Pulsed plasma jets (also called "SparkJets'") were investigated for use in controlling the mean separation location induced by shock wave-boundary layer interaction. These synthetic jet actuators are driven by electro-thermal heating from an electrical discharge in a small cavity, which forces the gas in the cavity to exit through a small hole as a high-speed jet. With this method of actuation, pulsed plasma jets can achieve pulsing frequencies on the order of kilohertz, which is on the order of the instability frequency of many lab-scale shock wave-boundary layer interactions (SWBLI). The interaction under investigation was generated by a 20° compression ramp in a Mach 3 flow. The undisturbed boundary layer is transitional with Re[subscript theta] of 5400. Surface oil streak visualization is used in a parametric study to determine the optimum pulsing frequency of the jet, the optimum distance of the jet from the compression corner, and the optimum injection angle of the jets. Three spanwise-oriented arrays of three plasma jets are tested, each with a different pitch and skew angle on the jet exit port. The three injection angles tested were 22° pitch and 45° skew, 20° pitch and 0° skew, and 45° pitch and 0° skew. Jet pulsing frequency is varied between 2 kHz and 4 kHz, corresponding to a Strouhal number based on separation length of 0.012 and 0.023. Particle image velocimetry is used to characterize the effect that the actuators have on the reattached boundary layer profile on the ramp surface. Results show that plasma jets pitched at 20° from the wall, and pulsed at a Strouhal number of 0.018, can reduce the size of an approximate measure of the separation region by up to 40% and increase the integrated momentum in the downstream reattached boundary layer, albeit with a concomitant increase in the shape factor. / text

No description available.

<p>High speed flows in engineering applications are often characterized by shock-wave/boundary layer interactions (SWBLI) and three-dimensional flows. This thesis aims to study commonly employed turbulence models in the context of configurations featuring SWBLI and 3D flows. Computational Fluid Dynamics (CFD) simulations of flow over a 20 degree isentropic compression curved fin were performed, and the results were compared with published experimental data. Two variations of the Spalart-Allmaras (SA) turbulence model, and two variations of the Menter Shear Stress Transport (SST) turbulence model were tested, and their ability to predict mean flow data was compared. All four models predict the correct flow structure, but the SA models display higher error in predicting Pitot pressure in the curved fin 3D boundary layer. CFD simulations of flow over a sharp fin were performed using the same four turbulence models, and mean flow data were compared to published experimental and computational data. Simulated flow profiles showed good accuracy in the regions away from the shock structure, but did not accurately predict flow in the supersonic regions in the vicinity of the shock wave. Time-accurate IDDES simulations of both configurations were performed, and neither configuration showed any deviation from the steady state solution.</p>

Shock wave-boundary layer interactions (SWBLIs) have been studied ex-tensively due to their practical importance in the design of high speed ve-hicles. These interactions, especially the ones leading to shock induced separation are typically unsteady in nature and can lead to large fluctuating pressure and thermal loads on the structure. The resulting shock oscil-lations are generally composed of high-frequency small-scale oscillations and low-frequency large-scale oscillations, the source of the later being a subject of intense recent debate. Motivated by these debates, we study in the present work, the SWBLI at a compression ramp and on a forward-facing step (FFS) at a Mach number of 2.5. In the case of compression ramps, a few ramp angles are studied ranging from small (10 degree) ramp angle to relatively large values of up to 28 degrees. The FFS configuration, which consists of a 90 degree step of height h, may be thought of as an extreme case of the compression ramp geometry, with the main geometri-cal parameter here being (h/δ), where δis the thickness of the oncoming boundary layer. This configuration is less studied and has some inherent advantages for experimentally studying SWBLI as the size of the separa-tion bubble is large. In the present experimental study, we use high-speed schlieren, unsteady wall pressure measurements, surface oil flow visualiza-tion, and detailed particle image velocimetry (PIV) measurements in two orthogonal planes to help understand the features of SWBLI in the com-pression ramp geometry and the forward-facing step case. The SWBLI at a compression ramp has been more widely studied, and our measurements show the general features that have been seen in earlier studies. The upstream boundary layer is found to separate close to the ramp corner forming a separation bubble. The streamwise length of the separa-tion bubble is found to increase with the ramp angle, with a consequent shift of the shock foot further upstream. At very small ramp angles up to 10 degrees, there is no evidence of separation, while at large ramp angles of 28 degrees, the separation bubble extends upstream to about 3.5δ(δ=boundary layer thickness). In all cases, the separation bubble is however very small in the wall normal direction, typically known to be about 0.1δ, and hence is difficult to directly measure in experiments using PIV. Shock foot measurements using PIV show that the shock has a spanwise ripple, which seems directly related to the high-and low-speed streaks in the in-coming boundary layer as recently shown by Ganapathisubramani et al. (2007). The forward-facing step configuration may be thought of as an extreme case of the compression ramp geometry, with a ramp angle of 90 degrees. This configuration has not been extensively studied, and is experimentally convenient due to the large separation bubbles formed ahead of the step. In the present work, extensive measurements of the mean and unsteady flow around this configuration have been done, especially for the case of h/δ=2, where his the step height. Pressure measurements in this case, show clear low-frequency motions of the shock at non-dimensional frequencies of about fh/U∞≈ 0.02. In this case, PIV measurements show the pres-ence of a large mean separation bubble extending to about 4hupstream and about 1hvertically. Instantaneous PIV measurements have been done in both cross-stream (streamwise and wall-normal plane) and in the span-wise (streamwise-spanwise) plane. Instantaneous cross-stream PIV mea-surements show significant variations of the shock location and angle, be-sides large variations in the recirculation region (or separation bubble), this being determined as the area having streamwise velocities less than zero. From a large set of individual PIV instantaneous fields, we can estimate the correlation of the measured shock location to both downstream effects like the area of the recirculation region, and upstream effects like the presence of high-/low-speed streaks in the oncoming boundary layer. We find that the shock location measured from data outside the boundary layer is more highly correlated to downstream effects as measured through the recircu-lation area compared to upstream effects in the boundary layer. However, we find that the shock foot within the boundary layer has ripples in the spanwise direction which are well correlated to the presence of high-/low-speed streaks in the incoming boundary layer. These spanwise ripples are however found to be small (less than one h) compared to the highly three-dimensional shape of the recirculation region with spanwise variation of the order of 3 step heights. In summary, the study shows that the separated region ahead of the step is highly three-dimensional. The shock foot within the boundary layer is found to have ripples that are well correlated to fluctuations in the in-coming boundary layer. However, we find that the large-scale nearly two-dimensional shock motions outside the boundary layer are not well cor-related to the fluctuations in the boundary layer, but are instead well cor-related with the spanwise-averaged separation bubble extent. Hence, the present results suggest that for the forward-facing step configuration, it is the downstream effect caused by the separation bubble that leads to the observed low-frequency shock motions.