Bulletin of the American Physical Society
68th Annual Meeting of the APS Division of Fluid Dynamics
Volume 60, Number 21
Sunday–Tuesday, November 22–24, 2015; Boston, Massachusetts
Session E22: Turbulent Boundary Layers: Wall Modeling |
Hide Abstracts |
Chair: Kurt Aikens, Houghton College Room: 210 |
Sunday, November 22, 2015 4:50PM - 5:03PM |
E22.00001: Inviscid Wall-Modeled Large Eddy Simulations for Improved Efficiency Kurt Aikens, Kyle Craft, Andrew Redman \hyphenpenalty=1000 The accuracy of an inviscid flow assumption for wall-modeled large eddy simulations (LES) is examined because of its ability to reduce simulation costs. This assumption is not generally applicable for wall-bounded flows due to the high velocity gradients found near walls. In wall-modeled LES, however, neither the viscous near-wall region or the viscous length scales in the outer flow are resolved. Therefore, the viscous terms in the Navier-Stokes equations have little impact on the resolved flowfield. Zero pressure gradient flat plate boundary layer results are presented for both viscous and inviscid simulations using a wall model developed previously.\footnote{K.~M.~Aikens, ``High-fidelity large eddy simulation for supersonic jet noise prediction,'' Ph.D.~thesis, Purdue University, 2014.} The results are very similar and compare favorably to those from another wall model methodology\footnote{S.~Kawai and J.~Larsson, Phys.~Fluids {\bf 24}, 015105 (2012).} and experimental data. Furthermore, the inviscid assumption reduces simulation costs by about 25\% and 39\% for supersonic and subsonic flows, respectively. Future research directions are discussed as are preliminary efforts to extend the wall model to include the effects of unresolved wall roughness. [Preview Abstract] |
Sunday, November 22, 2015 5:03PM - 5:16PM |
E22.00002: Application of the Integral Length-Scale Approximation to Wall Modelled LES Amirreza Rouhi, Ugo Piomelli, Alexandre Silva-Lopes A new length-scale for modelling the unresolved stresses in LES was proposed [Piomelli \textit{et al.}, \textit{J. Fluid Mech.}, \textbf{766}, 2015] in which the filter width is related to the turbulence statistics instead of the grid. The model constant is assigned by requiring that the unresolved, subfilter scales (SFS) support some percentage of the total stress. This model gave very good results in wall-resolved LES channel flow. When the same model is applied to wall-modelled simulations, however, significant errors result due to the requirement that the contribution of SFS to the transport be constant through the channel. Near the wall, the grid becomes larger than the mixing length of the flow, and the resolved eddies are not able to support the desired contribution to the transport. Better agreement is obtained by requiring that, as the wall is approached, the SFS contribute an increasing percentage of the momentum transport, reaching 100\% at the wall. The model was tested on channel flows at $Re_\tau = 5\times10^4$ and $5\times10^5$, using $256\times80\times128$ and $256\times160\times128$ grid points. The simulations predict the universal log-law very well. The model has the same cost as the Smagorinsky model and the robustness of the dynamic eddy-viscosity model. [Preview Abstract] |
Sunday, November 22, 2015 5:16PM - 5:29PM |
E22.00003: Multiple-relaxation-time lattice Boltzmann simulations of turbulent channel and pipe flows. Harish Opadrishta, Cheng Peng, Lian-Ping Wang The mesoscopic Lattice Boltzmann method (LBM) has become a reliable alternative for solving incompressible turbulent flows. However, the statistics of a simulated turbulent flow near a curved boundary may deviate from the physical rotational invariance (RI) of lattice coordinates. The main objective of this study is to compare the effects of different lattice models on the simulation results of turbulent flows, and explore ways to restore RI near a curved boundary. We will apply D3Q19 and D3Q27 multiple-relaxation-time LBM models to simulate turbulent pipe and channel flows. The statistics of the simulated flows are examined to quantify the nature of departures from RI. To help understand whether the departure is originated from the bounce-back scheme at the solid wall, we will perform simulations of a turbulent channel flow with walls orientated at an angle from the lattice grid, and test the use of an overset lattice grid near a pipe wall. The Chapman-Enskog analysis of these models will be performed to probe RI errors near a boundary. Our goal is to eventually perform an accurate direct numerical simulation of a turbulent pipe flow, and compare the results to previous simulations based on the Navier-Stokes equations. [Preview Abstract] |
Sunday, November 22, 2015 5:29PM - 5:42PM |
E22.00004: Assesment of turbulence models for boundary layers with pressure gradient and roughness Rabijit Dutta, Ugo Piomelli The performance of sand-grain-based roughness corrections for the SA, SST $k-\omega$ and $k-\epsilon$ models has been evaluated by comparing the model results with large eddy simulation (LES) data. Computations are performed for a turbulent boundary layer with both smooth and rough walls subjected to two different pressure-gradient conditions, namely, an adverse pressure gradient (APG) with separation and a realistic pressure-gradient situation encountered in a hydraulic turbine blade. A new roughness correction was developed for the SST $k-\omega$ model that gave improved results near separation. For the cases with smooth wall, RANS models give reasonable agreement in predicting skin friction coefficient ($c_f$) at the wall. RANS models predict too high Reynolds stresses in the separated region, which lead to earlier reattachment. For the rough wall computations, the RANS models predict that $c_f$ changes sign much later than the LES data. In the LES, however, the wall stress becomes negative inside the roughness sublayer, and the flow reversal does not correspond to the separation, which occurs much later, where the separation leaves the body, and the total stress above the roughness crest changes sign. The RANS models predict the position of this point more accurately. [Preview Abstract] |
Sunday, November 22, 2015 5:42PM - 5:55PM |
E22.00005: Investigation of pressure gradient aware wall modeling in LES Olivier Thiry, Gregoire Winckelmans, Matthieu Duponcheel This work focuses on the investigation of various wall modeling strategies for the simulation of high Reynolds number wall-bounded turbulent flows with acceleration and/or deceleration. Our code is based on fourth order finite differences, is momentum conserving, and is energy conserving up to fourth order. We here use a ``channel flow'' set-up, with no slip and wall modeling at the bottom, with slip at the top, and with blowing and/or suction at the top in order to generate the desired acceleration-deceleration profile. Two strategies are investigated and compared. Pressure gradient corrected algebraic models are first considered, and we investigate various local averaging techniques so as to avoid imposing mean profile laws pointwise. RANS sub-layer models are then also considered, where the turbulent viscosity is corrected to account for pressure gradient effects and for resolved LES fluctuations effects. A wall-resolved LES was also performed to provide a reference solution. [Preview Abstract] |
Sunday, November 22, 2015 5:55PM - 6:08PM |
E22.00006: Improved engineering models for turbulent wall flows Zhen-Su She, Xi Chen, Hong-Yue Zou, Fazle Hussain We propose a new approach, called \textit{structural ensemble dynamics} (SED), involving new concepts to describe the mean quantities in wall-bounded flows, and its application to improving the existing engineering turbulence models, as well as its physical interpretation. First, a revised $k-\omega $ model for pipe flows is obtained, which accurately predicts, for the first time, both mean velocity and (streamwise) kinetic energy $\left\langle {u'u'} \right\rangle $ for a wide range of the Reynolds number (\textit{Re}), validated by Princeton experimental data. In particular, a multiplicative factor is introduced in the dissipation term to model an anomaly in the energy cascade in a meso-layer, predicting the outer peak of $\left\langle {u'u'} \right\rangle $ agreeing with data. Secondly, a new one-equation model is obtained for compressible turbulent boundary layers (CTBL), building on a multi-layer formula of the stress length function and a generalized temperature-velocity relation. The former refines the multi-layer description - viscous sublayer, buffer layer, logarithmic layer and a newly defined bulk zone - while the latter characterizes a parabolic relation between the mean velocity and temperature. DNS data show our predictions to have a 99{\%} accuracy for several Mach numbers \textit{Ma}$=$2.25, 4.5, improving, up to 10{\%}, a previous similar one-equation model (Baldwin {\&} Lomax, 1978). Our results promise notable improvements in engineering models. [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