An experimental study on the effects of blade row interactions on aerodynamic loss mechanisms in a multistage compressor

While the gas turbine engine has existed for nearly 80 years, much of the complex aerodynamics which governs compressor performance is still not well understood. The unsteady flow field consists of periodic blade row interactions from the wakes and potential fields of each blade and vane. Vane clocking is the relative circumferential indexing of adjacent vane rows with the same vane count, and it is one method to change blade row interactions. Though the potential of performance benefits with vane clocking is known, the driving flow physics have yet to be identified. This research examines the effects of blade row interactions on embedded stator total pressure loss and boundary layer transition in the Purdue 3-stage axial compressor. The inlet guide vane, Stator 1, and Stator 2 all have 44 vanes which enable vane clocking of the embedded stage, while the rotors have different blade counts producing amplitude modulation of the unsteady interactions. A detailed investigation of corrected conditions is presented to establish repeatable, compressor performance year-round in a facility utilizing ambient inlet conditions. Without proper humidity accounting of compressor corrected conditions and an understanding of the potential for inlet temperature changes to affect clearances due to thermal growth, measurements of small performance changes in detailed research studies could be indiscernible. The methodology and implementation of a powder-paint flow visualization technique along with the illuminated flow physics are presented in detail. This method assists in understanding the loss development in the compressor by highlighting stator corner separations and endwall flow patterns. Effects of loading condition, rotor tip clearance height, and stator wake and rotor tip leakage interactions are shown with this technique. Vane clocking effects on compressor performance were quantified for nine loading conditions and six clocking configurations - the largest vane clocking dataset in the open literature. These data show that vane clocking effects are small at low loading conditions, including peak efficiency operation, but become stronger as loading increases, and then eventually lessen at near stall operation. Additionally, stator wake profiles and flow visualization reveal that total pressure loss changes are due to a corner separation modulation between clocking configurations. To further address these clocking trends, high-frequency response data were acquired at the Stator 2 inlet and along the Stator 2 surface. The unsteadiness at the Stator 2 inlet was quantified with detailed radial traverses for the different clocking configurations. These data show the effects of interactions between the Stator 1 wake and Rotor 2 tip leakage flow, which result in significantly different inlet flow conditions for Stator 2. The high unsteadiness and blockage region formed by the rotor tip leakage flow changes in size and shape between clocking configurations. Finally, measurements of the Stator 2 surface flows were acquired to investigate the vane clocking effects on unsteady surface pressures and boundary layer transition. These data reveal that Stator 2 performance is influenced by blade row interactions including rotor-rotor interactions, stator wake-rotor tip leakage flow interactions, and vane clocking