Modern gas turbine designs' ever-increasing turbine inlet temperature raises challenges for the nozzle guide vane cooling. Two typical endwall cooling schemes, jump cooling and louver cooling, result in different interactions between the injected coolant and the mainstream, leading to different cooling effects. This study investigates these two cooling schemes on the endwall cooling experimentally and numerically. Wind tunnel tests and the CFD simulations are carried out with engine-representative conditions of an exit Mach number of 0.85, an exit Reynolds number of 1.5×10^6, and an inlet Turbulence intensity of 16%. The jump cooling scheme experiments investigate two blowing ratios, 2.5 and 3.5, two density ratios, 1.2 and 1.95, and three endwall profiles with different NGV-turbine alignments. Four coolant mass flow ratios from 1.0% to 4.0% are tested for the louver cooling.
The results show that the cavity vortex, the horseshoe vortex, and the passage vortex are the main factors that prevent the upstream coolant from reaching the NGV passage. The jump cooling scheme generally provides high momentum to the cooling jets. As a result, the coolant at the design case density ratio of 1.95 and blowing ratio of 2.5 is sufficiently energized to penetrate the horseshoe vortex. It then forms a relatively uniform coolant film near the NGV passage inlet, leading to a minimum adiabatic cooling effectiveness of 0.4 throughout the passage. Reducing the coolant density or increasing the blowing ratio leads to higher coolant momentum, so the coolant jets can further suppress the horseshoe vortex. However, high momentum may cause coolant lift-off, mitigating the coolant reattachment. Therefore, the density ratio needs to be carefully balanced with the blowing ratio to optimize the cooling effect. This balance is also affected by the combustor-NGV misalignment, as a higher step height requires higher coolant momentum to overcome the step-induced vortices.
On the contrary, the louver cooling scheme provides less momentum to the coolant. The results showed that only by exceeding a coolant mass flow rate of 1~2% can the coolant form a uniform film which provides good coverage upstream of the NGV passage inlet. As for the cooling of the NGV passage, the mass flow rate ratio of the range investigated is not sufficient for desirable cooling performance. The pressure side endwall proves most difficult for the coolant to reach. In addition, the fishmouth cavity at the combustor-NGV passage causes a three-dimensional cavity vortex that transports the coolant in the pitch-wise direction. Moreover, the coolant transport pattern is dependent on the coolant blow rate.
Overall, the more-energized coolant film generated by the jump cooling tends to survive longer, but it is also more prone to lift-off. At the same time, the less-energized coolant film caused by the louver cooling is more susceptible to vortices and the discontinuity of the endwall geometry. However, it develops faster, especially in the lateral direction. The two schemes could be applied simultaneously for an ideal cooling system. The jump cooling can provide enough momentum for the coolant to persist in the NGV passage. Meanwhile, the louver cooling covers the upstream region before the jump cooling coolant reattaches to the endwall. / Doctor of Philosophy / Gas turbines, sometimes called combustion turbines, are widely used to generate power or propulsion for various applications. The three main components of a gas turbine are compressor, combustor, and turbine. Modern gas turbines run at a high turbine inlet temperature that exceeds the current metal limits to increase efficiency. However, this brings significant challenges to the cooling of the first stage of the turbine, the nozzle guide vane. In this research, two commonly used endwall cooling methods, jump cooling and louver cooling, are investigated under engine-representative conditions experimentally and numerically. In addition, flow physics is demonstrated to explain the endwall cooling performance, mainly the upstream boundary layer caused by the interaction between the mainstream and the coolant flow.
The results show that the cavity vortex, the horseshoe vortex, and the passage vortex are the main factors that prevent the upstream coolant from reaching the NGV passage. The jump cooling scheme provides high momentum to the cooling jets. As a result, the coolant in the design case is sufficiently energized to penetrate the horseshoe vortex, providing a desirable cooling effect in the NGV passage. Reducing the density ratio or increasing the blowing ratio can help the coolant jets further suppress the horseshoe vortex but also causes more lift-off, which adversely affects the cooling performance.
On the contrary, the louver cooling scheme provides less momentum to the coolant, forming a less energized coolant film. The lack of coolant causes the louver coolant film to provide good coverage immediately downstream of the louver scheme exit. However, due to unfavorable interaction with vortices and endwall discontinuity, the cooling effect decays quickly downstream.
Overall, the more-energized coolant film generated by the jump cooling tends to survive longer, but it is also more prone to lift-off. At the same time, the less-energized coolant film caused by the louver cooling is more susceptible to vortices and the discontinuity of the endwall geometry. However, it develops faster, especially in the lateral direction. The two schemes could be applied simultaneously for an ideal cooling system to work mutually beneficially.
Identifer | oai:union.ndltd.org:VTETD/oai:vtechworks.lib.vt.edu:10919/110442 |
Date | 03 June 2022 |
Creators | Mao, Shuo |
Contributors | Mechanical Engineering, Ng, Wing Fai, Borgoltz, Aurelien, Son, Chang Min, O'Brien, Walter F., Diller, Thomas E., Paul, Mark R. |
Publisher | Virginia Tech |
Source Sets | Virginia Tech Theses and Dissertation |
Language | English |
Detected Language | English |
Type | Dissertation |
Format | ETD, application/pdf |
Rights | In Copyright, http://rightsstatements.org/vocab/InC/1.0/ |
Page generated in 0.0019 seconds