Bulletin of the American Physical Society
64th Annual Meeting of the APS Division of Fluid Dynamics
Volume 56, Number 18
Sunday–Tuesday, November 20–22, 2011; Baltimore, Maryland
Session S12: Turbulence Simulation VI |
Hide Abstracts |
Chair: Kai Schneider, Universit\'e de Provence, Marseille Room: 315 |
Tuesday, November 22, 2011 3:05PM - 3:18PM |
S12.00001: Error Quantification in Simulations of Variable Density Low Mach Number Turbulent Flows Nicholas Robertson, Siddhartha Verma, Guillaume Blanquart In numerical simulations of low Mach number flows, the density is often expressed as a function of one or several transported scalar quantities through the use of an equation of state (EOS). As a result, these scalars play a critical role in controlling the accuracy of the overall simulation. This study aims at understanding and quantifying the sources of errors introduced in performing simulations of variable density low Mach number laminar and turbulent flows. As a first step, the relative importance of the order of accuracy of the scalar transport scheme and the velocity scheme is analyzed in canonical variable density test cases. Then, various implementations of the EOS are investigated in laminar flows. Particular importance is placed on the robustness, the convergence, and the accuracy of these implementations. Finally, the different scalar transport schemes and implementations of the EOS are combined and evaluated in simulations of turbulent planar jets. Following this analysis, guidelines for performing accurate and robust simulations of variable density low Mach number laminar/turbulent flows are proposed. [Preview Abstract] |
Tuesday, November 22, 2011 3:18PM - 3:31PM |
S12.00002: Large-Eddy Simulation of Crashback in a Ducted Propulsor Hyunchul Jang, Krishnan Mahesh Crashback is an operating condition to quickly stop a propelled vehicle, where the propeller is rotated in the reverse direction to yield negative thrust. The crashback condition is dominated by the interaction of free stream flow with strong reverse flow. Crashback causes highly unsteady loads and flow separation on blade surface. This study uses Large-Eddy Simulation to predict the highly unsteady flow field in crashback for a ducted propulsor. Thrust mostly arises from the blade surface, but most of side-force is generated from the duct surface. Both mean and RMS of pressure are much higher on inner surface of duct, especially near blade tips. This implies that side-force on the ducted propulsor is caused by the blade-duct interaction. Strong tip leakage flow is observed behind the suction side at the tip gap. The physical source of the tip leakage flow is seen to be the large pressure difference between pressure and suction sides. The conditional average during high amplitude event shows that the tip leakage flow and pressure difference are significantly higher. This work is supported by the United States Office of Naval Research under ONR Grant N00014-05-1-0003. [Preview Abstract] |
Tuesday, November 22, 2011 3:31PM - 3:44PM |
S12.00003: Investigating the effect of hull on marine propeller in crashback using LES Aman Verma, Hyunchul Jang, Krishnan Mahesh Crashback is an off-design condition where the marine vessel moves forward while the propeller is in reverse rotation, resulting in a highly unsteady separated flow. According to experiments, an upstream hull significantly increases the side-forces on a propeller in crashback below an advance ratio of $J=-0.7$. LES is performed for an open propulsor with and without hull at two advance ratios, $J=-1.0$ and $J=-0.5$. The experimentally observed behavior is reproduced using LES. At $J=-1.0$, two noticeable flow features are found with the hull - a recirculation zone located on the hull in the vicinity of the propeller and a vortex ring much closer to the propeller. In contrast, at $J = -0.5$, the recirculation zone is weaker and farther upstream of the propeller. As a result, the hull does not significantly alter the flow in the immediate vicinity of the propeller. For both advance ratios, side forces are mostly generated from the leading edge separation on suction side. However, high levels of side forces are also generated from trailing edge separation on suction side at $J=-1.0$. This work was supported by the United States Office of Naval Research under ONR Grant N00014-08-1-0433. [Preview Abstract] |
Tuesday, November 22, 2011 3:44PM - 3:57PM |
S12.00004: Marine Propeller Analysis with an Immersed Boundary LES methodology Seth Schroeder, Elias Balaras Modern marine propeller design and analysis techniques employ a wide range of computational methods to balance time constraints and accuracy requirements. Potential flow and RANS based methods have historically comprised the tools necessary for design and global performance analysis. However, for analysis where unsteady flow phenomena are of interest, eddy resolving methodologies are required. In the present study we use the large-eddy simulation (LES) approach coupled with an immersed-boundary (IB) method to perform computations of the flow around a rotating propeller at laboratory Reynolds numbers. Compared to classical boundary conforming strategies our formulation eliminates meshing overhead time and allows for body motion without additional treatments such as overset or dynamic meshing. The structured Cartesian grid also allows for a non-dissipative solver which conserves mass, momentum and energy. We will focus on the Italian Ship Model Basin (INSEAN) E1619 propeller and compare the predictions of our method to experimental results. The E1619 propeller is a 7-bladed submarine stock propeller that has been the subject of extensive experimental testing and previous computational studies. [Preview Abstract] |
Tuesday, November 22, 2011 3:57PM - 4:10PM |
S12.00005: Turbulence Simulation using many Graphics Processors Ali Khajeh-Saeed, J. Blair Perot Unsteady simulations of turbulence are performed using up to 64 graphics processors on the NSF XSede supercomputer, Lincoln, located at NCSA. For a $512^3$ simulations the performance of 16 GPUs (Tesla S1070) is about 45 times faster than that obtained with the same number of CPU cores of quad-core Intel Harpertown processors on the same machine. The code is optimized to use the fast shared-memory on the GPUs and to use communication/computation overlapping. Results show that the computation time is now so fast that even for large problems, with up to 8 million unknowns per GPU, the MPI communication time controls the scaling behavior of the CFD algorithm. [Preview Abstract] |
Tuesday, November 22, 2011 4:10PM - 4:23PM |
S12.00006: Development and testing of a user-friendly Matlab interface for the JHU turbulence database system Jason Graham, Edo Frederix, Charles Meneveau One of the challenges that faces researchers today is the ability to store large scale data sets in a way that promotes easy access to the data and sharing among the research community. A public turbulence database cluster has been constructed in which 27 terabytes of a direct numerical simulation of isotropic turbulence is stored (Li et al., 2008, JoT). The public database provides researchers the ability to retrieve subsets of the spatiotemporal data remotely from a client machine anywhere over the internet. In addition to C and Fortran client interfaces, we now present a new Matlab interface based on Matlab's intrinsic SOAP functions. The Matlab interface provides the benefit of a high-level programming language with a plethora of intrinsic functions and toolboxes. In this talk, we will discuss several aspects of the Matlab interface including its development, optimization, usage, and application to the isotropic turbulence data. We will demonstrate several examples (visualizations, statistical analysis, etc) which illustrate the tool. [Preview Abstract] |
Tuesday, November 22, 2011 4:23PM - 4:36PM |
S12.00007: Large eddy simulation of decaying homogeneous isotropic turbulence on a block-structured non-uniform grid using reconstruction Lauren Goodfriend, Tina Chow, Marcos Vanella, Elias Balaras Modeling turbulence directly requires too many computational resources for many realistic problems to be tractable, so methods have been developed to reduce the computational size of turbulence simulation. Large-eddy simulation and adaptive mesh refinement reduce the computational cost of turbulence modeling compared to direct numerical simulation, but they are rarely used together. Errors appear at grid refinement interfaces due to dependence of computed quantities on the filter width and insufficient smoothness of the solution at the grid scale. Here, approximate reconstruction of the unfiltered velocity field is used to decrease these errors in a simulation of decaying homogeneous isotropic turbulence advected past a grid refinement interface. Different explicit filter types and levels of reconstruction are tested. Reconstruction was found to improve results in cases using a binomial approximation to a Gaussian filter for a fine to coarse grid refinement interface, the natural transition for decaying turbulence. These results inform the use of large-eddy simulation on more complicated adaptive mesh refinement grids. [Preview Abstract] |
Tuesday, November 22, 2011 4:36PM - 4:49PM |
S12.00008: One-way coupling of nested grids in WRF-LES and local LES for turbulent boundary layer Tetsuro Tamura, Kojiro Nozawa This study proposes the method of one-way coupling between nested grids with different resolution for LES of turbulent boundary layer. Based on the difference of the momentum equations for the coarse and fine grids, we deduce the equation of the finer-grid component of velocity fluctuations. This equation has two types of SGS stress terms. One is the energy cascade from finer-grid GS turbulence to SGS turbulence and the other is the production term in this equation which cascades turbulence energy from coarser grid to finer grid. In order to validate the proposed method, we consider the a-priori test. Coarser-grid velocity is obtained by filtering the solution computed on finer grid. There are two types of finer-grid component of velocity, such as the predicted values by solving the proposed equation or the exact values. Comparison has been performed for the results of rough-wall turbulent boundary layer simulation. The spectra of velocity fluctuation obtained as sum of the predicted finer-grid component and the coarser-grid velocity appropriately fit to the -5/3 power law for the inertial sub-range. As posteriori test, the proposed method is also applied to atmospheric flows nesting WRF-LES and local LES. Coarser-grid velocity is obtained by WRF-LES. Finer-grid solutions are checked and the potential of the proposed method is investigated. [Preview Abstract] |
Tuesday, November 22, 2011 4:49PM - 5:02PM |
S12.00009: Designing LES of the High Reynolds Surface Layer to Account for Numerical Friction in the Algorithm. James Brasseur, Ganesh Vijayakumar, Matthew Churchfield, Adam Lavely, Eric Paterson, Patrick Moriarty Numerical friction stabilizes large-eddy simulation (LES), but also impacts accuracy. We explore this issue using a theory (Brasseur {\&} Wei 2010) where the LES is designed in a 3-parameter space that quantifies the level of friction in the SFS stress model (Re$_{LES})$, the relative content of resolved to SFS stress ($\Re )$, and surface layer resolution. To achieve law-of-the-wall in the mean, the LES must be in the ``high-accuracy zone'' (HAZ) of the $\Re -Re_{LES} $ parameter space. Using rough-wall channel flow and atmospheric boundary layer LES, we analyze simulations that are identical except for spectral vs. finite volume (FV) algorithms. Numerical friction shifts the LES away from the HAZ in the $\Re -Re_{LES} $ parameter space consistent with changes in mean shear-rate. The effective low pass filter from numerical friction shifts the total stress from resolved to subfilter-scale contributions, and effect that is more apparent when the spectral version of the LES is in the HAZ. A consequence is the enhancement of streamwise coherence in turbulence structure, particularly apparent in the integral scales. We shall discuss the requirements to adjust the FV LES to match a corresponding spectral LES in the HAZ, and differences in efficiency and accuracy. \textit{Support: NSF, DOE}. [Preview Abstract] |
Tuesday, November 22, 2011 5:02PM - 5:15PM |
S12.00010: Local relaxation and maximum entropy states in two-dimensional decaying turbulence Sergio Servidio, Minping Wan, William H. Matthaeus, Vicenzo Carbone The phenomenon of vortex merging in two-dimensional turbulence has been investigated through direct numerical simulations. The fast and local processes that occur during the turbulent relaxation of a randomly initialized system in periodic geometry have been examined. The analysis reveals that many of the coherent structures can be described by a local principle of maximization of entropy. The validity of this entropy principle has been further confirmed by time-dependent statistics using a contour-tracking technique. Implications for the description of persistent coherent vortices commonly observed in nature are suggested, including growing evidence for the wide applicability of maximum entropy-based relaxation principles. [Phys. Fluids 22, 125107 (2010)]. [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