Abstract

This experimental study characterizes the mean flow structure, distribution of Reynolds stresses, and the mechanisms affecting turbulence production in the entire rotor passage of an axial compressor. The stereo-PIV (Particle Image Velocimetry) measurements are performed in the JHU refractive index matched facility and cover an entire rotor passage of a 1.5 stage axial compressor. The rotor blades have an aspect ratio of 0.73 and a measured tip gap of 2.8% of the blade span. The primary focus is on pre-stall flow rate, but trends and distributions are compared to those obtained at a higher flow rate. While many flow features exist at both flow rates, they differ in size, magnitude, and location. The tip leakage vortex (TLV) roll up, breakdown, migration, and double leakage occur earlier, and the associated blockage covers a larger area and has a lower axial and higher circumferential velocity at pre-stall. Furthermore, instantaneous velocity samples and cavitation images at pre-stall showcase transient extreme events involving a backward leakage flow extending upstream of the rotor, spillage of the leakage flow and backflow vortices from one passage to the next around the leading edge (LE), and intermittent boundary layer separation on the blade suction side near the trailing edge. All of these intermittent phenomena affect the distribution of turbulence, including an extension of elevated turbulent kinetic energy (TKE) upstream of the rotor LE plane. The turbulence is particularly high around the TLV, at the interface between the backward leakage flow and the forward passage flow upstream of the TLV, and along the boundary of the blockage area radially inward of the TLV. Vortex breakdown expands the elevated turbulence area substantially. As a general observation, the turbulence is highly anisotropic and inhomogeneous. The spatial distribution and the evolution of the normal Reynolds stresses vary greatly with operating condition and among components. In most regions, the associated trends can be explained by examining the corresponding production rate terms. Finally, a comparison of the TKE and normal Reynolds stress distributions downstream of the rotor as well as the production rates across two compressors show striking similarities, demonstrating that the same flow features play dominant roles in both machines.

This content is only available via PDF.
You do not currently have access to this content.