An experimental investigation of a final HPC stage including pre-diffuser performance

Kyle Louis Miller, Purdue University

Abstract

The objectives of this study were to experimentally investigate the performance of two highly loaded OGV/pre-diffuser designs, including the effect of increased rotor tip clearance. Data were acquired in the Purdue Compressor/Pre-diffuser/Combustor Research Facility, where two highly loaded OGV designs were investigated, one of which utilized an integrated OGV/pre-diffuser design. The rotor tip clearance was varied from 1% to 2% blade height using different inserts over the rotor tip allowing for a trenched rotor configuration. Both OGV/pre-diffuser designs perform the same amount of turning and diffusion, but the main difference between the two designs is the location of the pre-diffuser. Design A incorporates a standard OGV/pre-diffuser setup where the flow exits the OGV axially and turns radially within the pre-diffuser. Design B, however, integrates the pre-diffuser within the OGV, which requires some radial turning to be performed within the OGV passage. The advantage of Design B is a shorter pre-diffuser, which reduces the weight and length of the engine. To aid the radial turning within the passage, the OGVs of Design B are positively leaned, imparting a radial shift to the streamlines. Results showed that both OGV designs achieved the same performance (efficiency and total pressure ratio) at design loading. The stall margin at design speed was also similar for both designs. However, as the loading was increased, Design B slightly outperformed Design A in both total pressure ratio and efficiency. Flow visualization results showed that the flow through the OGV of Design A exhibited greater radial migration of end-wall fluid to the mid-span regions. This resulted in suction side boundary layer separation at mid-span leading to additional loss at higher loading. Static pressure measurements within the pre-diffuser showed that the area ratio of Design B was too aggressive, resulting in boundary layer separation and poor pre-diffuser performance for Design B. The increased rotor tip clearance had negligible effects on performance at design loading. However, as the compressor was loaded up, the rotor experienced a 10% decrease in efficiency at near-stall, and a drop in peak total pressure ratio, 2.6% for Design A and 3.0% for Design B, for the increased rotor tip clearance. While the effects on the rotor were significant, both OGV designs showed that no additional pressure loss was incurred at design loading. Since neither OGV was significantly affected by the increased tip leakage flow, the net result was a smaller drop in stage total pressure ratio and a relatively small decrease in both surge margin and efficiency for the stage.

Degree

M.S.M.E.

Advisors

Key, Purdue University.

Subject Area

Mechanical engineering

COinS