Numerical investigation of shock-shock and shock-boundary layer interaction mechanisms and in terms of surface heat transfer at hypersonic flow

Recently, comprehensive studies have been carried out for developing hypersonic vehicles like aircraft, rockets, etc. in the atmosphere. Shockwave interactions are one of the most important challenges in designing such vehicles because of the high surface heating that can change the aerodynamic characteristics of the vehicle at high speeds. In the present study, shock-shock and shock-boundary layer interactions and high heat transfer rate resulting from such interactions are investigated by using Computational Fluid Dynamics (CFD) methods. In the context of the thesis, a double wedge geometry which is one of the commonly used generic geometries is chosen to compare with the experimental results in the literature. At first, the analyses in a case where constant angles of the double wedge are rigorously performed with different flow models. The obtained results are compared with experimental and other computational studies and it is determined that what flow model give accurate results. Next, the first angle of the double wedge is fixed and second angle is parametrically changed for a laminar flow in two- and three-dimensional computations. The flow physics is tried to recognize in a short- or long-term for various angles of the second wedge, and the high pressure and temperature changes on the surfaces of the double wedge are examined. One of the most significant contributions of the study to the literature is to show that the flow is time periodic after the experimental periods reached today at hypersonic speeds. First, in order to validate the computational model and solver, CFD analyses are carried out for inviscid, laminar and turbulence models in an unsteady two-dimensional flow over the double wedge geometry. The computations are performed using OpenFOAM software, which is an open-source CFD solver based on finite volume method. One of the solvers of OpenFOAM, rhoCentralFoam is a density-based Navier-Stokes solver. The solver is first and second order accurate, respectively, in time and space. The solver can capture contact discontinuities and shock waves such as strong discontinuities in the flow field without oscillation since Kurganov and Tadmor's central upwind scheme is used. A 7-coefficient NASA thermodynamics relation is used to model gas thermodynamic properties depending on the temperature. Both the computational model and the solver are validated by comparing obtained results to experimental measurements and other computational studies. In addition, solver validation study is examined for a different double wedge geometry which has experimental data. Regarding the results of validation study, it can be said that the assumption of the laminar flow is the most accurate approach when compared to the experimental measurements. Second, the first angle of the double wedge is kept constant and the second wedge angle is systematically changed from θ2=40° to 60° with 5° intervals. Shock interaction mechanisms for different aft angles in a two-dimensional low enthalpy flow, flow physics and variations in the heat transfer are investigated for t=327 µs by means of Mach contours and average heat flux distributions on the surface. It is observed that as the second ramp angle increases, the bow (BS) and transmitting (TS) shocks propagating from the triple point to the aft wedge strengthen, the shear layer extends towards the leading edge and numerous new shock waves occur. Edney Type V and its early stages are only observed as the shock interaction mechanism and the flow is not steady for the cases with θ2= 50°, 55° and 60°. It is observed that intensity and impingement angle of TS dictate the size of the separation bubble at TS impingement point. On the other hand, separation bubble size affects standoff distance of BS and thereby the intensity of TS. That means these two mechanisms create a two-way coupling between BS and the boundary layer on the ramp surface. The adverse pressure gradient caused by TS propagates upstream through the subsonic boundary layer and that triggers complex and unsteady flow as the angle of the θ2 increases. In addition, the results show that there is a threshold value and the value is in the range of θ2= 45° - 50° for two-way coupling mechanism. Moreover, the presence of time periodicity is investigated by performing computations up to 50 times the duration of the former two-dimensional flow (t=327 µs). In order to find the threshold value for two-way coupling mechanism, the cases with θ2=45° - 50° is solved 1° intervals for the long-term and it is determined that the threshold value is θ2=47°. Then, the aft angle is changed from θ2=50° to 60° with the intervals of 5°. The shock interaction mechanism and flow physics are described by tracking of instantaneous separation angle, instantaneous locations of separation shock (SS) and TS, temporal variations of pressure and heat flux at several stations and with the help of density gradient and Q contours that visualize the vortex structures. The long-term computations show that the the flow over the double wedge is steady under θ2=47° threshold value. Otherwise, the flow becomes time periodic beyond the threshold value. As the aft angle increased, the distance from the leading edge to separation point shortens, number of the vortices increases near the compression corner, separation region gets longer and thicker and duration of the TS's impingement on the aft wedge shortens. The obtained results indicate that the case with θ2=60° might be a limit for time periodic flow and there can be a relation between the time periodicity and the length of the second surface. Finally, the case with θ2=55° is first performed in three-dimensional computational domain in order to examine three-dimensional effects and whether the flow preserves two-dimensionality. The obtained three-dimensional results are compared with the experimental and computational studies for t=327 µs and the differences between the results obtained from two-dimensional computations are also examined. For both two- and three-dimensional results, the density gradient contours correspond to the Schlieren images from the experiment at the beginning of the simulation. However, the density gradient contours in three-dimensional results are moving away from Schlieren images in the subsequent instants. It is shown that this situation is due to three-dimensional relaxation (or relieving) effects. The surface heat flux distributions for three-dimensional simulation match the experimental measurements better than two-dimensional results, but impingement point of TS is observed to slip downstream direction according to two-dimensional results. After the assessment of the case with θ2=55° in the chapter of three-dimensional results, the aft angle is changed from θ2=45° to 60° with 5° intervals as in the previous chapters. As in the case of two-dimension, the flow is not steady for the cases with θ2=50°, 55° and 60°. The comparison of two- and three-dimensional results shows the following discrepancies: while almost the same shock structures are observed in the case with θ2=45° for all instants, same shock structures are only identified at about the first 100 µs in the cases with θ2>45°. When θ2 is increased, the distance between of SS and BS in three-dimensional simulation get longer with respect to two-dimensional simulation. Therefore, the location of the impingement point of TS in the case θ2=55° and 60° is significantly different from two-dimensional results. While there are several additional shock waves due to separation bubbles in the separation zone of two-dimensional simulation, less separation bubble and less related shock waves are observed in three-dimensional simulation. In all θ2 cases except for the case with θ2=45°, the flow is truly three-dimensional. Due to three-dimensional relaxation effect, in the spanwise direction from mid-span to sides, BS gets weaker, where the subsonic region is smaller, SS moves towards compression corner and separation region gets narrower and thinner over the surfaces for all the cases with θ2. Triple point also approaches the double wedge surfaces. Especially in the cases with θ2=55° and 60°, distinct striations, which are the characteristic sign of the presence of Görtler-like vortices, at the impingement lines of TS in spanwise direction (or at reattachment) are observed within the flow. Görtler-like vortices cause the highest heating at the TS impingement point to move upstream or downstream direction over the wedge surfaces. In brief, obtained results from three-dimensional computations reveal that the flow is unsteady and has three-dimensional nature including Görtler-like vortices. In conclusion, a two-way coupling between BS and the boundary layer on the ramp which triggers complex and unsteady flow as the angle of θ2 is increased is the major contributions to the literature. Moreover, it is determined that there is a threshold value for the two-way coupling and that value is θ2=47°. A notable contribution to the literature is to show that the flow is time periodic after the experimental periods reached today at hypersonic speeds. The time-periodic motion can impose a limiting factor in the design of a hypersonic vehicle. From this point of view, taking the findings of this study into account can be important in the design stage of such vehicles. Another of the literature contributions is that Görtler-like vortices, which are detected in three-dimensional computations, have the ability to transport high heating areas towards upstream and downstream. In the view of these results, detection and tracking of the Görtler-like vortices can be crucial in the design of the hypersonic vehicle.