Bulletin of the American Physical Society
65th Annual Meeting of the APS Division of Fluid Dynamics
Volume 57, Number 17
Sunday–Tuesday, November 18–20, 2012; San Diego, California
Session H5: Computational Fluid Dynamics V |
Hide Abstracts |
Chair: Sourabh Apte, Oregon State University Room: 24A |
Monday, November 19, 2012 10:30AM - 10:43AM |
H5.00001: Multi-scale Numerical Simulations of Magnetic Fluids Philip Yecko, Ruben Scardovelli, Holly Timme, A. David Trubatch We develop, validate and apply a new Height Function (HF) based Volume of Fluid (VOF) code to the simulation of ordinary and magnetic fluids in two-dimensional and three-dimensional axisymmetric geometries. The HF algorithm provides second-order accurate curvature and interface normal formulation, improving surface tension and magnetic stress accuracy. Motivated by our recent experimental results on ferrofluid rheology dominated by field-induced magnetic particle threads we have applied this code to the meso-scale problem of the interaction of flow with threads. The multiscale approach allows us to model solid particles embedded in bulk fluid, approximating the magnetic particle threads --which are several microns (thousands of magnetic nanoparticles) wide-- by means of a chain of several (from 3 to 11) pseudo-particles with magnetic properties intermediate between the ferrofluid and the magnetic particles. In the presence of an imposed uniform field, these pseudo-particles form chains (threads) along the field direction. By placing a single pseudo-particle chain in a uniform flow, simple shear flow and/or a pure straining flow we can directly examine the equilibrium chain orientation and hydrodynamic stresses which characterize the interaction between thread and shear. Our simulation results allow us to compute the energy dissipation and thus model the enhanced drag effected by a thread on fluid flow. [Preview Abstract] |
Monday, November 19, 2012 10:43AM - 10:56AM |
H5.00002: LES of turbulent boundary layer flow over irregular and multiscale topographies, and comparison with experimental data William Anderson, Kenneth Christensen Topographies featuring an irregular distribution of obstacles occur frequently in fluid machinery applications. Moreover, the distribution of characteristic size of these obstacles may also be broad. The hydrodynamic response of turbulence to such topographies is complicated, since both the flow and topography are composed of multiple length scales and identification of dominant scale is not obvious. Mejia-Alvarez and Christensen, 2010: Phys. Fluids, 22 015106 presented comprehensive experiments of developing and developed turbulent boundary layer flow over multiscale gas turbine blade topography. In the present work, results of large-eddy simulation of turbulent boundary layer flow over the topography considered by the experimentalists is presented. The topography is resolved with an immersed boundary method. The numerical and experimental results show reasonable agreement. The results are also used to make a posteriori evaluation of parameters used in classical relations linking the hydrodynamic roughness length to other statistics of the topography. [Preview Abstract] |
Monday, November 19, 2012 10:56AM - 11:09AM |
H5.00003: Numerical Study of Nusselt Number in a Heated Pipe With the Use of Variable-Order Resolution Kenneth Davis, Paul Fischer We present results for a numerical study of turbulent heat transfer in a pipe with a constant heat flux at the wall. Nusselt numbers are computed for Reynolds numbers between 5,000 and 15,000 over a wide range of Prandtl numbers and the results are compared to the Dittus-Boelter relation. The simulations are based on the spectral element method in which the velocity and pressure are represented by tensor-product polynomials of degree $N$ in each of $E$ elements. Typical values of $N$ are in the range 4 to 20 and, for this study, $E$ is between 4000 and 12000. We examine potential savings of using elevated resolution for the temperature field only, which is particularly interesting for the case $Pr > 1$. Specifically, for water flow, one has as Peclet to Reynolds number ratio of approximately six, which implies a need for elevated resolution of the temperature field. This study explores the relative convergence rates and costs when the polynomial order for the temperature field, $N_t$, is increased with respect to $N$. We identify the optimal ratio, $N_t/N$, as a function of Prandtl number as grid convergence is attained. [Preview Abstract] |
Monday, November 19, 2012 11:09AM - 11:22AM |
H5.00004: Parallel performance of iterative Poisson Solvers for Uniform and Structured Adaptive Grids on a Cray XT5 (Kraken) Marcos Vanella, Elias Balaras The Poisson equation solution is of great importance in areas as diverse as Gravitation and Electrostatics to fractional step methods for viscous incompressible flows. For uniform grids fast direct methods based on FFT are usually adopted, which scale well to thousands of processors. For block-wise structured adaptive grids, scaling depends on the amount of communication and load balancing inherent on the solution method. In the present study we report scaling tests on uniform and two-level grids of increasing size. A direct solver based on trigonometric transforms is used on the uniform grid cases, and three different iterative solvers are used on the adaptive cases. In particular, two multigrid algorithms and a bi-conjugate gradient stabilized algorithm, (BiPCGSTAB), preconditioned by one sweep of the Multigrid (no corrections applied in preconditioned) are utilized. All solvers are implemented within the Flash software infrastructure. The computations are performed on the Kraken, Cray XT5 supercomputer employing up to 40,000 cores. It is seen that interprocessor communication required by the data redistribution operations, and the base level fast solution has an important effect on the decline in parallel scaling. [Preview Abstract] |
Monday, November 19, 2012 11:22AM - 11:35AM |
H5.00005: Comparative Efficiency of Implicit, Explicit and Implicit-Explicit Strong Stability Preserving Methods Gabriel Moraes, Renan Teixeira, Leonardo Alves Several unsteady problems in transport phenomena require highly accurate solutions. Attempts to increase accuracy order of most numerical schemes, however, often weaken their linear and/or nonlinear stability. On the other hand, Strong Stability Preserving (SSP) methods are able to increase their accuracy order in time while still maintaining the overall stability properties of the original forward Euler method from which they were generated. This is achieved by, among other things, restricting the maximum allowed time step of these schemes. Hence, most recent studies have focused on developing optimal SSP schemes, i.e., minimally restrictive. Under this context, a significant variety of explicit and implicit formulations for multi-step and multi-stage marching schemes of different accuracy orders has been created. Despite their popularity, some open issues still remain. Linear stability of implicit schemes usually allows very large time steps, making them cost effective for many applications when compared to their explicit counterparts. Hence, the SSP time step restriction might render these schemes comparatively inefficient. The present study evaluates different explicit, implicit and implicit--explicit SSP time integration schemes for a series of test cases in an attempt to distinguish the most efficient scheme according to an error / computer time analysis. All three schemes employed a second order TVD flux limiter discretization for the spatial derivatives. [Preview Abstract] |
Monday, November 19, 2012 11:35AM - 11:48AM |
H5.00006: Superimposition of external oscillation to enhance heat transfer from objects in cross flow Raed Bourisli Laminar flow around objects gives rise to the recurrent build-up and release of vortices on alternate sides of the objects over a wide range of Reynolds numbers. Inherent disturbance of the otherwise uniform flow and temperature fields plays an important role in many structural, hydrodynamic as well as thermal aspects of situations where it is present. For example, the local disturbance of the velocity field leads to subsequent instability in the temperature field, causing variations in local Nusselt number, heat flux and surface temperature, among other things. One can take advantage of this phenomenon in many applications such as the cooling of electronic equipment. It is suggested here that the intensity of the outlined vortex shedding phenomenon can be deepened if an external movement is superimposed on the velocity of the structure or any nearby object. Numerical test of several objects rotated in-plane: cylinders, squares, triangles and horizontal plates, are performed. The key physical observation is the relative magnitudes of the heat transfer due to natural vortex shedding compared to the added value obtained by superimposing an additional external source of oscillation. A realistic case of electronics chips cooling is presented to show the effect of matching the natural frequency of vortex shedding by that of the inhomogeneity (\underline {Video: 0-9 s optimum; 9-18 non-}). In this case, vortex shedding from the plates plays a smaller role in disturbing the flow, hindering it at times. When the two frequencies coincide, however, in-phase shedding leads to more efficient heat transfer. [Preview Abstract] |
Monday, November 19, 2012 11:48AM - 12:01PM |
H5.00007: Statistic fluid dynamic of multiphase flow Hyunkyung Lim, James Glimm, Yijie Zhou, Xiangmin Jiao We study a turbulent two-phase fluid mixing problem from a statistical point of view. The test problem is high speed turbulent two-phase Taylor-Couette flow. We find extensive mixing in a transient state between an initial unstable and a final stable configuration. With chemical processing as a motivation, we estimate statistically surface area, droplet size distribution and transient droplet duration. [Preview Abstract] |
Monday, November 19, 2012 12:01PM - 12:14PM |
H5.00008: Comparative Study of Reynolds Averaged and Embedded Large Eddy Simulations of a High Pressure Turbine Stage Aleksandar Jemcov, Theodore Williams, Thomas Corke An Embedded Large Eddy Simulation (ELES) approach is used to simulate the flow path through a high pressure turbine stage. The turbine stage includes the entry duct, stationary inlet and exit guide vanes, and a rotor. The rotor blade design includes a squealer tip. The flowfield around the rotor is simulated using LES. A Reynolds Averaged Simulation (RAS) is used to simulate the rest of the flow domain. The interface between RAS and LES domains uses the RAS turbulence quantities as a means of obtaining length scales that are used in computing the vorticity that is required to trigger a proper energy cascade within the LES part of the flow field. The ELES approach allows for substantial computational savings since it allows for different mesh resolutions in various parts of the computational domain as needed. The objective of this work is to observe at a lower computational cost, the local flow features that cannot be resolved in a RAS approach. A comparative analysis between RAS and ELES approaches for this turbomachinery problem is then presented. [Preview Abstract] |
Monday, November 19, 2012 12:14PM - 12:27PM |
H5.00009: ABSTRACT WITHDRAWN |
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