The Physics of Transitional Jet Interactions

Abstract

An experimental and computational study has been undertaken to investigate the interaction between a transverse circular sonic jet and a Mach 6.7 cross flow on a sharp-edged flat plate with particular emphasis on the influence of the oncoming boundary layer state on the flowfield characteristics. This was achieved by generating steady wedges of turbulent flow in the otherwise laminar oncoming boundary layer. The first component of this investigation involved an experimental investigation of transitional phenomena on a sharp-edged flat plate in a free-piston shock tunnel. Four test conditions were examined spanning a unit Reynolds number range of 3.16- 7.82×106m-1 and a flow enthalpy range of 2.86-3.20MJ/kg. The onset of boundary layer transition and the characteristics of the dynamic structures within the transitional region for these conditions were consistent with previously reported results. The main investigation examined the response of the jet interaction (JI) flowfield in response to mixed laminar/turbulent oncoming flow. The test condition used for this work had a unit Reynolds number of 4.86×106m-1 and a flow enthalpy of 3.02MJ/kg. The interaction was generated by nitrogen gas from a circular 3mm diameter sonic nozzle at total jet-to-freestream pressure ratios of between 237 and 412. The upstream separated region was found to be very sensitive to the upstream boundary layer state with simultaneous regions of aminar-like and turbulent-like separation. The shock structures immediately around the expanding plume were relatively insensitive, although some change in shape was noted. Highly variable localised regions of high heat transfer were observed around the injector. The Computational Fluid Dynamic (CFD) study found a previously unexplained coupling between the upstream and downstream vortex structures. It also demonstrated that the forces and moments generated by the jet could be strongly influenced by both the location and width of the boundary layer turbulence relative to the size of the interaction. The agreement between the experimental and computational surface pressures results was good. The heat transfer comparisons were limited due to the inability of the steady-state code to accurately model the large scale transient structures present in the physical flowfield which contribute to this process.