Horseshoe Vortex Control with Leading Edge Endwall Boundary Layer Removal

Abstract

Endwall boundary layer removal was applied at the leading edge of a faired cylinder protruding normal from a flat plate. Flow visualization was used to qualitatively understand the impact of boundary layer removal on the formation of the leading edge endwall horseshoe vortex (HV). Total pressure and velocity data were measured using a Kiel probe and particle image velocimetry (PIV). Total pressure losses were measured for five suction rates (0, 6, 11, 15, and 23%). Suction rate is defined as the fraction of the total boundary layer mass flow removed through the suction slot. The total boundary layer mass flow is the boundary layer mass flow that would pass through the cylinder region in an undisturbed flow (without the cylinder). The losses decreased by ~30% with a suction rate of 11.0%. Vorticity contours from the PIV data compared the horseshoe vortex structure with and without endwall suction. Suction effectively removed the adverse pressure gradient responsible for the formation of the horseshoe vortex. Consequently, when suction is applied the vortex system does not form. Boundary layer velocity profiles extracted from the PIV data confirmed that boundary layer separation did not occur when suction was applied. In fact, the near wall boundary layer accelerated due to the suction. This acceleration resulted in the formation of a vortex pair in the region of the suction slot. This vortex pair increased in size and magnitude with increasing suction rates above 11.0%. It also migrated away from the suction slot. It was hypothesized that this vortex pair is responsible for the decreasing effectiveness of suction on the total pressure losses for suction rates greater than 11.0%. PIV results also showed that suction reduced the corner vortex responsible for elevated heat transfer in the corner region. The knowledge gained from this study will be used in future studies to eliminate the HV system in a turbine cascade.