Bulletin of the American Physical Society
63rd Annual Meeting of the APS Division of Fluid Dynamics
Volume 55, Number 16
Sunday–Tuesday, November 21–23, 2010; Long Beach, California
Session CC: Turbulence Modeling II |
Hide Abstracts |
Chair: Guillermo Araya, Swansea University UK Room: Long Beach Convention Center 102A |
Sunday, November 21, 2010 1:00PM - 1:13PM |
CC.00001: ABSTRACT WITHDRAWN |
Sunday, November 21, 2010 1:13PM - 1:26PM |
CC.00002: A pressure-strain correlation closure model with improved consistency with the Rapid Distortion Theory of Turbulence Ananda Mishra, Sharath Girimaji This work introduces the development of a new pressure-strain correlation model that is consistent with rapid distortion theory in two-dimensional strain and rotation dominated mean flows. Based on a modal (rather than statistical) analysis of the RDT equations and system bifurcation characteristics, small but important changes to the current pressure-strain correlation models are proposed. The closure procedure yields a direct relationship between the model coefficients and the RDT statistical data in the all-important intermediate regime of evolution. The new model coefficients depend on the mean-flow invariants and many of the current models can be recovered as special cases. The predictions of the new model are evaluated in a variety of canonical test cases. [Preview Abstract] |
Sunday, November 21, 2010 1:26PM - 1:39PM |
CC.00003: Bridging Approach to variable-resolution turbulence simulations -- Unification of PANS and PITM closures Sharath Girimaji, Robert Rubinstein Variable-resolution turbulence simulation schemes can be classified into two general categories: hybrid methods and bridging approaches. Hybrid computations entail Reynolds averaged Navier-Stokes (RANS) calculations in some flow regions and large eddy simulations (LES) in others. There exists no clear consensus on the criterion for switching from RANS to LES. The bridging methods, on the other hand, seamlessly transition from one flow resolution to another. There are currently two major bridging approaches -- Partially-Averaged Navier Stokes (PANS) and Partially Integrated Turbulence Model (PITM). While the two bridging closure expressions share many similarities, there are notable differences as well. In this work, we generalize the PITM derivation to include additional features and demonstrate complete consistency with PANS closure. The unification of the two methods confirms the mathematical rigor of the bridging approach and highlights its advantage over hybrid methods. [Preview Abstract] |
Sunday, November 21, 2010 1:39PM - 1:52PM |
CC.00004: Modeling near-wall turbulence behavior for variable-resolution bridging methods Dasia Reyes, Sharath Girimaji High-fidelity near-wall modeling is crucial for the success of all variable-resolution turbulence modeling approaches. In this work, we revisit the log-law analysis in the context of Partially Averaged Navier-Stokes (PANS) bridging method. In the PANS method, the cut-off resolution is parameterized in terms of ratios of unresolved-to-resolved kinetic energy and dissipation. The behavior of the unresolved stresses (second-order central moments) and dissipation, as well as the resolved fields subject to different cut-off lengthscales are examined. The modeling of the related turbulent transport term is also considered. Although this work in performed in the PANS context, its findings can benefit closure modeling for all hybrid methods. [Preview Abstract] |
Sunday, November 21, 2010 1:52PM - 2:05PM |
CC.00005: Evaluation of Partially-Averaged Navier-Stokes (PANS) bridging method in turbulent channel flows Branislav Basara, Sharath Girimaji, Pavlovic Pavlovic The Partially-Averaged Navier-Stokes (PANS) variable-resolution turbulence computational method is intended for seamless bridging between Reynolds-Averaged Navier-Stokes (RANS) and Direct Numerical Solution (DNS). While the success of PANS has been well documented in separated flows with largescale instabilities, its performance in wall-bounded flows is yet to be demonstrated. Toward that end, we perform channel flow simulations using the PANS $\zeta $-f closure which is based on the near-wall RANS $\zeta$-f model. The filter width is controlled by specifying the appropriate control parameters, unresolved-to-total ratios of turbulent kinetic energy and frequency. Computations are performed with three different grids (filter-widths) for the case of Re$_{\tau}$ = 650. Direct Numerical Simulation and Large eddy simulation data are used for comparison and evaluation. The results clearly demonstrate that PANS performs well in near-wall turbulent flows as well. [Preview Abstract] |
Sunday, November 21, 2010 2:05PM - 2:18PM |
CC.00006: Turbulence Closures Uncertainties in Shock Boundary Layer Interactions Michael Emory, Rene Pecnik, Gianluca Iaccarino Reynolds averaged closures have limited predictive capabilities when applied to the problem of shock boundary layer interaction. Several modifications to RANS models have been proposed in the literature, including compressibility corrections, limiters, and alternative forms of the turbulence production terms. Our objective is to characterize the errors introduced by the various approximations used in typical two-equation models, such as turbulence isotropy, linear stress-strain relationship, dissipation rate, etc. by isolating each contribution separately. We use the barycentric map to alter the turbulence anisotropy and introduce realizable Reynolds stress perturbations to investigate the effect of potential modeling errors on the resulting wall pressure and shock position. This method of measuring structural uncertainty is compared against two traditional uncertainty quantification approaches which evaluate the effect of boundary condition and model coefficient variability on the quantities of interest. We also discuss how appropriate input uncertainty ranges are determined. [Preview Abstract] |
Sunday, November 21, 2010 2:18PM - 2:31PM |
CC.00007: Adaptive wall functions for moving walls using the {\it k}-$\omega$ turbulence model John Axerio-Cilies, Gianluca Iaccarino An adaptive wall function for the {\it k}-$\omega$ model is derived for moving walls starting from a wall-resolved RANS computation of the flow over a moving flat plate with zero pressure gradient. The wall function is implemented via lookup tables for the turbulence quantities and the friction velocity $u_\tau$. The reference well-resolved, grid-converged RANS numerical solutions are obtained using the {\it k}-$\omega$ turbulence model with wall integration on very fine grids ($y+ <1$). Selecting a reference frame such that $U_{\infty}\geq 0$ yields three distinct velocity profile regimes: $U_{\infty}\geq 0 < U_w$, $U_{\infty}\geq U_w \geq 0$, and $U_w>U_{\infty}\geq0$. It is shown that adaptive wall functions are appropriate for all three velocity profile regimes as well as different Reynolds numbers when the near wall grid resolution is not sufficient ($y+>1$). For very fine grids ($y+<1$) this approach yields results consistent with the wall integration solution. Finally, the performance of the proposed adaptive wall functions is investigated for the complex flow around a rotating Formula 1 tire. The complexity of this flow arises from the impingement and jetting at the front of the tire, strong pressure gradients, and the large separated region behind the tire. [Preview Abstract] |
Sunday, November 21, 2010 2:31PM - 2:44PM |
CC.00008: Two-Equation Turbulence Model Predictions of Transition in Buoyancy- and Shock-Driven Flows Bryan Johnson, Oleg Schilling Two-equation Reynolds-averaged Navier--Stokes (RANS) models are generally regarded as relevant only for fully-developed turbulent flow. It is shown here that the early-time evolution of these models captures the turbulent transition for Rayleigh--Taylor instability and shock--turbulence interaction. When the fluctuation energy is much less than the mean flow energy, turbulent diffusion is negligible and the equations can be integrated analytically for a steady mean flow. For an incompressible flow, the turbulent kinetic energy grows exponentially at the physical growth rate (with appropriate model coefficients). For a shock-driven flow, the turbulent kinetic energy is amplified over the advection time across the shock, with an amplification factor equivalent to the physical amplification factor. Once turbulent diffusion becomes important, the turbulent quantities across the mixing layer are generally insensitive to the initial evolution. The primary consequence of varying model coefficients and initial conditions in the linear regime is a shift in the time at which the mixing layer begins to develop. [Preview Abstract] |
Sunday, November 21, 2010 2:44PM - 2:57PM |
CC.00009: Bayesian Calibration and Comparison of RANS Turbulence Models for Channel Flow Todd Oliver, Robert Moser A set of RANS turbulence models---including Baldwin-Lomax, Spalart-Allmaras, $k$-$\epsilon$, and $\overline{v^2}$-$f$---are calibrated and compared in the context of fully-developed channel flow. Specifically, a Bayesian calibration procedure is applied to infer the parameter values for each turbulence model from channel flow DNS data. In this process, uncertainty arises both from uncertainty in the data and inadequacies in the turbulence models. Various stochastic models of the turbulence model inadequacy are formulated, and the impacts of different uncertainty modeling choices are examined. The calibrated turbulence models are compared in terms of two items: posterior plausibility and predictions of quantities of interest such as centerline velocity and the location of the maximum Reynolds shear stress. The posterior plausibility indicates which model is preferred by the data according to Bayes' theorem, while the predictions allow assessment of how strongly the model differences impact the quantities of interest. The implications of these comparisons for turbulence model validation will be discussed. This work is supported by the Department of Energy [National Nuclear Security Administration] under Award Number [DE-FC52-08NA28615]. [Preview Abstract] |
Sunday, November 21, 2010 2:57PM - 3:10PM |
CC.00010: Reproducing second order statistics of turbulent flows using linearized Navier-Stokes equations with forcing Mihailo Jovanovic, Tryphon Georgiou We study the problem of reproducing second order statistics of turbulent flows using linearized Navier-Stokes (NS) equations with forcing. This forcing is represented by a stochastic excitation that enters into the equations as an additive spatio- temporal body force. For homogeneous isotropic turbulence, we show that the steady-state velocity correlation tensors can be exactly matched by the linearized NS equations subject to a temporally white solenoidal forcing with appropriately selected second order statistics. For turbulent channel flows, however, forcing of the linearized equations by colored-in-time stochastic process is required. The forcing spectra, which are consistent with DNS data, are obtained from a solution of the maximum entropy optimization problem. We show how this forcing can be generated as an output of a spatio-temporal filter driven by white-in-time stochastic process with appropriately selected second order statistics. Our results can be used to model forcing correlations in, (i) receptivity analysis of a free-stream turbulence induced transition in boundary layers and (ii) design of flow estimators and controllers for turbulence suppression in wall-bounded shear 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