Engineering topics

Reynolds, W. C.

Publications and source records attributed to Reynolds, W. C..

At least 37 records · Page 2

NASA NTRS1987

Advanced in turbulence physics and modeling by direct numerical simulations

The advent of direct numerical simulations of turbulence has opened avenues for research on turbulence physics and turbulence modeling. Direct numerical simulation provides values for anything that the scientist or modeler would like to know about the flow. An overview of some recent advances in the physical understanding of turbulence and in turbulence modeling obtained through such simulations is presented.

Reynolds, W. C.↗

NASA NTRS1987

Passive scalar entrainment and mixing in a forced, spatially-developing mixing layer

Numerical simulations are performed for the forced, spatially-developing plane mixing layer in two and three dimensions. Transport of a passive scalar field is included in the computation. This, together with the allowance for spatial development in the simulations, affords the opportunity for study of the asymmetric entrainment of irrotational fluid into the layer. The inclusion of a passive scalar field provides a means for simulating the effect of this entrainment asymmetry on the generation of 'products' from a 'fast' chemical reaction. Further, the three-dimensional simulations provide useful insight into the effect of streamwise structures on these entrainment and 'fast' reaction processes. Results from a two-dimensional simulation indicate 1.22 parts high-speed fluid are entrained for every one part low-speed fluid. Inclusion of streamwise vortices at the inlet plane of a three-dimensional simulation indicate a further increase in asymmetric entrainment - 1.44:1. Results from a final three-dimensional simulation are presented. In this case, a random velocity perturbation is imposed at the inlet plane. The results indicate the 'natural' development of the large spanwise structures characteristic of the mixing layer.

Lowery, P. S.↗

NASA NTRS1985

Unsteady turbulent boundary-layer experiments with rapidly changing free-stream conditions

Selected data sets produced by the Stanford Unsteady Turbulent Boundary Layer Program are compared to analytic solutions of the Stokes equations for different kinds of imposed, organized unsteadiness. Good agreement is found, qualifying the facility as one that faithfully produces flows of different prescribed unsteady conditions. Data sets from the facility are therefore considered especially suitable for comparison with predictions of turbulence models for unsteady flows. New data describing the development of an unsteady separation bubble are presented.

Brereton, G. J.↗

NASA NTRS1985

On the structure of homogeneous turbulence

Full turbulence simulation has been conducted of homogeneous turbulence subject to irrotational strains and under relaxation from these. Effects of the mode and strength of the imposed mean strain rate on the Reynolds stress, dissipation and vorticity fields are analyzed, yielding new insight. Tensorial models for the relaxation process are proposed for the vorticity dissipation and Reynolds stress anisotropies.

Lee, M. J.↗

NASA NTRS1985

Unsteady turbulent boundary layers - Some effects of abrupt free-stream velocity changes

The response of a turbulent boundary layer to abrupt changes in free-stream velocity is studied. Streamwise velocity measurements are examined during the abrupt events and after them, as the flow relaxes to its equilibrium condition. The measured velocity field is found to be in good agreement with the analytic solution of the Stokes equation for boundary conditions describing abrupt changes. While the flow recovers to its equilibrium state, changes in integral, boundary-layer parameters scale on the local free-stream time of flight.

Brereton, G. J.↗

NASA NTRS1984

Discrete orthogonal function expansions for non-uniform grids using the fast Fourier transform

A technique for applying discrete Fourier series to infinite domains is presented. The technique uses mappings designed to minimize truncation error and can be applied to solve mixed initial boundary value problems among others. The method is alias-free and yields consistent differentiation and integration operators. The mapping-induced truncation error is explicitly expressible and small in nearly all cases of interest. The method is illustrated for three problems involving convection, diffusion, and vortex interaction.

Cain, A. B.↗

NASA NTRS1983

Improved turbulence models based on large eddy simulation of homogeneous, incompressible turbulent flows

The physical bases of large eddy simulation and subgrid modeling are studied. A subgrid scale similarity model is developed that can account for system rotation. Large eddy simulations of homogeneous shear flows with system rotation were carried out. Apparently contradictory experimental results were explained. The main effect of rotation is to increase the transverse length scales in the rotation direction, and thereby decrease the rates of dissipation. Experimental results are shown to be affected by conditions at the turbulence producing grid, which make the initial states a function of the rotation rate. A two equation model is proposed that accounts for effects of rotation and shows good agreement with experimental results. In addition, a Reynolds stress model is developed that represents the turbulence structure of homogeneous shear flows very well and can account also for the effects of system rotation.

Bardino, J.↗

NASA NTRS1983

Simulation of turbulent shear flows at Stanford and NASA-Ames - What can we do and what have we learned?

The capabilities and limitations of large eddy simulation (LES) and full turbulence simulation (FTS) are outlined. It is pointed out that LES, although limited at the present time by the need for periodic boundary conditions, produces large-scale flow behavior in general agreement with experiments. What is more, FTS computations produce small-scale behavior that is consistent with available experiments. The importance of the development work being done on the National Aerodynamic Simulator is emphasized. Studies at present are limited to situations in which periodic boundary conditions can be applied on boundaries of the computational domain where the flow is turbulent.

Reynolds, W. C.↗

NASA NTRS1983

Transient response of a turbulent boundary layer

A unique feature of the present ensemble-averaged measurements of a turbulent boundary layer's transient response to a spontaneous change in the free stream velocity distribution, is that the test boundary layer is a standard, steady, flat plate turbulent boundary layer at the entrance to the unsteady region, and is then subjected to sudden changes in free stream velocity distribution in the test section. These water tunnel tests were controlled by minicomputer. It is noted that the boundary layer development was relatively slow, with a characteristic time that was greater than the free stream time-of-flight by a factor of as much as 3. Response varied dramatically across the boundary layer, and the evolution of the turbulent stress field occurred on the same time scale as that of the ensemble-averaged velocity field.

Parikh, P. G.↗

NASA NTRS1982

Dynamics of an unsteady turbulent boundary layer

The present investigation is concerned with the response of a well-defined, steady, flatplate turbulent boundary layer subjected subsequently to two types of oscillations of the free-stream velocity in the test section. It is found that the mean velocity profile in the boundary layer is unaffected by imposed free-stream oscillations in the range of frequencies employed. It is the same as the one measured with a free-stream velocity distribution held steady at its mean value. The behavior of the mean velocity field is a consequence of two observations. The time-averaged Reynolds stress distribution across the boundry layer is unaffected by the imposed oscillations. The Reynolds stresses arising from the organized velocity fluctuations under imposed oscillatory conditions are negligible compared to Reynolds stress due to the random fluctuations.

Parikh, P. G.↗

NASA NTRS1981

A three-dimensional simulation of transition and early turbulence in a time-developing mixing layer

The physics of the transition and early turbulence regimes in the time developing mixing layer was investigated. The sensitivity of the mixing layer to the disturbance field of the initial condition is considered. The growth of the momentum thickness, the mean velocity profile, the turbulence kinetic energy, the Reynolds stresses, the anisotropy tensor, and particle track pictures of computations are all examined in an effort to better understand the physics of these regimes. The amplitude, spectrum shape, and random phases of the initial disturbance field were varied. A scheme of generating discrete orthogonal function expansions on some nonuniform grids was developed. All cases address the early or near field of the mixing layer. The most significant result shows that the secondary instability of the mixing layer is produced by spanwise variations in the straining field of the primary vortex structures.

Cain, A. B.↗

NASA NTRS1981

Dynamic behavior of an unsteady trubulent boundary layer

Experiments on an unsteady turbulent boundary layer are reported in which the upstream portion of the flow is steady (in the mean) and in the downstream region, the boundary layer sees a linearly decreasing free stream velocity. This velocity gradient oscillates in time, at frequencies ranging from zero to approximately the bursting frequency. For the small amplitude, the mean velocity and mean turbulence intensity profiles are unaffected by the oscillations. The amplitude of the periodic velocity component, although as much as 70% greater than that in the free stream for very low frequencies, becomes equal to that in the free stream at higher frequencies. At high frequencies, both the boundary layer thickness and the Reynolds stress distribution across the boundary layer become frozen. The behavior at higher amplitude is quite similar. At sufficiently high frequencies, the boundary layer thickness remains frozen at the mean value over the oscillation cycle, even though flow reverses near the wall during a part of the cycle.

Parikh, P. G.↗

NASA NTRS1981

Mixing of a passive scalar in isotropic and sheared homogeneous turbulence

In order to calculate the velocity and scalar fields, the three dimensional, time-dependent equations of motion and the diffusion equation were solved numerically. The following cases were treated: isotropic, homogeneous turbulence with decay of a passive scalar; and homogeneous turbulent shear flow with a passive scalar whose mean varies linearly in the spanwise direction. The solutions were obtained at relatively low Reynolds numbers so that all of the turbulent scales could be resolved without modeling. Turbulent statistics such as integral length scales, Taylor microscales, Kolmogorov length scale, one- and two-point correlations of velocity-velocity and velocity-scalar, turbulent Prandtl/Schmidt number, r.m.s. values of velocities, the scalar quantity and pressure, skewness, decay rates, and decay exponents were calculated. The results are compared with the available expermental results, and good agreement is obtained.

Shirani, E.↗

NASA NTRS1981

Numerical simulation of a compressible homogeneous, turbulent shear flow

A direct, low Reynolds number, numerical simulation was performed on a homogeneous turbulent shear flow. The full compressible Navier-Stokes equations were used in a simulation on the ILLIAC IV computer with a 64,000 mesh. The flow fields generated by the code are used as an experimental data base, to examine the behavior of the Reynols stresses in this simple, compressible flow. The variation of the structure of the stresses and their dynamic equations as the character of the flow changed is emphasized. The structure of the tress tensor is more heavily dependent on the shear number and less on the fluctuating Mach number. The pressure-strain correlation tensor in the dynamic uations is directly calculated in this simulation. These correlations are decomposed into several parts, as contrasted with the traditional incompressible decomposition into two parts. The performance of existing models for the conventional terms is examined, and a model is proposed for the 'mean fluctuating' part.

Feiereisen, W. J.↗

NASA NTRS1981

Large-eddy simulation of the shear-free turbulent boundary layer

The shear-free turbulent boundary layer is calculated by the large-eddy simulation technique. The filtered Navier-Stokes equations are used; the method of integration employs Fourier expansions in the homogeneous directions and finite differences in the cross-stream direction. Results indicate that the simulation is capable of predicting the primary Reynolds-number effects.

Biringen, S.↗

NASA NTRS1981

International Conference on Numerical Methods in Fluid Dynamics, 7th, Stanford University, Stanford and Moffett Field, CA, June 23-27, 1980, Proceedings

Topics discussed include polygon transformations in fluid mechanics, computation of three-dimensional horseshoe vortex flow using the Navier-Stokes equations, an improved surface velocity method for transonic finite-volume solutions, transonic flow calculations with higher order finite elements, the numerical calculation of transonic axial turbomachinery flows, and the simultaneous solutions of inviscid flow and boundary layer at transonic speeds. Also considered are analytical solutions for the reflection of unsteady shock waves and relevant numerical tests, reformulation of the method of characteristics for multidimensional flows, direct numerical simulations of turbulent shear flows, the stability and separation of freely interacting boundary layers, computational models of convective motions at fluid interfaces, viscous transonic flow over airfoils, and mixed spectral/finite difference approximations for slightly viscous flows.

Reynolds, W. C.↗

NASA NTRS1981

Dynamic behavior of an unsteady turbulent boundary layer

Experiments on an unsteady turbulent boundary layer are reported in which the upstream portion of the flow is steady (in the mean) and in the downstream region, the boundary layer sees a linearly decreasing free stream velocity. This velocity gradient oscillates in time, at frequencies ranging from zero to approximately the bursting frequency. For the small amplitude, the mean velocity and mean turbulence intensity profiles are unaffected by the oscillations. The amplitude of the periodic velocity component, although as much as 70 percent greater than that in the free stream for very low frequencies, becomes equal to that in the free stream at higher frequencies. At high frequencies, both the boundary layer thickness and the Reynolds stress distribution across the boundary layer become frozen. The behavior at higher amplitude is quite similar. At sufficiently high frequencies, the boundary layer thickness remains frozen at the mean value over the oscillation cycle, even though flow reverses near the wall during a part of the cycle. Previously announced in STAR as N81-28392

Parikh, P. G.↗

NASA NTRS1980

Improved subgrid-scale models for large-eddy simulation

The paper analyzes models for subgrid-scale turbulence. The analysis indicates that there is sufficient information in the resolved scales to determine some characteristics of the complete flow field. The kinetic energy of the small-scale motions can be decomposed into two parts: one results from the large scales and is correlated with them, and the other part is uncorrelated which leads to a two-component eddy-viscosity model. The 'production equals dissipation' argument does not apply to the small scales in the decay of turbulence because it does not account for the uncorrelated component. The exchange between the large and small scales takes place mainly between the smallest scales of the former and the largest scales of the latter; this argument is the basis of a new model shown to be superior to the Smagorinsky model (1963).

Bardina, J.↗