Bulletin of the American Physical Society
61st Annual Meeting of the APS Division of Fluid Dynamics
Volume 53, Number 15
Sunday–Tuesday, November 23–25, 2008; San Antonio, Texas
Session BR: CFD: Continuum Mechanics |
Hide Abstracts |
Chair: Peter Diamessis, Cornell University Room: 203A |
Sunday, November 23, 2008 10:30AM - 10:43AM |
BR.00001: Wall Boundary Conditions for Two Point Correlations Michael Nilsson, J. Blair Perot The reaction of isotropic homogeneous turbulent flow to the instantaneous insertion of a wall is investigated using direct numerical simulation (DNS). At small time-scales, the viscous terms have yet to affect the flow. Therefore, the wall boundary condition implemented is a free-slip wall. Analysis of the velocity fields is performed using two-point correlations taken over planes parallel to the wall placement. In this talk, the resulting two-point correlations are presented, and a model to predict the behavior of the two-point correlations is also proposed. The strengths of the model are presented, in relation to the orientation of the wall placement in the domain. It will be shown that for correlations involving the wall-normal velocity component, the predictive model performs exceptionally well at high resolutions. The decomposition of the two-point correlations was performed in order to gain a better understanding of the effects of the boundary conditions, and is presented. The investigation into the role of vorticity for further development of the predictive model involving the non-normal terms will also be discussed. [Preview Abstract] |
Sunday, November 23, 2008 10:43AM - 10:56AM |
BR.00002: New algorithms for the direct numerical simulation of turbulent flows past compliant bodies Anish Karandikar, Thomas Bewley This work develops an efficient and accurate new method for the DNS of laminar and turbulent flow past a circular cylinder with a deformable (compliant) surface. We study the interaction of the incompressible flow with the compliant cylinder. From the outset, this is defined as an optimization problem, in which we seek to minimize aeroacoustic noise generated by dipole sound sources on the compliant surface at low Mach numbers. We build on a unique method developed in our lab for simulating turbulent flow in a channel with compliant walls. This method is accurate and efficient for large surface deformations. We adapt this method for the cylindrical polar coordinate system to study flow past a compliant cylinder. In this method, a time-dependent coordinate transformation is used to map the deformed flow domain to a regular computational domain. The governing Navier Stokes equations are formulated in the cylindrical polar form and not the contravariant form, as the latter is computationally expensive to simulate. The compliant surface is modeled by a simple spring-mass-damper system. As surface compliance is increased, a decrease in the peak lift coefficient for the compliant cylinder is observed both in the laminar 2D case at $Re=80$, as well as the turbulent 3D case at $Re=300$. On the other hand, the frequency of vortex shedding and the time-average drag both increase with surface compliance. [Preview Abstract] |
Sunday, November 23, 2008 10:56AM - 11:09AM |
BR.00003: Assessent of elliptic solvers for the pressure Poisson equation J.P. Strodtbeck, J.B. Polly, J.M. McDonough It is well known that as much as 80\% of the total arithmetic needed for a solution of the incompressible Navier--Stokes equations can be expended for solving the pressure Poisson equation, and this has long been one of the prime motivations for study of elliptic solvers. In recent years various Krylov-subspace methods have begun to receive wide use because of their rapid convergence rates and automatic generation of iteration parameters. However, it is actually total floating-point arithmetic operations that must be of concern when selecting a solver for CFD, and not simply required number of iterations. In the present study we recast speed of convergence for typical CFD pressure Poisson problems in terms of CPU time spent on floating-point arithmetic and demonstrate that in many cases simple successive-overrelaxation (SOR) methods are as effective as some of the popular Krylov-subspace techniques such as BiCGStab(l) provided optimal SOR iteration parameters are employed; furthermore, SOR procedures require significantly less memory. We then describe some techniques for automatically predicting optimal SOR parameters. [Preview Abstract] |
Sunday, November 23, 2008 11:09AM - 11:22AM |
BR.00004: A hybrid parallel solver for flow and scalar transport in complex geometries Dazhi Yu, Tony Ladd We have developed a hybrid parallel solver to simulate flow and transport in an explicit pore topography. The lattice Boltzmann (LB) method was used to determine the fluid velocity, and a finite difference scheme was used to solve for the scalar transport. At high Peclet numbers the scalar field can vary more rapidly than the fluid velocity, especially near solid boundaries. Thus the resolution of velocity and scalar fields can be chosen independently. Solid-fluid interfaces of arbitrary geometric complexity can be described by the distance of each surface point from a regular grid just inside the solid. These surface points lie on grid lines intersecting the surface. At each point, precipitation/dissolution reactions can occur, based on the local scalar concentration. During erosion or precipitation, the points move normal to the surface, and are interpolated back to the grid lines using a cubic Bezier surface constructed around the old intersection point. Because of the large aspect ratio of typical fractures, we require a high fidelity finite difference method to solve for the scalar field. We have designed a scheme that solves along characteristics (like the LB solver), which is much more accurate than a Lax-Wendroff method when the flow is oriented away from the grid lines. [Preview Abstract] |
Sunday, November 23, 2008 11:22AM - 11:35AM |
BR.00005: Enhancement of PANS model's performance by introducing advanced numerical techniques Branislav Basara The Partially-Averaged Navier-Stokes (PANS) approach is a recently proposed method by Girimaji (2003), which changes seamlessly from RANS to the direct numerical solution of the Navier-Stokes equations (DNS) as the unresolved-to-total ratios of kinetic energy and dissipation are varied. The parameter which determines the unresolved-to-total kinetic energy ratio fk is defined based on the grid spacing. The PANS asymptotic behaviour goes smoothly from RANS to DNS with decreasing fk. In the work of Basara, Krajnovic and Girimaji (2008), it was shown that a dynamic update of the PANS key parameter fk by changing at each point and at the end of every time step is the promising approach to provide the optimum modeling on employed computational meshes. This work is extended here by introducing numerical techniques which efficiently increase a grid resolution and with that, decrease the parameter fk. This is achieved by employment of the effective non-reflecting boundary conditions and cutting computational domains, and by adaptive grids which allow keeping in advance prescribed value of the parameter fk. The results will show benefits of using the advanced numerical techniques in conjunction with PANS method. [Preview Abstract] |
Sunday, November 23, 2008 11:35AM - 11:48AM |
BR.00006: Development of Boundary Condition Independent Reduced Order Thermal Models using Proper Orthogonal Decomposition Arun Raghupathy, Karman Ghia, Urmila Ghia Compact Thermal Models (CTM) to represent IC packages has been traditionally developed using the DELPHI-based \textbf{(DE}velopment of \textbf{L}ibraries of \textbf{PH}ysical models for an \textbf{I}ntegrated design) methodology. The drawbacks of this method are presented, and an alternative method is proposed. A reduced-order model that provides the complete thermal information accurately with less computational resources can be effectively used in system level simulations. Proper Orthogonal Decomposition (POD), a statistical method, can be used to reduce the order of the degree of freedom or variables of the computations for such a problem. POD along with the Galerkin projection allows us to create reduced-order models that reproduce the characteristics of the system with a considerable reduction in computational resources while maintaining a high level of accuracy. The goal of this work is to show that this method can be applied to obtain a boundary condition independent reduced-order thermal model for complex components. The methodology is applied to the 1D transient heat equation. [Preview Abstract] |
Sunday, November 23, 2008 11:48AM - 12:01PM |
BR.00007: Improving Proper Orthogonal Decomposition Robustness by Optimal Sampling Michel Bergmann, Angelo Iollo This talk focuses on improving the robustness of the functional subspace built using Proper Orthogonal Decomposition (POD). Since a POD basis is able to give an optimal representation of the kinetic energy included in the snapshots database generated with some given input parameters, this same basis is not adapted to represent flow dynamics generated with other input parameters. Our aim is thus to build a POD basis that accurately represents the solution over a desired input parameter subspace by enlarging the database. We present a systematic method to sample the input parameter subspace. The basic idea is to add to the existing database, snapshots of the solution for which the POD approximation error is maximal. This is the Greedy sampling. The approach we follow is similar: it is based on finding the centroid of a region around the point where an estimate of the POD approximation error is maximal. We show numerical evidence that the Navier-Stokes residuals are a reliable estimate of the POD approximation error. Results relative to a 2D confined square cylinder wake flow are presented. The input parameter subspace is represented by an interval of Reynolds numbers that corresponds to periodical laminar flows. We show that a judicious choice of the sampling Reynolds numbers leads to a POD basis that minimizes the average approximation error on the chosen interval. [Preview Abstract] |
Sunday, November 23, 2008 12:01PM - 12:14PM |
BR.00008: A Spectrally-accurate Gridless Method for the Navier-Stokes Equations with Moving Boundaries Syad Hussain, Jerzy M. Floryan A gridless, spectrally-accurate algorithm for the Stokes flow with moving boundaries is presented. The boundaries of the flow domain move inside the fixed computational domain. Spatial discretization uses Fourier and Chebyshev expansions in the streamwise and transverse directions respectively. Temporal discretization is based on first and second order implicit formulations. The boundary conditions on the moving boundaries are imposed using the immersed boundary conditions concept. Numerical tests confirm the spectral accuracy in space and theoretically-predicted accuracy in time. [Preview Abstract] |
Sunday, November 23, 2008 12:14PM - 12:27PM |
BR.00009: A bulk viscosity approach for shock capturing on unstructured grids Mohammad Shoeybi, Nils Johan Larsson, Frank Ham, Parviz Moin The bulk viscosity approach for shock capturing (Cook and Cabot, JCP, 2005) augments the bulk part of the viscous stress tensor. The intention is to capture shock waves without dissipating turbulent structures. The present work extends and modifies this method for unstructured grids. We propose a method that properly scales the bulk viscosity with the grid spacing normal to the shock for unstructured grid for which the shock is not necessarily aligned with the grid. The magnitude of the strain rate tensor used in the original formulation is replaced with the dilatation, which appears to be more appropriate in the vortical turbulent flow regions (Mani et al., 2008). The original form of the model is found to have an impact on dilatational motions away form the shock wave, which is eliminated by a proposed localization of the bulk viscosity. Finally, to allow for grid adaptation around shock waves, an explicit/implicit time advancement scheme has been developed that adaptively identifies the stiff regions. The full method has been verified with several test cases, including 2D shock-vorticity entropy interaction, homogenous isotropic turbulence, and turbulent flow over a cylinder. [Preview Abstract] |
Sunday, November 23, 2008 12:27PM - 12:40PM |
BR.00010: A Method for Fast Computation of FTLE Fields Steven Brunton, Clarence Rowley An efficient method for computing finite time Lyapunov exponent (FTLE) fields is investigated. FTLE fields, which measure the stretching between nearby particles, are important in determining transport mechanisms in unsteady flows. Ridges of the FTLE field are Lagrangian Coherent Structures (LCS) and provide an unsteady analogue of invariant manifolds from dynamical systems theory. FTLE field computations are expensive because of the large number of particle trajectories which must be integrated. However, when computing a time series of fields, it is possible to use the integrated trajectories at a previous time to compute an approximation of the integrated trajectories initialized at a later time, resulting in significant computational savings. This work provides analytic estimates for accumulated error and computation time as well as simulations comparing exact results with the approximate method for a number of interesting flows. [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