpolations. The interface is explicitly tracked using the arbitrary Lagrangian-Eulerian (ALE) technique, wherein the grid at the interface is constrained to move with the interface. [Preview Abstract] |
Monday, November 20, 2017 4:57PM - 5:10PM |
L2.00005: Effect of Finite Particle Size on Convergence of Point Particle Models in Euler-Lagrange Multiphase Dispersed Flow. Samaun Nili, Chanyoung Park, Raphael T. Haftka, Nam H. Kim, S. Balachandar Point particle methods are extensively used in simulating Euler-Lagrange multiphase dispersed flow. When particles are much smaller than the Eulerian grid the point particle model is on firm theoretical ground. However, this standard approach of evaluating the gas-particle coupling at the particle center fails to converge as the Eulerian grid is reduced below particle size. We present an approach to model the interaction between particles and fluid for finite size particles that permits convergence. We use the generalized Faxen form to compute the force on a particle and compare the results against traditional point particle method. We apportion the different force components on the particle to fluid cells based on the fraction of particle volume or surface in the cell. The application is to a one-dimensional model of shock propagation through a particle-laden field at moderate volume fraction, where the convergence is achieved for a well-formulated force model and back coupling for finite size particles. Comparison with 3D direct fully resolved numerical simulations will be used to check if the approach also improves accuracy compared to the point particle model. [Preview Abstract] |
Monday, November 20, 2017 5:10PM - 5:23PM |
L2.00006: Dynamic mesh adaptation for multi-material simulations using weighted condition number relaxation Patrick Greene, Sam Schofield, Robert Nourgaliev A mesh smoothing method designed to cluster cells near a dynamically evolving interface is presented. The method is based on weighted condition number mesh relaxation with the weight function being computed from either a level set or volume fraction representation of the interface. The weight function is expressed as a Taylor series based discontinuous Galerkin (DG) projection, which makes the computation of the derivatives of the weight function needed during the condition number optimization process a trivial matter. The method retains the excellent smoothing capabilities of condition number relaxation, while providing a method for clustering mesh cells near regions of interest requiring increased resolution. The algorithm has recently been implemented in one of LLNL’s arbitrary Lagrangian Eulerian (ALE) production codes. Results for a number of multi-material problems are presented, which will demonstrate the method’s great potential as a mesh relaxer for ALE simulations. [Preview Abstract] |
Monday, November 20, 2017 5:23PM - 5:36PM |
L2.00007: An update on the Eulerian formulation for the simulation of soft solids in fluids Suhas Jain S, Ali Mani Soft solids in fluids find wide applications in science, especially in the study of biological tissues and membranes. In this study an incompressible 2D Eulerian Finite volume solver has been developed on a fully collocated grid. We have adopted the Reference Map Technique by Valkov et. al (J. Appl. Mech., 82, 2015) as an approach to fully resolve hyperelastic solids in a fluid on an Eulerian grid. Multiple improvements for this technique are assessed. The extrapolation of the reference map field outside the solid region is performed using a cost effective Least Square Approach. Following recent adoptions, level-set field is constructed using the reference map field at every time step. These modifications allow simulations without artificial viscosity in solid regions while maintaining numerical robustness and have completely eliminated the striations of the interface that was seen before, hence eliminating the additional routines that were required for the smoothing of the interface. An approximate projection method has been used to project the velocity field onto a divergence free field. Cost and accuracy analysis of the solver has been performed. Further details of the computational techniques used and the results will be discussed in the presentation. [Preview Abstract] |
Monday, November 20, 2017 5:36PM - 5:49PM |
L2.00008: A continuum treatment of sliding in Eulerian simulations of solid-solid and solid-fluid interfaces Akshay Subramaniam, Niranjan Ghaisas, Sanjiva Lele A novel treatment of sliding is developed for use in an Eulerian framework for simulating elastic-plastic deformations of solids coupled with fluids. In this method, embedded interfacial boundary conditions for perfect sliding are imposed by enforcing the interface normal to be a principal direction of the Cauchy stress and appropriate consistency conditions ensure correct transmission and reflection of waves at the interface. This sliding treatment may be used either to simulate a solid-solid sliding interface or to incorporate an internal slip boundary condition at a solid-fluid interface. Sliding laws like the Coulomb friction law can also be incorporated with relative ease into this framework. Simulations of sliding interfaces are conducted using a 10\textsuperscript{th} order compact finite difference scheme and a Localized Artificial Diffusivity (LAD) scheme for shock and interface capturing. 1D and 2D simulations are used to assess the accuracy of the sliding treatment. The Richmyer-Meshkov instability between copper and aluminum is simulated with this sliding treatment as a demonstration test case. [Preview Abstract] |
Monday, November 20, 2017 5:49PM - 6:02PM |
L2.00009: A genuinely discontinuous approach for multiphase EHD problems Mahesh Natarajan, Olivier Desjardins Electrohydrodynamics (EHD) involves solving the Poisson equation for the electric field potential. For multiphase flows, although the electric field potential is a continuous quantity, due to the discontinuity in the electric permittivity between the phases, additional jump conditions at the interface, for the normal and tangential components of the electric field need to be satisfied. All approaches till date either ignore the jump conditions, or involve simplifying assumptions, and hence yield unconvincing results even for simple test problems. In the present work, we develop a genuinely discontinuous approach for the Poisson equation for multiphase flows using a Finite Volume Unsplit Volume of Fluid method. The governing equation and the jump conditions without assumptions are used to develop the method, and its efficiency is demonstrated by comparison of the numerical results with canonical test problems having exact solutions. [Preview Abstract] |
Monday, November 20, 2017 6:02PM - 6:15PM |
L2.00010: Dual domain material point method for multiphase flows Duan Zhang Although the particle-in-cell method was first invented in the 60's for fluid computations, one of its later versions, the material point method, is mostly used for solid calculations. Recent development of the multi-velocity formulations for multiphase flows and fluid-structure interactions requires the Lagrangian capability of the method be combined with Eulerian calculations for fluids. Because of different numerical representations of the materials, additional numerical schemes are needed to ensure continuity of the materials. New applications of the method to compute fluid motions have revealed numerical difficulties in various versions of the method. To resolve these difficulties, the dual domain material point method is introduced and improved. Unlike other particle based methods, the material point method uses both Lagrangian particles and Eulerian mesh, therefore it avoids direct communication between particles. With this unique property and the Lagrangian capability of the method, it is shown that a multiscale numerical scheme can be efficiently built based on the dual domain material point method. In this talk, the theoretical foundation of the method will be introduced. Numerical examples will be shown. [Preview Abstract] |
Monday, November 20, 2017 6:15PM - 6:28PM |
L2.00011: Slug-flow dynamics with phase change heat transfer in compact heat exchangers with oblique wavy walls Kenichi Morimoto, Hidenori Kinoshita, Ryo Matsushita, Yuji Suzuki With abundance of low-temperature geothermal energy source, small-scale binary-cycle power generation system has gained renewed attention. Although heat exchangers play a dominant role in thermal efficiency and the system size, the optimum design strategy has not been established due to complex flow phenomena and the lack of versatile heat transfer models. In the present study, the concept of oblique wavy walls, with which high j/f factor is achieved by strong secondary flows in single-phase system, is extended to two-phase exchangers. The present analyses are based on evaporation model coupled to a VOF technique, and a train of isolated bubbles is generated under the controlled inlet quality. R245fa is adopted as a low boiling-point working media, and two types of channels are considered with a hydraulic diameter of 4 mm: (i) a straight circular pipe and (ii) a duct with oblique wavy walls. The focus is on slug-flow dynamics with evaporation under small capillary but moderate Weber numbers, where the inertial effect as well as the surface tension is of significance. A possible direction of the change in thermo-physical properties is explored by assuming varied thermal conductivity. Effects of the vortical motions on evaporative heat transfer are highlighted. [Preview Abstract] |
Monday, November 20, 2017 6:28PM - 6:41PM |
L2.00012: Scalable Methods for Eulerian-Lagrangian Simulation Applied to Compressible Multiphase Flows David Zwick, Jason Hackl, S. Balachandar Multiphase flows can be found in countless areas of physics and engineering. Many of these flows can be classified as dispersed two-phase flows, meaning that there are solid particles dispersed in a continuous fluid phase. A common technique for simulating such flow is the Eulerian-Lagrangian method. While useful, this method can suffer from scaling issues on larger problem sizes that are typical of many realistic geometries. Here we present scalable techniques for Eulerian-Lagrangian simulations and apply it to the simulation of a particle bed subjected to expansion waves in a shock tube. The results show that the methods presented here are viable for simulation of larger problems on modern supercomputers. [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. |
© 2024 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
100 Motor Pkwy, Suite 110, Hauppauge, NY 11788
(631) 591-4000
Office of Public Affairs
529 14th St NW, Suite 1050, Washington, D.C. 20045-2001
(202) 662-8700