Investigation and improvement of supersonic intake flow characteristics using boundary layer control techniques

Abstract

In this study, a supersonic air intake is aimed to be analysed using an open source Navier-Stokes solver, HiSA. The interaction between shock waves and boundary layers is a significant factor to consider when designing supersonic inlets for aircraft. Bleed systems have traditionally been used to improve the stability and efficiency of these inlets by removing the boundary layer to prevent flow separation due to adverse pressure gradients. However, determining the impact of the boundary layer bleed system on the performance of the supersonic inlet remains a challenging problem. To study the effects of such systems, a literature study is performed on the subject. Design and effect of the bleed mechanisms are studied. It is seen that the overall pressure recovery values can be increased by a significant amount using bleed systems. The design considerations for a rectangular supersonic intake with a circular cross-section are summarized. Emphasis is placed on the creation of a watertight 3D CAD model of the intake geometry and the generation of a suitable mesh to ensure accurate simulation results. Additionally, a bleed system is designed for the intake, with the purpose of controlling flow conditions and preventing performance loss. The design considerations for the bleed system encompass the analysis of the bleed entry area, bleed exit area, and bleed wall angles, in terms of their effects on the overall performance and functionality. The bleed system serves important purposes, including boundary layer control, regulation of total pressure recovery, and prevention of flow separation. Proper design and optimization of the bleed system are deemed crucial for achieving efficient and reliable operation of the supersonic intake. The bleed entry area, bleed exit area, and bleed wall angles are systematically varied through a parametric analysis. This analysis aims to provide insights into the sensitivity of the bleed system's performance to these parameters and identify optimal design configurations. Performance evaluation metrics, such as total pressure recovery, flow uniformity, and flow separation characteristics, is defined as benchmarks for comparing different bleed system designs. To verify the solver settings which is used in this study, a validation study is performed. Two cases are used for the validations. In the first case, a zero-gradient solution for the inlet is obtained. Then, a back pressure case with a pressure ratio of 7 is computed for comparison. The obtained results matched the reference study's numerical solutions almost perfectly, but differed significantly from the reference study's experimental results, possibly due to different back pressure boundary conditions applied at the intake end. To test the discrepancy, a 3D mesh was generated and a computational analysis was performed, resulting in good accuracy with the 2D results. Therefore, another case was attempted using a similar geometry and CFD approach as a reference study, resulting in good agreement with the experimental results. The results of the study demonstrated the potential of CFD in predicting the performance of supersonic inlets with good accuracy. Next, a mesh independency study is performed both for two and three dimensional meshes. For the two-dimensional mesh, a fine grid is selected since the computational costs are not high in two dimensional solutions. While the medium grid is selected for the three-dimensional studies resulting from the same fact. Computation times and costs are a much important factor in three dimensional analyses. Then the computational studies are begun for the intake geometry. First, two dimensional analyses are performed at different back pressure values. Apart from the case with BPR=5.0 a strong separation is seen at the ramp of the intake while the separation occurred on the cowl at BPR=5.0. As the BPR increased, the intake shifted towards the critical condition from the supercritical conditions. At BPR=5.9, the intake was operating in the critical condition. As the BPR is increased to 6.2, the intake shifted to the subcritical conditions in clean configuration. It should also be noted that the intake did not operate at the subcritical condition for none of the cases when the bleed geometry was present. It is observed that intake flow structure was steady in most cases apart from a few outliers such as the case with BPR=5.0. It is seen that the bleed geometry was successful at preventing flow separation at the ramp if the bleed entry area and bleed exit are large enough. However, in most of these cases, the separation instead occurred at the cowl surface. Despite this situation, the bleed geometry improved the flow uniformity in most cases by delaying the flow separation. The total pressure recovery values show that the bleed configuration was most successful at preventing performance loss at critical and subcritical conditions. Then the three-dimensional studies are performed. It is observed that, for the most part, the solutions agree with their two-dimensional counterparts. However, it is seen that the intake reached the critical and subcritical conditions earlier compared to the two-dimensional solutions. The clean geometry reached subcritical conditions at BPR=5.9 while it was 6.2 at the two-dimensional cases. Depending on the geometry, the bleed geometries reached the subcritical condition at either BPR=6.0 or BPR=6.2, whereas subcritical conditions were not observed in two-dimensional bleed cases. This situation is thought to be caused by the circular cross-section of the intake diffuser and exit which restricts the flow more than predicted in two-dimensional cases. In the three-dimensional cases, it is observed that as the intake approached the critical and subcritical conditions, the flow started to show unsteady behaviour. Similar to the two-dimensional cases, it is observed that the bleeding system was successful in the suction of the SWBLI induced separation. However, once again, this caused the separation to occur on the cowl surface for the most of the cases. It was also observed that the flow was more with bleed systems compared to the clean geometry at near critical and subcritical conditions.