Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “CHANNEL FLOW”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 37 records · Page 2
All sourcesNASA NTRSDOE OSTI
NASA NTRS1991

Numerical simulation of spatially-evolving instability in plane channel flow

The spatial stability of plane channel flow is analyzed using a three-dimensional, time-dependent spectral/finite difference code (Danabasoglu et al., 1990) which integrates numerically the Navier-Stokes equations. The study centers on inflow disturbance amplitudes effects on the secondary instability. The resolution requirements along the spacewise direction, which become critical before the breakdown stage, are of particular interest. A direct comparison is made with the experiments of Nishioka et al. (1980).

Danabasoglu, G.↗
NASA NTRS1995

A new non-eddy viscosity subgrid-scale model and its application to channel flow

To date, most large-eddy simulations (LES) have been carried out with eddy viscosity subgrid scale (SGS) models, with only a few exceptions that used the mixed model. Even though the assumptions behind Smagorinsky's model are rather stringent, it has been applied successfully to a variety of turbulent flows. This success is attributed to the ability of eddy viscosity models to drain energy from large scales, thus simulating the dissipative nature of turbulence. Most SGS models are absolutely dissipative, i.e. they remove energy from the large scales at every instant. However, SGS stresses may transfer energy back to the large scales intermittently; this reverse transfer or backscatter is especially important in geophysical flows and in transition. In a fully developed channel flow, there is reverse flow of energy from small to large scales near the walls, but eddy viscosity models are unable to account for this important feature. The dynamic localization eddy viscosity model of Ghosal et al. (1995) allows backscatter by co-evolving an auxiliary equation for the SGS energy; however, the computational cost is considerably larger than for conventional SGS models (Cabot 1994). In this report, a new non-eddy viscosity model based on local approximation of total quantities in terms of filtered ones is introduced; the scale similarity model of Bardina (1983) is a special case of this model. This procedure does not require the assumption of homogeneity, permits backscatter of energy from small to large scales, and is readily implemented in finite difference codes. The results of applying the proposed model to second order finite volume simulation of plane channel flow at high Reynolds numbers (Re(sub b) = 38,000) is described in this report. Greater emphasis is placed on the high Reynolds number flow since it provides a more rigorous test of the SGS model and its potential application. The results are compared to ones produced by the conventional and dynamic Smagorinsky models and the spectral LES of Piomelli (1993).

Shah, K. B.↗
NASA NTRS1984

Direct numerical simulation of curved turbulent channel flow

Low Reynolds number, mildly curved, turbulent channel flow has been simulated numerically without subgrid scale models. A new spectral numerical method developed for this problem was used, and the computations were performed with 2 million degrees of freedom. A variety of statistical and structural information has been extracted from the computed flow fields. These include mean velocity, turbulence stresses, velocity skewness, and flatness factors, space time correlations and spectra, all the terms in the Reynolds stress balance equations, and contour and vector plots of instantaneous velocity fields. The effects of curvature on this flow were determined by comparing the concave and convex sides of the channel. The observed effects are consistent with experimental observations for mild curvature. The most significant difference in the turbulence statistics between the concave and convex sides was in the Reynolds shear stress. This was accompanied by significant differences in the terms of the Reynolds shear stress balance equations. In addition, it was found that stationary Taylor-Gortler vortices were present and that they had a significant effect on the flow by contributing to the mean Reynolds shear stress, and by affecting the underlying turbulence.

Moser, R. D.↗
NASA NTRS1984

Final stages of transition to turbulence in plane channel flow

This paper involves a numerical simulation of the final stages of transition to turbulence in plane channel flow at a Reynolds number of 1500. Three-dimensional incompressible Navier-Stokes equations are numerically integrated to obtain the time evolution of two- and three-dimensional finite-amplitude disturbances. Computations are performed on the CYBER-203 vector processor for a 32 x 51 x 32 grid. Solutions indicate the existence of structures similar to those observed in the laboratory and characteristics of the various stages of transition that lead to final breakdown. In particular, evidence points to the formation of a upside-down-V-shaped vortex and the subsequent system of horseshoe vortices inclined to the main flow direction as the primary elements of transition. Details of the resulting flow field after breakdown indicate the evolution of streaklike formations found in turbulent flows. Although the flow field does approach a steady state (turbulent channel flow), the introduction of subgrid-scale terms seems necessary to obtain fully developed turbulence statistics.

Biringen, S.↗
NASA NTRS1983

A numerical simulation of transition in plane channel flow

This paper involves a numerical simulation of the final stages of transition to turbulence in plane channel flow at a Reynolds number of 7500. Three-dimensional, incompressible Navier-Stokes equations are numerically integrated to obtain the time-evolution of two- and three-dimensional finite-amplitude disturbances. Computations are performed on the CYBER-203 vector processor for a 32 x 33 x 32 grid. Solutions indicate the existence of structures similar to those observed in the laboratory and which are characteristic of various stages of transition that lead to final breakdown. Details of the resulting flow field after breakdown indicate the evolution of streak-like formations found in turbulent flows. Although the flow field does approach a steady-state (turbulent channel flow), implementation of subgrid-scale terms are necesary to obtain proper turbulent statistics.

Biringen, S.↗
NASA NTRS1991

On the effect of riblets in fully developed laminar channel flows

The effect of longitudinal riblet surfaces on viscous drag in fully developed laminar channel flows was investigated. Unlike turbulent flows, drag reduction was not obtained in the laminar flows. Results were independent of Reynolds number. Wall-shear rates on most regions of the cross-sectional perimeter of riblets were smaller than that of corresponding plane channel flow even though the net drag was increased.

Choi, Haecheon↗
NASA NTRS1994

Instability of streamwise vortices in plane channel flows

We present analysis and numerical experiments on the instability of streamwise vortices in 'minimal channel' flows and argue that this instability is a key feature in the observed intermittent cycle of formation, break-up, and re-formation of these structures. The base flow is a three-component, two-dimensional pair of counter-rotating rolls with axes aligned along the direction of the mean shear. While it is not a steady solution to the Navier-Stokes equations, we show numerically that this flow is unstable on a fast time scale to a secondary, three-dimensional Floquet mode. The growth of the secondary instability does not saturate in a new equilibrium, but continues until highly unstable local shear layers form and the entire flow breaks down into turbulence. Our analysis is motivated in part by the strong similarities between the intermittent turbulent cycle in minimal channel flows and one studied, both experimentally and in computations, in Couette-Taylor flow.

Coughlin, K.↗
NASA NTRS1993

Large eddy simulations of time-dependent and buoyancy-driven channel flows

The primary goal of this work has been to assess the performance of the dynamic SGS model in the large eddy simulation (LES) of channel flows in a variety of situations, viz., in temporal development of channel flow turned by a transverse pressure gradient and especially in buoyancy-driven turbulent flows such as Rayleigh-Benard and internally heated channel convection. For buoyancy-driven flows, there are additional buoyant terms that are possible in the base models, and one objective has been to determine if the dynamic SGS model results are sensitive to such terms. The ultimate goal is to determine the minimal base model needed in the dynamic SGS model to provide accurate results in flows with more complicated physical features. In addition, a program of direct numerical simulation (DNS) of fully compressible channel convection has been undertaken to determine stratification and compressibility effects. These simulations are intended to provide a comparative base for performing the LES of compressible (or highly stratified, pseudo-compressible) convection at high Reynolds number in the future.

Cabot, William H.↗
NASA NTRS1993

Optimal feedback control of turbulent channel flow

Feedback control equations were developed and tested for computing wall normal control velocities to control turbulent flow in a channel with the objective of reducing drag. The technique used is the minimization of a 'cost functional' which is constructed to represent some balance of the drag integrated over the wall and the net control effort. A distribution of wall velocities is found which minimizes this cost functional some time shortly in the future based on current observations of the flow near the wall. Preliminary direct numerical simulations of the scheme applied to turbulent channel flow indicates it provides approximately 17 percent drag reduction. The mechanism apparent when the scheme is applied to a simplified flow situation is also discussed.

Bewley, Thomas↗
NASA NTRS1988

Turbulent structure in low-concentration drag-reducing channel flows

A two-component laser-Doppler velocimeter was used to obtain simultaneous measurements of the velocity components parallel and normal to the wall in two fully developed well-mixed low-concentration drag-reducing channel flows and one turbulent channel flow. For the drag-reducing flows, the average time between bursts was found to increase. Although the basic structure of the fundamental momentum transport event is shown to be the same in these drag-reducing flows, the lower-threshold Reynolds-stress-producing motions were found to be damped, while the higher-threshold motions were not. It is suggested that some strong turbulent motions are needed to maintain extended polymer molecules, which produce a solution with properties that can damp lower threshold turbulence and thereby reduce viscous drag.

Luchik, T. S.↗
NASA NTRS1988

On the generation of mean flows by the interaction of Goertler vortices and Tollmien-Schlichting waves in curved channel flows

There are many fluid flows where the onset of transition can be caused by different instability mechanisms which compete in the nonlinear regime. The interaction of a centrifugal instability mechanism with the viscous mechanism which causes Tollmein-Schlichting waves is discussed. The interaction between these modes can be strong enough to drive the mean state; the interaction is investigated in the context of curved channel flows so as to avoid difficulties associated with boundary layer growth. Essentially it was found that the mean state adjusts itself so that any modes present are neutrally stable even at finite amplitude. In the first instance, the mean state driven by a vortex of short wavelength in the absence of Tollmein-Schlichting waves is considered. It is shown that for a given channel curvature and vortex wavelength, there is an upper limit to the mass flow rate which the channel can support as the pressure gradient is increased. When the Tollmein-Schlichting waves are present, then the nonlinear differential equation to determine the mean state is modified.

Bassom, Andrew P.↗
NASA NTRS1995

Wavenumber selection for small-wavelength Goertler vortices in curved channel flows

The problem of wavenumber selection for fully nonlinear, small-wavelength Goertler vortices in a curved channel flow is considered. These types of Goertler vortices were first considered by Hall & Lakin (1988) for an external boundary layer flow. They proved particularly amenable to asymptotic description, it was possible to consider vortices large enough so that the mean flow correction driven by them is as large as the basic state, and this prompted the authors to consider them in a curved channel flow as an initial application of the phase-equation approach to Goertler vortices. This involves the assumption that the phase variable of these Goertler vortices varies on slow spanwise and time scales, then an analysis of both inside and outside the core region, to which vortex activity is restricted, leads to a system of partial differential equations which can be solved numerically for the wavenumber. The authors consider in particular the effect on the wavenumber of the outer channel wall varying on the same slow spanwise scale as the phase variable.

Dando, Andrew↗
NASA NTRS1993

Dynamic localization and second-order subgrid-scale models in large eddy simulations of channel flow

The objective here is to test the Dynamic Localization (DL) model in a wall-bounded channel flow for numerical stability and accuracy of results. Algebraic stress models suggest that the model for the residual subgrid-scale (SGS) Reynolds stress and scalar flux should generally have terms comprising most of the unique products of the resolved strain (S) and rotation (R) tensors with S and the resolved scalar gradient. The standard dynamic SGS model uses a simple (Smagorinsky) base model for the residual Reynolds stress, which is made proportional to S, and down-gradient base models for residual scalar fluxes; these correspond to the lowest, 'first-order' terms in algebraic stress models. Temporal scaling terms in these base models are formed from the magnitude of the resolved strain rate. While this is appropriate for simple shear flows, it may not be appropriate for more complicated flows (relevant to geophysical and astrophysical problems) that include any combination of shear, rotation, buoyancy, etc. On the other hand, the coefficient in the dynamic SGS model readily adjusts itself to different flow conditions and may adequately take account of these effects without the need for more complicated base models. Cabot (1993) has begun to test the dynamic SGS model in buoyant flows (Rayleigh-Benard and internally heated convection) with and without buoyancy terms explicitly included in the scaling terms of the base model; no great differences were found in large eddy simulation (LES) results for the different base model scalings. The second objective in this work is to test base models with additional, 'second-order' terms (e.g., S(sup 2) and RS for the residual Reynolds stress). These terms have been found to improve large-scale flow predictions by kappa-epsilon models in the presence of rotation and shear. Second-order base models will be tested here in the LES of channel flow with and without solid-body rotation and compared with results from the standard first-order base models to determine if there are significant differences or improvements in results that would warrant the added complexity of the second-order base models.

Cabot, William H.↗
NASA NTRS1982

Study of incompressible turbulent channel flow by large eddy simulation

Incompressible turbulent channel flow is investigated by large eddy simulation using improved numerical methods and boundary conditions. In downstream and spanwise directions, cyclic boundary conditions are imposed for velocity and pressure, and two types of boundary conditions near the wall are used and compared. One type is based on the logarithmic law of the mean velocity near the wall and has a slip boundary condition where the molecular-viscous term is neglected. The other type is based on a no-slip boundary condition, where fine mesh spacing near the wall is used to take account of the molecular viscosity. Although the present study employs a coarse mesh (16 x 16 x 21), its results are in good agreement with those of Moin and Kim (1981).

Horiuti, K.↗
NASA NTRS1989

On the shape and dynamics of wall structures in turbulent channel flow

Turbulence-producing events in turbulent channel flow were found to be predominantly associated with a symmetric vortical structures rather than pairs of counter-rotating structures. An asymmetry-preserving averaging scheme was devised, allowing a picture of the 'average' structure that more closely resembles the instantaneous one to be obtained. In addition, these structures were found to persist for long distances with little change while convecting downstream.

Guezennec, Yann G.↗
NASA NTRS1987

A method for calculating turbulent boundary layers and losses in the flow channels of turbomachines

An interactive inviscid core flow-boundary layer method is presented for the calculation of turbomachine channel flows. For this method, a one-dimensional inviscid core flow is assumed. The end-wall and blade surface boundary layers are calculated using an integral entrainment method. The boundary layers are assumed to be collateral and thus are two-dimensional. The boundary layer equations are written in a streamline coordinate system. The streamwise velocity profiles are approximated by power law profiles. Compressibility is accounted for in the streamwise direction but not in the normal direction. Equations are derived for the special cases of conical and two-dimensional rectangular diffusers. For these cases, the assumptions of a one-dimensional core flow and collateral boundary layers are valid. Results using the method are compared with experiment and good quantitative agreement is obtained.

Schumann, Lawrence F.↗