Transitional Bubble in Periodic Flow Phase Shift

Abstract

One particular characteristic observed in unsteady shear layers is the phase shift relative to the main flow. In attached boundary layers this will have an effect both on the instantaneous skin friction and heat transfer. In separation bubbles the contribution to the drag is dominated by the pressure distribution. However, the most significant effect appears to be the phase shift on the transition process. Unsteady transition behaviour may determine the bursting of the bubble resulting in an un-recoverable full separation. An early analysis of the phase shift was performed by Stokes for the incompressible boundary layer of an oscillating wall and an oscillating main flow. An amplitude overshoot within the shear layer as well as a phase shift were observed that can be attributed to the relatively slow diffusion of viscous stresses compared to the fast change of pressure. Experiments in a low speed facility with the boundary layer of a flat plate were evaluated in respect to phase shift. A pressure distribution similar to that on the suction surface of a turbomachinery aerofoil was superimposed generating a typical transitional separation bubble. A periodically unsteady main flow in the suction type wind tunnel was introduced via a rotating flap downstream of the test section. The experiments covered a range of the three similarity parameters of momentum-loss-thickness Reynolds-number of 92 to 226 and Strouhal-number (reduced frequency) of 0.0001 to 0.0004 at the separation point, and an amplitude range up to 19 %. The free stream turbulence level was less than 1% .Upstream of the separation point the phase shift in the laminar boundary layer does not appear to be affected significantly bay either of the three parameters. The trend perpendicular to the wall is similar to the Stokes analysis. The problem scales well with the wave velocity introduced by Stokes, however, the lag of the main flow near the wall is less than indicated analytically. The separation point comes closest to the Stokes analysis but the phase is still 20 degrees lower at the wall.