Engineering topics
Publications and source records attributed to Smits, A. J..
A reattaching supersonic shear layer was perturbed by transverse air injection, and the effects were investigated using flow visualization. It was found that shear layer mixing was increased and that the shock system was strongly disturbed due to the heightened three dimensionality of the flow. Mixing enhancement was also achieved in an incompressible downstream-facing step flow using the same blowing procedure.
Measurements of wall pressure fluctuations were made as part of an ongoing program to investigate the control of a turbulent, reattaching shear layer at Mach 2.9. The flow was disturbed near separation by air injection normal to the plane of the shear layer. This perturbation was found to dramatically increase the intensity of pressure fluctuations in the vicinity of reattachment. A local increase in shear layer growth rate was also observed in the perturbed flow.
A study was made of the wall pressure fluctuations in the reattachment region of a supersonic free shear layer. The free shear layer was formed by the separation of a Mach 2.9 turbulent boundary layer from a backward facing step. Reattachment occurred on a 20-deg ramp. By adjusting the position of the ramp, the base pressure at the step was set equal to the freestream pressure, and the free shear layer formed in the absence of any turning. An array of flush-mounted, miniature, high-frequency pressure transducers was used in the vicinity of the reattachment region to make multichannel measurements of the fluctuating wall pressure. Contrary to previous observations of this flow, the reattachment region was found to be highly unsteady, and the pressure fluctuations were found to be large. The overall behavior of the wall pressure loading is similar in scale and magnitude to the unsteadiness of the wall pressure field in compression ramp flows at the same Mach number.
The research performed was an experimental program to help develop turbulence models for shock wave boundary layer interactions. The measurements were taken in a Mach 3, 16 deg compression corner interaction, at a unit Reynolds number of 63 x 10(exp 6)/m. The data consisted of heat transfer data taken upstream and downstream of the interaction, hot wire measurements of the instantaneous temperature and velocity fluctuations to verify the Strong Reynolds Analogy, and single- and double-pulsed Rayleigh scattering images to study the development of the instantaneous shock/turbulence interaction.
An experimental program to investigate the control of a turbulent, reattaching shear layer at Mach 2.9 is described. In preliminary experiments, schlieren photography and Rayleigh scattering were used to visualize the effects of air injection normal to the plane of the shear layer. Localized blowing was qualitatively found to increase the growth rate of the shear layer, the intensity of the turbulence, and the unsteadiness of the reattachment shock. However, uniformly distributed air injection did not appear to affect the flow strongly. These results suggest that three-dimensional disturbances are more effective for flow control than two-dimensional disturbances, and that there is a strong connection between the incoming turbulence and the shock motion in this flow.
The primary aim is to investigate the mechanisms which cause the unsteady wall-pressure fluctuations in shock wave turbulent shear layer interactions. The secondary aim is to find means to reduce the magnitude of the fluctuating pressure loads by controlling the unsteady shock motion. The particular flow proposed for study is the unsteady shock wave interaction formed in the reattachment zone of a separated supersonic flow. Similar flows are encountered in many practical situations, and they are associated with high levels of fluctuating wall pressure. Wall pressure fluctuations were measured in the reattachment region of the supersonic free shear layer. The free shear layer was formed by the separation of a Mach 2.9 turbulent boundary layer from a backward facing step. Reattachment occurred on a 20 deg ramp. By adjusting the position of the ramp, the base pressure was set equal to the freestream pressure, and the free shear layer formed in the absence of a separation shock. An array of flush-mounted, miniature, high-frequency pressure transducers was used to make multichannel measurements of the fluctuating wall pressure in the vicinity of the reattachment region. Contrary to previous observations of this flow, the reattachment region was found to be highly unsteady, and the pressure fluctuations were found to be significant. The overall behavior of the wall pressure loading is similar in scale and magnitude to the unsteadiness of the wall pressure field in compression ramp flows at the same Mach number. Rayleigh scattering was used to visualize the instantaneous shock structure in the streamwise and spanwise direction. Spanwise wrinkles on the order of half the boundary layer thickness were observed.
This paper presents measurements in a Mach 3 turbulent boundary layer which was subjected to a short region of concave surface curvature. Mean velocity and time-averaged turbulence measurements are presented to indicate the severity of the distortion, although the focus of this study is on how the large-scale organized motions are affected by the distortion. A 12 deg increase in the inclination angle of organized motions is observed, and correlation measurements indicate an increase in the streamwise length scale. VITA conditional sampling is used to identify strong positive streamwise mass-flux gradients as the large-scale structure responsible for the measured structure angle and to determine the ensemble-averaged flowfield associated with the mass-flux gradient structures.
A wide range of recent work on compressible turbulent boundary layers is described. Special attention was paid to flows with rapid changes in pressure including flows with shock waves, curved walls, and expansions. The application of rapid distortion theory to flows transversing expansion and shock waves is reviewed. This is followed by an account of experiments aimed at elucidating the large scale structures present in supersonic boundary layers. The current status of laser-Doppler and hot-wire anemometry in supersonic flow is discussed, and a new interferometric technique for the determination of wall-stress is described. The use of small pressure transducers to deduce information about the structure of zero pressure-gradient and severely perturbed boundary layers is investigated. Finally, there is an extension of the data presentation of AGARDographs 223, 253 and 263 to cover rapidly distorted boundary layers.
An experimental investigation concentrating on the change in the large-scale organized motions of a supersonic turbulent boundary layer subjected to a short region of concave surface curvature is presented. Two flows were studied, a relatively weak and a relatively strong perturbation. In both flows the freestream Mach number was 2.89 and the Reynolds number based on momentum thickness was 81,900. Hot-wire anemometry and schlieren photography provided the principle tools for this investigation. Hot-wire data were analyzed using spectral and correlation methods and conditional sampling techniques. Changes in structure angle and results of conditionally sampled velocity signals are discussed. The strong perturbation is found to increase the structure angle, and in both flows the length scales of the organized motions are altered.
The calibration and operation of a constant-temperature crossed-wire probe in supersonic flow is considered. Crossed-wire probes offer considerable advantages over single, inclined wires: the kinematic shear stress can be derived from a single point measurement; the rms quantities can be derived from the same measurement, and the instantaneous quantities can be obtained as a continuous function of time. However, using a crossed-wire probe in supersonic flow is subject to the following practical difficulties: the problem of flow interference, where the shock waves from one wire and its supports interfere with the flow over the other wire; the necessity for high frequency response to resolve the spectral content, and the sensitivity of the results to small changes in the calibration constants. In the present contribution, each of these problems is addressed. Practical solutions are suggested, and some encouraging results are presented.
This paper presents and discusses experimental data from an investigation of organized motions in a supersonic turbulent boundary layer. Conditional sampling of crossed-wire and multiple normal-wire signals is performed. A comparison is made between events detected using the VITA conditional sampling technique and those found by thresholding the UV signal. Based on this comparison, limitations of the VITA technique are discussed. The conditional sampling results indicate that most organized motions are consistent with hairpin vortices.
Calculations were performed to investigate the supersonic flow of a turbulent boundary layer over short regions of concave surface curvature. Upstream of each curved surface the freestream Mach number was 2.9. Three different constant radii of curvature models were investigated to cover a range of curvatures and turning angles. The numerical technique solved the full, Reynolds-averaged Navier-Stokes equations using two different turbulence models: the algebraic eddy viscosity model due to Baldwin and Lomax (1978), and the one-equation model to Rubesin (1976). The calculations were compared with the experimental data of Taylor and Smits (1984), and the agreement was very satisfactory, especially for the computations using the one-equation model.
An experimental investigation is conducted of the effects of bulk compression and streamline curvature on a turbulent supersonic boundary layer whose incoming Mach number is 2.87. Two different, constant-radius curved wall models were used; the total turning angle in both cases was fixed at 8 deg. While the amplification of the normal stress was approximately the same in both models, the amplification of the shear stress was greater in the one with delta(o)/R value 0.1 than in the model with a value of 0.02. The flow over the latter model appears to be especially complex.
The processes present in a turbulent boundary layer (BL) which experiences a sudden change in the surface roughness are discussed in terms of numerical models in comparison with empirical data. Consideration is limited to flow perturbations which preserve the BL approximation. Generalized continuity, momentum and turbulent energy transport equations are reviewed. The perturbed BL has two regions: one with large velocity gradients and an outer zone with smaller gradients. Expressions are defined for the velocity and temperature profiles and the roughness length after a step change in surface roughness, which causes the disturbance. A pertrurbation could also arise from a heat flux input of a pressure gradient. The effects of the shape of the perturbing step and the presence of several disturbing edges are explored. Finally, it is shown that closure forms and constants for self-preserving flows are worthy of further development because of their demonstrable abilities to describe simple perturbed flows.
A major effort to improve the accuracies of turbulence measurement techniques is described including the development and testing of constant temperature hot-wire anemometers which automatically compensate for frequency responses. Calibration and data acquisition techniques for normal and inclined wires operated in the constant temperature mode, flow geometries, and physical models to explain the observed behavior of flows are discussed, as well as cooperation with computational groups in the calculation of compression corner flows.
Previously cited in issue 05, p. 591, Accession no. A83-17917
Previously cited in issue 15, p. 2345, Accession no. A82-31948
Mean flow measurements have been made in a supersonic turbulent boundary layer experiencing a short region of concave curvature and an adverse pressure gradient. The incoming boundary layer had a Mach number of 2.87 and a unit Reynolds number of 6.3 x 10 to the 7th/m. Flow over constant radii curvatures of delta (0)/R = 0.10 and 0.02 with a fixed turning angle of 8 deg were investigated. Results indicate that both flows remain two-dimensional; length scales increase over their equilibrium values; the mean flow behaves similarly to subsonic flows experiencing concave curvature and lateral divergence; and boundary layer behavior is strongly similar to a corresponding 8 deg ramp flow.