Modelling Considerations for a Transonic Fan
Abstract
The objective of this work is to provide a computational baseline for modelling the flow physics in the tip region of a transonic fan. A transonic fan was donated by Honeywell Aerospace to the Purdue University High-Speed Compressor Research Laboratory for the purposes of studying casing treatments and inlet distortion under the Office of Naval Research Power and Propulsion Program. The purpose of casing treatment is to extend the stall margin of the fan without being detrimental to fan efficiency. Hence, before an effective casing treatment can be designed, understanding the instabilities that lead to stall or surge and understanding the flow field near the rotor tip at different operating conditions is necessary. The behavior of the flow field was studied at design speed using steady simulations for near stall, peak efficiency, and choke operating conditions. The details of the passage shock, tip leakage vortex, and the shock-vortex interaction were investigated. The passage shock moves forward in the rotor passage as the loading increases, until eventually becoming unstarted near stall. The tip leakage vortex convects from the rotor tip leading edge to the pressure side of the adjacent blade, and its trajectory becomes parallel to the rotor inlet plane as the loading increases. The shock-vortex interaction does not cause the tip leakage vortex to breakdown, although distortion of the shock front and diffusion of the tip leakage vortex is significant near stall. To validate this computational model, steady simulations were used to conduct a grid convergence study. A single passage mesh of 8 million elements is sufficient to capture the flow qualitatively, but a mesh of at least 22 million elements is recommended to lower discretization error if quantitative details are important. A brief comparison of turbulence models is made, and the SST model was found to predict stronger radial flows than the BSL-EARSM and BSL-RSM models. However, the SST model still captures the flow features qualitatively, and the more complex models would be too costly for iterative design simulations. The importance of unsteady effects was also considered for a point near peak efficiency. Near peak efficiency, the effect of shock oscillations near the rotor shroud are small. Compared to steady simulations, the unsteady simulation predicts a slightly stronger horseshoe vortex at the hub and a passage shock closer to the rotor leading edge. The tip leakage vortex trajectory appears to be the same between the steady and unsteady simulations. The modelling decisions made in this research are currently only based on comparison between simulations. This model will be calibrated with experimental data in the future to provide a more accurate view of the flow physics inside this transonic fan. The modelling decisions made in this research are currently only based on comparison between simulations. This model will be calibrated with experimental data in the future to provide a more accurate view of the flow physics inside this transonic fan.
Degree
M.Sc.
Advisors
Key, Purdue University.
Subject Area
Design|Fluid mechanics|Marketing|Mechanical engineering|Mechanics
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.