Bulletin of the American Physical Society
72nd Annual Meeting of the APS Division of Fluid Dynamics
Volume 64, Number 13
Saturday–Tuesday, November 23–26, 2019; Seattle, Washington
Session P19: Advanced Turbulence Models II |
Hide Abstracts |
Chair: Guillaume Blanquart, Caltech Room: 401 |
Monday, November 25, 2019 5:16PM - 5:29PM |
P19.00001: Colouring turbulence with the nonlinear terms Sean Symon, Simon Illingworth, Ivan Marusic We study the behaviour of the nonlinear terms in turbulent channel flow from direct numerical simulations (DNS) at low and moderate Reynolds numbers. These terms correspond to the nonlinear forcing of the (linear) resolvent-based models of McKeon \& Sharma 2010. Spectral proper orthogonal decomposition (SPOD) is used to extract the most energetic structures and identify the power spectrum of the most energetic modes. We also compare the nonlinear terms to the equivalent nonlinear forcing supplied by eddy viscosity-enhanced resolvent models. At spatial wavenumbers where there is energetic activity, the nonlinear terms share many similarities with their velocity fluctuation counterparts. The nonlinear terms are also a good match with predictions from eddy viscosity at wavenumber pairs where the resolvent operator including eddy viscosity is low-rank. The implications for estimation techniques such as spectral linear stochastic estimation and the Kalman filter will be discussed. [Preview Abstract] |
Monday, November 25, 2019 5:29PM - 5:42PM |
P19.00002: 3D diffuser flow predictions using lag models. Rajarshi Biswas, Paul Durbin Lag RANS models employ the usage of a scalar parameter designed to represent the stress-strain misalignment in turbulent flows. The parameter is used to scale the eddy viscosity particularly in the near-wall region. This proves effective for improved prediction of 2D flows with separation. Two novel models, Lag k-epsilon and Lag k-omega are tested for a 3D diffuser configuration. The geometry is parameterized by the inlet aspect ratio(AR). LES predictions show the separation bubble switches position from one wall to the other as AR is increased. Typical RANS formulations such as k-omega SST predict the flow field inaccurately. The lag models are tested for six different ARs and compared against commonly used RANS models. [Preview Abstract] |
Monday, November 25, 2019 5:42PM - 5:55PM |
P19.00003: Linear Navier-Stokes based model for turbulent channels with unstable stratification Anagha Madhusudanan, Simon Illingworth, Ivan Marusic Studies have shown that the linearized Navier-Stokes equations model the coherent large-scale structures in turbulent wall-bounded flows reasonably well. In the present work we aim to understand if this linear model can be extended to study the coherent large-scale structures that have been experimentally and numerically observed in turbulent Rayleigh--B{\'e}nard--Poiseuille flows [e.g., Chauhan et al., 2013, Pirozzoli et al., 2017]. In particular, we concentrate on two features of these structures. First, we look at the wall-normal coherence of the streamwise constant modes. And second, we study the inclination angle of the large-scale structures in these flows. These features are then compared to the available results from numerical and experimental studies. \\ \newline \underline{References} \\ K. Chauhan, N. Hutchins, J. Monty, and I. Marusic. Structure inclination angles in the convective atmospheric surface layer. \textit{Boundary-layer meteorol.}, 147(1):41--50, 2013. \\ S. Pirozzoli, M. Bernardini, R. Verzicco, and P. Orlandi. Mixed convection in turbulent channels with unstable stratification. \textit{J. Fluid Mech.}, 821:482--516, 2017. [Preview Abstract] |
Monday, November 25, 2019 5:55PM - 6:08PM |
P19.00004: Restricting integral length scale growth in triply periodic turbulence simulations Limbert Palomino, Chandru Dhandapani, Jeff Rah, Guillaume Blanquart The 3D periodic box is an essential tool for studying turbulence. It is both an apposite canonical configuration for homogeneous isotropic turbulence and a computationally efficient configuration to simulate. Unfortunately, without an active mean of generating turbulence, the turbulent kinetic energy decays over time due to viscous dissipation. Through the years, various methods of forcing the Navier-Stokes have been proposed to maintain this statistically sta- tionary turbulence, including spectral and linear forcing. Although linear forc- ing schemes fully capture the physics of turbulence, as the simulation evolves in time, the largest eddies in the simulation grow to the order of the computational domain size. The current study characterizes this growth in terms of both the integral length scale and the corresponding energy spectra. Furthermore, we propose a modified linear forcing technique that is analogous to a re-scaling of the computational domain at each time step. This provides more active control over the integral length scale and eddy growth in the simulations. [Preview Abstract] |
Monday, November 25, 2019 6:08PM - 6:21PM |
P19.00005: The Macroscopic Forcing Method and incorporating non-locality into macroscopic models Jessie Liu, Ali Mani The Macroscopic Forcing Method (MFM) (Mani and Park (2019), arXiv:1905.08342) is a newly-developed technique that can be used to determine differential operators associated with turbulence closure. The results of MFM can then be used to improve macroscopic models, i.e. models that describe averaged quantities such as Reynolds-averaged Navier-Stokes (RANS) models. One issue is that often the found differential operators are non-local and may be difficult to incorporate into existing models. We present application of MFM to an example problem involving scalar transport and temporal non-locality. We then present a method for easily incorporating the effects of temporal non-locality into the macroscopic model for the averaged-scalar transport. [Preview Abstract] |
Monday, November 25, 2019 6:21PM - 6:34PM |
P19.00006: On new symmetry-generated RANS models Dario Klingenberg, Martin Oberlack, Dominik Pluemacher We apply new insights into turbulent statistics obtained through Lie-symmetry analysis to the problem of RANS modeling. Symmetries mirror key physical principles from governing equations. The symmetries of the infinite hierarchy of multi-point correlation equations fall into two categories: Symmetries of classical mechanics, which have direct counterparts in the unaveraged Navier-Stokes equations, and statistical ones, which only arise when adopting a statistical view on turbulence. It was shown that those of the latter type encode crucial phenomena, in particular intermittency and non-Gaussianity (Waclawczyk 2014). In a modeling context, symmetries provide constraints on the model equations, because unless the model equations contain precisely the same symmetries as the exact equations, the model is prone to exhibiting nonphysical behavior. We therefore present a new modeling framework that allows algorithmically generating turbulence models based on symmetries. We apply this method to obtain a prototype RANS model that is agreement with not only the classical, but also the statistical symmetries, which existing models fail to accomplish. It turns out to be necessary to introduce as a new model variable an additional velocity field with specific advantageous symmetry properties. [Preview Abstract] |
Monday, November 25, 2019 6:34PM - 6:47PM |
P19.00007: RANS-Equation and the Reynolds Momentum Flux: Homogeneous Decay in a Coriolis Field and a Magnetic Field Charles Petty, Andre Benard The Reynolds-averaged Navier-Stokes (RANS-) equation for constant property Newtonian fluids is an exact unclosed equation due to the explicit appearance of the normalized Reynolds stress (NRS) and the turbulent kinetic energy. The NR-stress is a dyadic-valued operator. This operator must be real, symmetric, non-negative, and non-objective (i.e., the eigenvalues depend on the temporal frame-of-reference). These mathematical properties must be satisfied for all turbulent flows in all inertial and non-inertial frames. Karuna S. Koppula (see, Koppula et al., 2009,2011,2013) developed a non-linear algebraic mapping of the NR-stress into itself that satisfies all of the foregoing mathematical properties (URAPS $=$ Universal, Realizable, Anisotropic Prestress). It is noteworthy that the RANS/URAPS closure predicts that the Coriolis acceleration causes an anisotropic re-distribution of turbulent kinetic energy among the three components of the fluctuating velocity field in a rotating homogeneous decay. The effect of the magnetic component associated with the Lorentz field has a similar effect. Koppula, K. S., A. Benard, and C. A. Petty, 2009, ``Realizable Algebraic Reynolds Stress Closure'', Chem. Eng. Sci., 64, 4611-4624. Koppula, K. S., A. Benard, and C. A. Petty, 2011, ``Turbulent Energy Redistribution in Spanwise Rotating Channel Flows'', Ind. Eng. Chem. Res., 50 (15), 8905-8916. Koppula, K.S., S. Muthu, A. Benard, and C. A. Petty, 2013, ``The URAPS closure for the normalized Reynolds stress'', Physica Scripta, T155. [Preview Abstract] |
Monday, November 25, 2019 6:47PM - 7:00PM |
P19.00008: Multi-level stochastic refinement for turbulent time series and fields Michael Sinhuber, Jan Friedrich, Rainer Grauer, Gregory P. Bewley, Michael Wilczek Many high-dimensional systems exhibit complex spatio-temporal dynamics. Typically, this complexity comes along with strong multi-scale correlations and scale-dependent deviations from Gaussianity, which requires multi-time-multi-point statistics for a full characterization. In many practical cases, obtaining sufficiently resolved data is prohibitively expensive or even impossible. While a statistical characterization of such systems is often sufficient, many applications where turbulence plays a key role require a complete spatio-temporal realization of the system. Examples range from modeling particle propagation in fusion plasmas to wind field modeling for wind energy applications. To address this challenge, we develop a stochastic refinement method that generates finely resolved data sets from readily available, coarsely sampled turbulence data as well as synthetic datasets within the scope of classical turbulence models. This is done by utilizing scale-Markovian properties of turbulence and scale-dependent three-point velocity statistics. We test our approach both for wind tunnel data from the VDTT at the Max-Planck-Institute for Dynamics and Self-Organization in G\"{o}ttingen as well as for simulated turbulent fields based on classical turbulence 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. |
© 2020 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
1 Research Road, Ridge, NY 11961-2701
(631) 591-4000
Office of Public Affairs
529 14th St NW, Suite 1050, Washington, D.C. 20045-2001
(202) 662-8700