Bulletin of the American Physical Society
60th Annual Meeting of the Divison of Fluid Dynamics
Volume 52, Number 12
Sunday–Tuesday, November 18–20, 2007; Salt Lake City, Utah
Session FB: Computational Fluid Dynamics II |
Hide Abstracts |
Chair: Elias Balaras, University of Maryland Room: Salt Palace Convention Center 150 D-F |
Monday, November 19, 2007 8:00AM - 8:13AM |
FB.00001: Sharp Interface Immersed-Boundary/Level-Set Cartesian Grid Method for Large-Eddy Simulation of Two-Phase Flows with Surface-Piercing Moving Bodies Jianming Yang, Frederick Stern A sharp interface Cartesian grid method for the large-eddy simulation of two-phase flows interacting with surface-piercing moving bodies is presented. The method is based on a sharp interface immersed boundary formulation for fluid flows with moving boundaries and a level set based ghost fluid method for two-phase interface treatment. A four-step fractional step method is adopted and a Lagrangian dynamic Smagorinsky subgrid-scale model is used for large-eddy simulations. The combination of immersed boundary method for solid/fluid boundaries and ghost-fluid method for fluid/fluid interfaces is discussed in detail. A variety of test cases with different scales ranging from bubble dynamics to ship hydrodynamics are performed for verification and validation purpose. Several examples of interest such as water exit and entry of a circular cylinder, landslide generated waves, and ship waves are demonstrated to showcase the accuracy and efficiency of our method. Approaches for extending it to high Reynolds number ship flows by means of wall-layer modeling are also discussed. [Preview Abstract] |
Monday, November 19, 2007 8:13AM - 8:26AM |
FB.00002: An embedded boundary method with adaptive mesh refinement Elias Balaras, Marcos Vanella, Partick Rabenold An embedded boundary method with adaptive mesh refinements for fluid-structure interaction problems is presented. In the fluid solver the grid does not need to conform to a complex body, which is allowed to move through the fixed grid undergoing large displacements. To adaptively refine/derefine the mesh a single-block, staggered, Cartesian grid solver is employed on a hierarchy of sub-grids with varying spatial resolution. Each of these sub-grid blocks has a structured topology, and is part of a tree data structure that covers the entire computational domain. The Paramesh toolkit is used for the implementation of the adaptive refinement process. The package creates and maintains the hierarchy of sub-grid blocks, with each block containing a fixed number of grid points. A strong coupling scheme is adopted, where the fluid and the structure are treated as elements of a single dynamical system, and all of the governing equations are integrated simultaneously, and interactively in the time domain. A demonstration of the accuracy and range of applicability of the method will be given for a variety of laminar and turbulent flow problems. [Preview Abstract] |
Monday, November 19, 2007 8:26AM - 8:39AM |
FB.00003: Turbulent channel flow with random roughness on one wall Jorge Bailon-Cuba, Stefano Leonardi, Luciano Castillo A direct numerical simulation for a turbulent channel flow with a two-dimensional {\it large scale random roughness} distribution at its lower wall has been performed. The roughness elements are wedges of random height. The Navier-Stokes equations are discretized using staggered central second-order finite-differences, and the roughness is treated by the immersed boundary technique. The roughness geometry has been modified removing the small wedges. Also, a profile of uniformly spaced wedges with the same longitudinal area as the original case has been considered for comparison. Results show that by increasing the width of the cavity the pressure at the stagnation point increases. Also this is considerably affected only when the higher wedges are removed. Suppression of the small wedges also produces a decrease of the skin friction coefficient, $C_{f}(x)$, if the separated region around these increases its length. However, if the flow reattaches to the smooth wall the effect is an increase. The two modified geometries allow a more complete development of the mean stream-wise velocity profile, $U^{+}(y^{+})$, in the log-law and viscous regions. Also these two configurations, allow (until $y/h \approx 0.5$) a higher increase of the RMS-fluctuations $u, v,$ and $w$. [Preview Abstract] |
Monday, November 19, 2007 8:39AM - 8:52AM |
FB.00004: A More Effective Method for DNS of Deformable Solid Particles and Fibers with Sharp Fluid-Solid Boundary Jingshu Wu, Cyrus Aidun We have developed a new method for coupling the lattice-Boltzmann (LB) equations in the fluid phase with the lattice-Spring (LS) discretization of the deformable solid phase which we refer to as the `Immersed Grid' (IG) method. This approach allows direct numerical simulation of large number of deformable particles and fibers suspended in fluid. The LB equation for the fluid is solved on a fixed regular lattice where the deformable suspended particles are fit into a Lagrangian grid with sharp fluid-solid boundary. The fluid-solid interactions are enforced by the IG method where the fluid velocity field is solved by adding the force density term into the LB analysis as the particle movement is captured from the Newtonian dynamics equations. The LS method is applied to calculate the solid particle deformation. Compared to the regular ``bounce-back'' method for LB-LS coupling, we will show that this method is more stable and smooth. In the regular ``bounce-back'' method, the fluid-solid boundaries are approximated at the midpoints on the fluid grid, with sudden jump from node to node causing small fluctuations on the interaction forces. Compared to Immersed Boundary method, which uses artificial penalty parameters to simulate stiff particles, the new method can handle large deformations. The methodology is validated by comparing to experimental and analytical results. [Preview Abstract] |
Monday, November 19, 2007 8:52AM - 9:05AM |
FB.00005: A viscous vortex particle method for deforming bodies, with application to biolocomotion J. Zhang, J. Eldredge Bio-inspired mechanics of locomotion generally consist of the interaction of flexible structures with the surrounding fluid to generate propulsive forces. These mechanics have attracted new interest in recent years, particularly as candidates for novel autonomous vehicles. Hence, accurate numerical simulation tools are needed to explore such interactions in detail. In this work, we extend a viscous vortex particle method (VVPM) to continuously deforming two-dimensional bodies. The VVPM is a high-fidelity Navier-Stokes computational method which captures the fluid motion through evolution of vorticity-bearing computational particles. The kinematics of the deforming body surface are accounted for via a surface integral in the Biot-Savart velocity. The spurious slip in each time step is removed by computing an equivalent vortex sheet and allowing it to flux to adjacent particles. Particles of both uniform and variable size are utilized, and their relative merits considered. The placement of this method in the larger class of immersed boundary methods is explored. For validation purposes, we investigate a periodically deforming circular cylinder immersed in a stagnant fluid, for which an analytical solution can be derived when the deformations are small. We show that the computed vorticity and velocity of this motion are both in excellent agreement with the analytical solution. Finally, we explore the fluid dynamics of a simple fish undergoing undulatory motion in its backbone, to demonstrate the application of the method to biomorphic locomotion investigations. [Preview Abstract] |
Monday, November 19, 2007 9:05AM - 9:18AM |
FB.00006: An Implicit Immersed Method for Compressible Solid Interacting with Slightly Compressible Viscous Fluid X. Sheldon Wang In this paper, an implicit immersed method is employed for compressible solids interacting with slightly compressible viscous fluid. Mixed finite element formulations are implemented for both fluid and solid domains. The velocity unknowns for fluid domain are attached to a fixed Eulerian mesh or an arbitrary Eulerian-Lagrangian (ALE) mesh. In the implicit formulation, at each Newton-Krylov iteration, the nodal velocity unknowns of immersed solids are mapped/interpolated from the surrounding background fluid mesh. Both fluid and solid pressure unknowns are directly calculated. Although the immersed solid has the same displacement/velocity as the fictitious fluid occupying the same domain (kinematic matching), in general the bulk moduli and the pressures are different. In order to satisfy the inf-sup conditions, for both fluid and solid domains, high-order mixed elements, namely, 9-4c elements, satisfying the inf-sup conditions are used. Numerical examples are used to compare with traditional fluid-structure interaction approaches. Pros and cons of explicit vs implicit and incompressible vs compressible immersed methods are also discussed in this paper. [Preview Abstract] |
Monday, November 19, 2007 9:18AM - 9:31AM |
FB.00007: Nonreflecting Boundary Conditions Based on Nonlinear Multidimensional Characteristics Q. Liu, O.V. Vasilyev The nonlinear multidimensional characteristics-based nonreflecting boundary conditions are proposed for compressible Navier-Stokes equations with/without scalar transport equations. This model is consistent with the physics of flows and transport properties. Based on the theory of characteristics, the multidimensional flows can be decomposed into acoustic, entropy, and vorticity waves. In order to obtain appropriate nonreflecting boundary conditions, the corresponding characteristic variables of incoming waves are set to zeros, and the source terms of the incoming acoustic wave are partially damped. The plane waves are analyzed to obtain the optimal damping coefficient. This novel approach substantially minimizes the spurious wave reflections of pressure, density, temperature, velocity, and vorticity from the artificial boundaries, where strong multidimensional flow effects exist. The proposed method has advantages of simplicity, robustness, and numerical accuracy, due to the fact that it conforms to the flow physics. The methodology is tested on two benchmark problems: cylindrical acoustic waves propagation and the wake flow after cylinder with periodic strong vortex convected out of the computational domain. These numerical simulations yield accurate results and verify that the method substantially improves the 1-D characteristics-based nonreflecting boundary conditions for the complex multidimensional flows. [Preview Abstract] |
Monday, November 19, 2007 9:31AM - 9:44AM |
FB.00008: Modeling flow through inline tube bundles using an adaptive immersed boundary method Chunlei Liang, Xiaoyu Luo, Boyce Griffith Fluid flow and its exerted forces on the tube bundle cylinders are important in designing mechanical/nuclear heat exchanger facilities. In this paper, we study the vortex structure of the flow around the tube bundle for different tube spacing. An adaptive, formally 2$^{nd}$ order immersed boundary (IB) method is used to simulate the flow. One advantage of the IB method is its great flexibility and ease in positioning solid bodies in the fluid domain. Our IB approach uses a six-point regularized delta function and is a type of continuous forcing approach. Validation results obtained using the IB method for two-in-tandem cylinders compare well with those obtained using the finite volume or spectral element methods on unstructured grids. Subsequently, we simulated flow through six-row inline tube bundles with pitch-to-diameter ratios of 2.1, 3.2, and 4, respectively, on structured adaptively refined Cartesian grids. The IB method enables us to study the critical tube spacing when the flow regime switches from the vortex reattachment pattern to alternative individual vortex shedding. [Preview Abstract] |
Monday, November 19, 2007 9:44AM - 9:57AM |
FB.00009: Sources of force oscillations from an immersed boundary method for moving-body problems Jongho Lee, Jungwoo Kim, Haecheon Choi, Kyung-Soo Yang Although the immersed boundary method has been successfully applied to stationary-body problems, it produces force oscillations when applied to moving-body problems. In the present study, we identify two important sources of force oscillations from an immersed boundary method. One is from pressure discontinuity in space across the immersed boundary, which is caused by discrete momentum forcing applied at the grid points where solid becomes fluid with body motion. The other is from velocity discontinuity in time according to the body motion. It is shown that the mass source/sink proposed by Kim, Kim \& Choi (2001, JCP) reduces the force oscillations by alleviating the pressure discontinuity in space. The velocity discontinuity occurs in time at the grid points where fluid becomes solid with body motion. The amount of velocity discontinuity significantly depends on local grid size and becomes smaller with smaller grid size. [Preview Abstract] |
Monday, November 19, 2007 9:57AM - 10:10AM |
FB.00010: Numerical Simulation of Thermal Discharge Flows from Real-Life Diffusers H.S. Tang A three-dimensional Reynolds-averaged Navier-Stokes computational fluid dynamics (CFD) model developed by the authors and coworkers is employed to accurately simulate turbulent mixing in the near-fields of thermal discharges from real-life diffusers. A domain decomposition method with multi-level embedded overset grids is employed to handle the complexity of the realistic configurations as well as to efficiently account for the large disparity in length scales of the ambient river reaches and the discharge diffusers. An algebraic mixing length model with a Richardson-number correction for buoyancy effects is used for turbulence closure. The governing equations are solved with a second-order-accurate, finite-volume, artificial compressibility method. The model is validated in simulating a temperature stratified shear flow and a negatively buoyant wall jet, and the computed results are shown to be in good overall agreement with experimental measurements. In order to demonstrate the potential of the numerical model as a powerful engineering simulation tool, it is applied in turbulent initial mixing of thermal discharges loaded from both single-port and multi-port diffusers in a prismatic channel and a natural river. Comparisons of the CFD results with those obtained by two empirical mixing zone models widely used in practice today show that the numerical modeling practices yield very similar results in terms of both dilution rates and plume shapes. [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