Utilization of Additive Manufacturing in the Development of Stationary Diffusion Systems for Aeroengine Centrifugal Compressors
Abstract
Rising costs and volatility in aviation fuel and increased regulations resulting from climate change concerns have driven gas turbine engine manufacturers to focus on reducing fuel consumption. Implementing centrifugal compressors as the last stage in an axial engine architecture allows for reduced core diameters and higher fuel efficiencies. However, a centrifugal compressor's performance relies heavily on its stationary diffusion system. Furthermore, the highly unsteady and turbulent flow field exhibited in the diffusion system often causes CFD models to fall short of reality. Therefore, rapid validation is required to match the speed at which engineers can simulate different diffuser designs utilizing CFD. One avenue for this is through the use of additive manufacturing in centrifugal compressor experimental research.This study focused on implementing a new generation of the Centrifugal Stage for Aerodynamic Research (CSTAR) at the Purdue Compressor Research Lab that utilizes an entirely additively manufactured diffusion system. In addition, the new configuration was used to showcase the benefits of additive manufacturing (AM) in evaluating diffusion components. Two diffusion systems were manufactured and assessed. The Build 2 diffusion system introduced significant modifications to the diffusion system compared to the Build 1 design. The modifications included changes to the diffuser vane geometry, endwall divergence, and increased deswirl pinch and vane geometries. The Build 2 diffusion system showed performance reductions in total and static pressure rise, flow range, and efficiencies. These results were primarily attributed to the changes made to the Build 2 diffuser. The end wall divergence resulted in end wall separation that caused increased losses. The changes to the diffuser vane resulted in increased throat blockage and lower pressure rise and mass flow rate.In addition to the experimental portion of this study, a computational study was conducted to study the design changes made to the Build 2 diffusion system. A speedline at 100% corrected rotational speed was solved, and the results were compared to experimental data. The simulated data matched the overall stage and diffusion system performance relatively well, but the internal flow fields of the diffusion components, namely the diffuser, were not well predicted. This was attributed to 16 using the SST turbulence model over BSL EARSM. The BSL EARSM model more accurately predicted the diffuser flow field to the SST model.
Degree
M.E.
Advisors
Key, Purdue University.
Subject Area
Design|Industrial engineering
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.