Study of separating compressible turbulent boundary-layers

The prediction of incipient separation due to shock wave interaction with a boundary layer is a critical prerequisite in the design of air-breathing engine inlets and combustors. These shock wave/boundary layer interactions occur in both two-dimensional and three-dimensional forms. In the 1970's Korkegi developed empirical correlations to predict the incipient separation of turbulent boundary layers based on available data from other researchers. In an attempt to understand the mechanism of separation of a boundary layer when subjected to a shock wave, an analysis based on momentum arguments has been formulated and has been compared with the Korkegi correlations. The aim of this work is to investigate experimentally and numerically the suitability of both predictions at the upper Mach number limit of the Korkegi correlation. The conditions studied are a hypersonic, perfect gas at Mach 9.5. Experiments were conducted in a small reflected shock tunnel with a model designed to produce a turbulent boundary layer along the inner wall of a circular duct. Incident shocks of varying strength were produced by placing cones of different semivertex angles along the centreline of the duct, near the exit. Static wall pressure and heat transfer rates were measured along the duct inner surface. Numerically these conditions are simulated using a finite volume formulation for the solution of the compressible Navier-Stokes equations. Large-Eddy Simulation (LES) is used with the Approximate Deconvolution Model subgrid scale (SGS) closure to include the effects of turbulence. Reasonable agreement was observed between experimental and numerical results for 8°, 15° and 20° semivertex cone angles. Interaction between the separated boundary layer and the cone was identified numerically. At this stage the selected semivertex cone angles have been inconclusive in determining suitability of either separation correlation