Showing papers by "Parviz Moin published in 1988"

Eddies, streams, and convergence zones in turbulent flows

Abstract: Recent studies of turbulent shear flows have shown that many of their important kinematical and dynamical properties can be more clearly understood by describing the flows in terms of individual events or streamline patterns These events or flow regions are studied because they are associated with relatively large contributions to certain average properties of the flow, for example kinetic energy, Reynolds stress, or to particular processes in the flow, such as mixing and chemical reactions, which may be concentrated at locations where streamlines converge for fast chemical reactions (referred to as convergence or C regions), or in recirculating eddying regions for slow chemical reactions The aim of this project was to use the numerical simulations to develop suitable criteria for defining these eddying or vortical zones The C and streaming (S) zones were defined in order to define the whole flow field It is concluded that homogeneous and sheared turbulent flow fields are made up of characteristic flow zones: eddy, C, and S zones A set of objective criteria were found which describe regions in which the streamlines circulate, converge or diverge, and form high streams of high velocity flow

Reynolds-stress and dissipation-rate budgets in a turbulent channel flow

Abstract: The budgets for the Reynolds stresses and for the dissipation rate of the turbulence kinetic energy are computed using direct simulation data of a turbulent channel flow. The budget data reveal that all the terms in the budget become important close to the wall. For inhomogeneous pressure boundary conditions, the pressure-strain term is split into a return term, a rapid term, and a Stokes term. The Stokes term is important close to the wall. The rapid and return terms play different roles depending on the component of the term. A split of the velocity pressure-gradient term into a redistributive term and a diffusion term is proposed, which should be simpler to model. The budget data is used to test existing closure models for the pressure-strain term, the dissipation rate, and the transport rate. In general, further work is needed to improve the models.

Model consistency in large eddy simulation of turbulent channel flows

Abstract: Combinations of filters and subgrid scale stress models for large eddy simulation of the Navier-Stokes equations are examined by a priori tests and numerical simulations. The structure of the subgrid scales is found to depend strongly on the type of filter used, and consistency between model and filter is essential to ensure accurate results. The implementation of consistent combinations of filter and model gives more accurate turbulence statistics than those obtained in previous investigations in which the models were chosen independently from the filter. Results and limitations of the a priori test are discussed. The effect of grid refinement is also examined.

Stochastic estimation of organized turbulent structure - Homogeneous shear flow

Abstract: A generalization of the conditional-eddy concept is proposed in which the conditional event specifies the local kinematic state in terms of the velocity and the deformation. Results are presented for stochastically estimated conditional eddies given the local kinematics. The equation governing the probability density function of a kinematic state has been derived for constant-property incompressible flow, providing a link between coherent flow structures corresponding to the conditional eddies and the modelling of turbulent transport. The primary contributions to the second-quadrant and fourth-quadrant Reynolds-stress events in homogeneous shear flow are shown to come from flow induced through the 'legs' and close to the 'heads' of upright and inverted 'hairpins', respectively.

Karhunen-Loeve expansion of Burgers' model of turbulence

Abstract: Characteristics of the Karhunen–Loeve expansion of a strongly inhomogeneous random process possessing small viscous length scales and a large outer scale have been investigated in relation to the application of the expansion to turbulent flow fields. Monte Carlo simulations of a randomly forced Burgers’ equation with zero velocity boundary conditions generate the random process numerically and the Karhunen–Loeve (KL) eigenfunctions and the eigenvalue spectra are computed for different Reynolds numbers. The eigenfunctions possess thin viscous boundary layers at the walls and are independent of Reynolds number in the core, where the random process is quasihomogeneous. The eigenfunctions and eigenvalues of the outer, large scale motions obey a principle of Reynolds number similarity. Eigenvalue spectra contain much of the energy in the first few modes, but they are as broad as ordinary trigonometric power spectra. The rate at which the expansion converges to within 90% of the total energy decreases with increasing Reynolds numbers and the expansion of the mean plus the fluctuation converges more rapidly than the expansion of the fluctuation alone.

Direct numerical simulation of a three-dimensional turbulent boundary layer

Abstract: The effects of transverse strain on an initially two‐dimensional turbulent boundary layer are studied in a direct numerical simulation of a planar channel flow with impulsively started transverse pressure gradient. Consistent with experiments in three‐dimensional boundary layers, the simulation shows a decrease in the Reynolds shear stress with increasing transverse strain. Also, the directions of the Reynolds shear stress vector and the mean velocity gradient vector were found to differ. In addition, the simulation shows a drop in the turbulent kinetic energy. Terms in the Reynolds stress transport equations were computed. The balances indicate that the decrease in turbulent kinetic energy is a result of a decrease in turbulence production, along with an increase in turbulent dissipation. Intuitive reasoning and current turbulence models would predict an increase in kinetic energy along with increases in production and dissipation rates as a result of increased mean‐flow strain rate. Later in the evolution of the flow, both turbulence production and dissipation increase.

Ejection mechanisms in the sublayer of a turbulent channel

Abstract: The structure of the vorticity field in the viscous wall layer of a turbulent channel is studied by examining the results of a fully resolved direct numerical simulation. It is shown that this region is dominated by intense three-dimensional shear layers in which the dominant vorticity component is spanwise. The advection and reproduction processes of these structures are examined and shown to be consistent with the classical generation mechanism for two-dimensional Tollmien-Schlichting waves. This process is fundamentally different from the usually accepted mechanism involving hairpin vortices.

Dynamics of coherent structures in a plane mixing layer

Abstract: An incompressible, time developing 3-D mixing layer with idealized initial conditions was simulated numerically. Consistent with the suggestions from experimental measurements, the braid region between the dominant spanwise vortices or rolls develops longitudinal vortices or ribs, which are aligned upstream and downstream of a roll and produce spanwise distortion of the rolls. The process by which this distortion occurs is explained by studying a variety of quantities of dynamic importance (e.g., production of enstrophy, vortex stretching). Other quantities of interest (dissipation, helicity density) are also computed and discussed. The currently available simulation only allows the study of the early evolution (before pairing) of the mixing layer. New simulations in progress will relieve this restriction.

Annual Research Briefs, 1987

Abstract: Lagrangian techniques have found widespread application to the prediction and understanding of turbulent transport phenomena and have yielded satisfactory results for different cases of shear flow problems. However, it must be kept in mind that in most experiments what is really available are Eulerian statistics, and it is far from obvious how to extract from them the information relevant to the Lagrangian behavior of the flow; in consequence, Lagrangian models still include some hypothesis for which no adequate supporting evidence was until now available. Direct numerical simulation of turbulence offers a new way to obtain Lagrangian statistics and so verify the validity of the current predictive models and the accuracy of their results. After the pioneering work of Riley (Riley and Patterson, 1974) in the 70's, some such results have just appeared in the literature (Lee et al, Yeung and Pope). The present contribution follows in part similar lines, but focuses on two particle statistics and comparison with existing models.

Dynamical interpretation of conditional patterns

Abstract: While great progress is being made in characterizing the 3-D structure of organized turbulent motions using conditional averaging analysis, there is a lack of theoretical guidance regarding the interpretation and utilization of such information. Questions concerning the significance of the structures, their contributions to various transport properties, and their dynamics cannot be answered without recourse to appropriate dynamical governing equations. One approach which addresses some of these questions uses the conditional fields as initial conditions and calculates their evolution from the Navier-Stokes equations, yielding valuable information about stability, growth, and longevity of the mean structure. To interpret statistical aspects of the structures, a different type of theory which deals with the structures in the context of their contributions to the statistics of the flow is needed. As a first step toward this end, an effort was made to integrate the structural information from the study of organized structures with a suitable statistical theory. This is done by stochastically estimating the two-point conditional averages that appear in the equation for the one-point probability density function, and relating the structures to the conditional stresses. Salient features of the estimates are identified, and the structure of the one-point estimates in channel flow is defined.