Date of Award

January 2016

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Mechanical Engineering

First Advisor

Forrest Ames

Abstract

The purpose of this study is to experimentally investigate the effects of high free-stream turbulence on an aft loaded vane designed with a large leading edge. The large leading edge vane design was chosen to help reduce the heat transfer levels on the leading edge of the vane, while also generating a large enough area in the leading edge to allow the vane be cooled internally using a double wall cooling design. Heat transfer measurements were acquired on a linear cascade using a constant heat flux technique to help determine the effects of turbulence over this vane design. The cascade that was used was designed as a four vane, three passage linear cascade with adjustable bleed flows and tailboards which are adjustable to the periodic streamlines of the blade to blade analysis. The heat transfer measurements were taken at a wide range of Reynolds numbers, ranging from 500,000 to 2,000,000, and seven different turbulence levels. The turbulence levels range from a low turbulence condition of 0.7% to a high turbulence condition of 17.4%. The other turbulence conditions are a small grid far condition (Tu = 3.5%), a small grid near condition (Tu = 7.9%), a large grid condition (Tu = 8.0%), an aero-combustor with a decay spool condition (Tu = 9.3%), and an aero-combustor closely placed to the cascade (Tu = 13.7%). Inlet and exit pressure distributions along with vane pressure distributions were taken to help ensure the aerodynamic accuracy of the cascade. The heat transfer levels taken at these turbulence levels were correlated in terms of the approach flow Reynolds number and the turbulence condition. This was then compared to recent cylindrical leading edge test surface data using the TRL parameter. The surface heat transfer measurements that were taken were based off of the exit Reynolds numbers and displayed in terms of the Stanton number. These were then compared to predictive comparisons generated from a boundary layer calculation (STAN 7), using an algebraic turbulence model (ATM), and a transition model (Mayle). At the low turbulence levels the predictions show a close resemblance to the heat transfer distributions based on the exit Stanton number. As the turbulence levels increase the predictions tend to under predict at the stagnation region of the vane and tend to predict early transition on the suction surface of the vane. The inaccuracies in the transition prediction indicate a need to account for the complex curvature effects on the suction surface. Also to help accurately predict the heat transfer the physics of the turbulent response in the stagnation region would have to be included in the modeling. Also, these inaccuracies show some of the challenges faced when using engineering turbulence models. Later, the heat transfer distributions were then compared to suction surface heat transfer data taken with a heated and unheated endwall to show the impact of secondary flows on heat transfer in this region. The heat transfer distributions taken with a heated endwall showed a dramatic increase in the heat transfer levels at the edge of the vane and should provide a more relevant representation of the heat transfer levels in this region.

Download

Share

COinS