Characterization of a large scale hypersonic shock tunnel and investigation of the effect of roughness on a large cone boundary layer flow

Abstract

High speed flight is ridden with engineering challenges, some of which have been studied for decades with only minor improvements. Boundary layer transition to turbulence in highMach number flow is one such topic, which, although has seen impressive improvements, remains far from staple engineering prediction. The boundary layer transition at high speeds is critical due to the high levels of heating that the turbulent boundary layer brings, but the same turbulent layer is also sought after for its resistance towards separation in adverse pressure gradients. This advantage is capitalized by Scramjet inlets where boundary layer separation can lead to unstart of the engine, and so the boundary layer must not be laminar when entering the engine. Given the hypersonic flight conditions of low unit length Reynolds numbers at high altitudes, most high Mach number flights cannot expect to have natural boundary layer transition and therefore must actively trip the boundary layer towards transition. This effect of a given roughness on the boundary layer at aMach 8 hypersonic flowover a generic axisymmetric forebody is the prime objective of this study. The study is performed at a low Reynolds number condition, which is realistic for a Mach 8 flight, to test an unfavorable condition for producing a turbulent boundary layer. The present study employs an 800mmlong sharp cone as the test body for the boundary layer experiments. The long model is housed in a recently characterized large scale shock tunnel. Experiments are conducted in Mach 8 flow at a low Reynolds number of 3.2 million per m. To trip the flow, roughness in the form of diamond shaped isolated 3-dimensional elements are primarily used, given the previous success from other studies which make it the most effective trip shape in flows with such speeds. For flows with boundary layer edge conditions that are hypersonic, the trip height requirements have been previously found to be greater than the local velocity boundary layer thickness. For the present study, the trip heights have been varied up to 5 times the local boundary layer thickness. For all the cases studied, there is no clear hint of a transitional boundary layer. The baseline case without trips, is found to be fully laminar. The cases with trips do affect the boundary layer, and this is observed by the heat flux variations behind the trips, but in no case does the heat flux variation clearly indicate development of a transitional boundary layer. This result brings out the importance of Reynolds numbers at high altitude hypersonic flight conditions, presenting a case of challenging tripping environment that may not even lead to a turbulent boundary layer. Moreover, the axisymmetric nature of the body makes transition further difficult, where previously obtained data in other studies showthat the flat plate case is relatively easier to bring to boundary layer transition. Finally, a comparison of the data is made with a few existing roughness correlations for transition prediction. It is found that the correlations do not work well in the hypersonic boundary layer edge flow conditions and that they over predict the effect of the Reynolds numbers. In carrying out this boundary layer study, a large scale shock tunnel has been calibrated and characterized. Apart from the regular calibration, a new analysis of the shock tube data is presented that helps to study the available test gas slug in such a facility. The new analysis brings out the various non-ideal effects involved in the deterioration of the performance of a facility and can be applied to any such facility without the need for any special equipment or measurement. For the large scale shock tunnel, this analysis helps to quantify the test gas slug as a function of operational incident shockMach number alone. This information is further found useful in accurate test time prediction in non-tailored mode of tunnel operation but may be applied to tailored conditions as well. Finally, using a previously established early driver gas arrival mechanism, attempts are made to predict and verify the driver gas arrival in the tunnel mode of operation. This process has been found to be partially successful, with further work required in this potential application of the presented test gas slug analysis.