Engineering topics
Johnston, J. P.
Publications and source records attributed to Johnston, J. P..
Experiments on near-wall structure of three-dimensional boundary layers
Investigations of three-dimensional turbulent boundary layers have shown basic differences between two- and three-dimensional flows. These differences can significantly impact the modeling of three-dimensional flows since many flow models are based on results from two-dimensional boundary layers. In many cases the shear stress vector direction has been shown to lag relative to the direction of the mean velocity gradient as the cross flow grows downstream. Coincidence of these vectors is necessary for a scalar eddy viscosity modeling assumption. A second effect is a reduction in magnitude of the shear stress and/or the shear stress to turbulence energy ratio, a(sub 1). This reduction has been observed in several experiments. Recent numerical simulations also indicate wall-layer structural differences between two- and three-dimensional boundary layers. The differences in structure between two- and three-dimensional boundary layers was also addressed in the experiment of Littell & Eaton. The experiment used two-point correlations to investigate the vortical structures in a three dimensional boundary layer on a spinning disk. It was found that each sign of longitudinal vortex is equally likely to exist, but one sign of vorticity is associated with a structure which is better at producing ejections. The goal of the current investigation is to study the structure of the inner layers. Among other questions, the differences between the effects deduced from the three-dimensional flow simulations and the effects seen in experiments can be examined. The research concentrates on the structure of the wall-layer through flow visualization and direct turbulence measurements down to y(+) = 5.
The effects of stabilizing and destabilizing longitudinal curvature on the structure of turbulent, two-stream mixing layers
The construction and development of the multi-component traversing system and associated control hardware and software are presented. A hydrogen bubble/laser sheet flow visualization technique was developed to visually study the characteristics of the mixing layers. With this technique large-scale rollers arising from the Taylor-Gortler instability and its interaction with the primary Kelvin-Helmholtz structures can be studied.
Film cooling on a convex wall: Heat transfer and hydrodynamic measurements for full and partial coverage
Turbine-blade cooling is an important issue for high-efficiency turbine engines, and discrete-hole injection is widely used as a cooling method. In the present study, detailed measurements were made of the heat transfer and hydrodynamics of a film-cooled flow on a convex wall, both for full and partial coverage. Two important parameters were altered: the blowing ratio, m, and the number of rows of injection holes. Three values of m were tested: m = 0.2, 0.4, and 0.6. In the blown region, m = 0.4 results in the lowest Stanton numbers of the three blowing ratios tested. This indicates that the value of m = 0.4 is near optimum on the convex wall from the point of view of cooling effect by injection. In the recovery region, Stanton numbers gradually approach the no injection values. Although the heat-transfer behavior during recovery from injection looks relatively complicated, the behavior of Stanton number can be explained in terms of two mechanisms: recovery from the thermal effect of injection and recovery from the turbulence augmentation. This interpretation of the data is supported by the hydrodynamic and temperture-profile measurements. For partial blowing cases, the data follow the full-coverage values inside the blown region. In the unblown region, both in the curved and in the flat plate, the effect of the number of blown rows is clearly seen. Hydrodynamic boundary-layer profiles were measured with the aid of a triple hot-water probe. Three mean-velocity components and six turbulence quantities were simultaneously measured, and inside the blown region strong three-dimensionality was observed.
Effect of initial conditions on turbulent reattachment downstream of a backward-facing step
The reattachment of a fully turbulent, two-dimentional shear layer downstream of a backward-facing step has been studied experimentally. The work examines the effect of variations in inlet conditions on the process of reattachment. A series of experiments was conducted in a low-speed wind tunnel using specialized instrumentation suited to the highly turbulent reversing flow near reattachment. Accurate characterization of the time-mean features of the reattaching flows was possible. Assuming linear scaling normalized on distance from reattachment, distributions of normalized pressure coefficient and forward flow fraction, and time-averaged skin friction coefficient appear universal for two-dimensional reattachment, independent of initial conditions and step height, for given duct geometry (area ratio) and for high step-height Reyolds numbers with thin separating boundary layers. The results suggest universal flow structure in the reattachment zone.
Experimental study of flow reattachment in a single-sided sudden expansion
The reattachment of a fully turbulent, two dimensional, separated shear layer downstream of a single-sided sudden expansion in a planar duct flow was examined experimentally. The importance of changing the structure of the separated shear layer on the reattachment process itself was examined. For all cases, the Reynolds number based on step height was greater than 20,000, the expansion ratio was 5/3, and the inlet boundary layer was less than one-half step height in thickness. A crucially important phase was the development of a pulsed wall probe for measurement of skin friction in the reattachment region, thus providing an unambiguous definition of the reattachment length. Quantitative features of reattachment - including streamwise development of the mean and fluctuating velocity field, pressure rise, and skin friction - were found to be similar for all cases studied when scaled by the reattachment length. A definition of the reattachment zone is proposed.
Turbulent boundary-layer flow and structure on a convex wall and its redeveloment on a flat wall
Although the number of data sets which show the effects of convex curvature on turbulence has increased in the last 10 years, the development of really good calculational models is being held back by a lack of reliable data. The pesent investigation is concerned with a set of experiments which provide a litle more information regarding the response of a boundary-layer flow to convex curvature. Careful measurements were conducted to show the response of the boudnary layer to a sudden change from flat wall to curvature, and from curvature to flat wall. It was found that curvature effects were clearly apparent one or two boundary-layer thicknesses downstream of the start of curvature, while the disappearance of the curvature effects on a flat wall was an extremely slow process.
A mixing-length model for the prediction of convex curvature effects on turbulent boundary layers
A mixing-length model is developed for the prediction of turbulent boundary layers with convex streamwise curvature. For large layer thickness ratio, delta/R greater than 0.05, the model scales mixing length on the wall radius of curvature, R. For small delta/R, ordinary flat wall modeling is used for the mixing-length profile with curvature corrections, following the recommendations of Eide and Johnston (1976). Effects of streamwise change of curvature are considered; a strong lag from equilibrium is required when R increases downstream. Fifteen separate data sets were compared, including both hydrodynamic and heat transfer results. Six of these computations are presented and compared to experiment.
Reattaching turbulent shear layers with perturbed structure
The reattachment of a fully turbulent, two-dimensional shear layer downstream of a backward-facing step has been studied experimentally. The work examines the effect of modifications in shear layer structure upstream of reattachment. A series of experiments were conducted in a low-speed wind tunnel using specialized instrumentation suited to the highly turbulent, reversing flow near reattachment. Accurate characterization of the time-mean features of the reattaching flows was possible. Assuming linear scaling normalized on distance from reattachment, distributions of the normalized pressure coefficients, the time-averaged skin friction coefficient, and the forward flow fraction appear universal for two-dimensional reattachment, independent of initial conditions and steep height, for given duct geometry (area ratio) and for high step height Reynolds numbers. The results suggest universal flow structure in the reattachment zone.
Turbulent boundary layer heat transfer experiments: Convex curvature effects including introduction and recovery
Measurements were made of the heat transfer rate through turbulent and transitional boundary layers on an isothermal, convexly curved wall and downstream flat plate. The effect of convex curvature on the fully turbulent boundary layer was a reduction of the local Stanton numbers 20% to 50% below those predicted for a flat wall under the same circumstances. The recovery of the heat transfer rates on the downstream flat wall was extremely slow. After 60 cm of recovery length, the Stanton number was still typically 15% to 20% below the flat wall predicted value. Various effects important in the modeling of curved flows were studied separately. These are: the effect of initial boundary layer thickness, the effect of freestream velocity, the effect of freestream acceleration, the effect of unheated starting length, and the effect of the maturity of the boundary layer. An existing curvature prediction model was tested against this broad heat transfer data base to determine where it could appropriately be used for heat transfer predictions.
A modal separation measurement technique for broadband noise propagating inside circular ducts
A measurement technique which separates broadband noise propagating inside circular ducts into the acoustic duct modes is developed. The technique is also applicable to discrete frequency noise. The acoustic modes are produced by weighted combinations of the instantaneous outputs of microphones spaced around the duct circumference. The technique is compared with the cross spectral density approach presently available and found to have certain advantages, and disadvantages. Considerable simplification of both the new technique and the cross spectral density approach occurs when no correlation exists between different circumferential mode orders. The properties leading to uncorrelated modes and experimental tests which verify this condition are discussed. The modal measurement technique is applied to the case of broadband noise generated by flow through a coaxial obstruction (nozzle or orifice) in a pipe. Different circumferential mode orders are shown to be uncorrelated for this type of noise source.
Modal content of noise generated by a coaxial jet in a pipe
Measurements of the hydrodynamic pressure fluctuation levels showed that, for circular obstructions with a flow passage diameter less than one-half of the pipe diameter, hydrodynamic pressure fluctuations were small compared with the acoustic pressure, at locations greater than ten pipe diameters downstream of the restrictions. Thus, in the study currently reported, which was concentrated on measurement of the modal characteristics of the sound field downstream of the restrictions, special techniques to avoid contamination by hydrodynamic 'noise' were not necessary. The major results obtained in the study are scaling laws for the shape of the modal pressure spectrum and for the overall efficiency of noise generation. The characteristics of the modal pressure spectra are discussed. It is shown that the primary parameter controlling the spectrum shape is the frequency ratio (Helmholtz number/Strouhal number). The efficiency levels when plotted as a function of Mach number show a strong dependence on restriction size
Turbulent boundary layer heat transfer experiments: Convex curvature effects, including introduction and recovery
Heat transfer rates were measured through turbulent and transitional boundary layers on an isothermal, convexly curved wall and downstream flat plate. The effect of convex curvature on the fully turbulent boundary layer was a reduction of the local Stanton numbers 20-50% below those predicted for a flat wall under the same circumstances. The recovery of the heat transfer rates on the downstream flat wall was extremely slow. After 60 cm of recovery length, the Stanton number was still typically 15-20% below the flat wall predicted value. Various effects important in the modeling of curved flows were studied separately. These are: (1) the effect of initial boundary layer thickness; (2) the effect of freestream velocity; (3) the effect of freestream acceleration; (4) the effect of unheated starting length; and (5) the effect of the maturity of the boundary layer. Regardless of the initial state, curvature eventually forced the boundary layer into an asymptotic curved condition. The slope, minus one, is believed to be significant.
Turbulent boundary layer on a convex, curved surface
The effects of strong convex curvature on boundary layer turbulence were investigated. The data gathered on the behavior of Reynolds stress suggested the formulation of a simple turbulence model. Three sets of data were taken on two separate facilities. Both rigs had flow from a flat surface, over a convex surface with 90 deg of turning, and then onto a flat recovery surface. The geometry was adjusted so that, for both rigs, the pressure gradient along the test surface was zero - thus avoiding any effects of streamwise acceleration on the wall layers. Results show that after a sudden introduction of curvature, the shear stress in the outer part of the boundary layer is sharply diminished and is even slightly negative near the edge. The wall shear also drops off quickly downstream. In contrast, when the surface suddenly becomes flat again, the wall shear and shear stress profiles recover very slowly towards flat wall conditions.
An experimental investigation of two large annular diffusers with swirling and distorted inflow
Two annular diffusers downstream of a nacelle-mounted fan were tested for aerodynamic performance, measured in terms of two static pressure recovery parameters (one near the diffuser exit plane and one about three diameters downstream in the settling duct) in the presence of several inflow conditions. The two diffusers each had an inlet diameter of 1.84 m, an area ratio of 2.3, and an equivalent cone angle of 11.5, but were distinguished by centerbodies of different lengths. The dependence of diffuser performance on various combinations of swirling, radially distorted, and/or azimuthally distorted inflow was examined. Swirling flow and distortions in the axial velocity profile in the annulus upstream of the diffuser inlet were caused by the intrinsic flow patterns downstream of a fan in a duct and by artificial intensification of the distortions. Azimuthal distortions or defects were generated by the addition of four artificial devices (screens and fences). Pressure recovery data indicated beneficial effects of both radial distortion (for a limited range of distortion levels) and inflow swirl. Small amounts of azimuthal distortion created by the artificial devices produced only small effects on diffuser performance. A large artificial distortion device was required to produce enough azimuthal flow distortion to significantly degrade the diffuser static pressure recovery.
Investigation of a reattaching turbulent shear layer Flow over a backward-facing step
The paper studies incompressible flow over a backward-facing step in order to investigate the flow characteristics in the separated shear layer, the reattachment zone, and the redeveloping boundary layer after reattachment. It is shown that turbulent intensities and shear stress reach maxima in the reattachment zone, followed by rapid decay near the surface after reattachment. In addition, it is found that downstream of reattachment, the flow returns very slowly to the structure of an ordinary turbulent boundary layer.
Experiment on the turbulent boundary layer over convex walls and its recovery to flat-wall conditions
The experiments were performed at a ratio of the thickness at which velocity is 99% of the potential flow velocity to the surface radius of curvature of about 0.09 on a convex wall, where the curvature effects are strong. The data show that a flat-wall boundary layer is strongly affected by curvature only a few boundary layer thicknesses downstream of the start of convex curvature. On the other hand, in flow over a flat surface downstream of convex curvature, the effects of curvature persist downstream for ten or more boundary layer thicknesses. It appears that the boundary layer over a convex wall may have a multilayer character which is slow to return to normal, flat-plate equilibrium.
Modal content of noise generated by a coaxial jet in a pipe
Noise generated by air flow through a coaxial obstruction in a long, straight pipe was investigated with concentration on the modal characteristics of the noise field inside the pipe and downstream of the restriction. Two measurement techniques were developed for separation of the noise into the acoustic duct modes. The instantaneous mode separation technique uses four microphones, equally spaced in the circumferential direction, at the same axial location. The time-averaged mode separation technique uses three microphones mounted at the same axial location. A matrix operation on time-averaged data produces the modal pressure levels. This technique requires the restrictive assumption that the acoustic modes are uncorrelated with each other. The measured modal pressure spectra were converted to modal power spectra and integrated over the frequency range 200-6000 Hz. The acoustic efficiency levels (acoustic power normalized by jet kinetic energy flow), when plotted vs. jet Mach number, showed a strong dependence on the ratio of restriction diameter to pipe diameter. The acoustic energy flow analyses based on the thermodynamic energy equation and on the results of Mohring both resulted in orthogonality properties for the eigenfunctions of the radial mode shape equation. These orthogonality relationships involve the eigenvalues and derivatives of the radial mode shape functions.