| Summary: | 博士 === 國立臺灣大學 === 機械工程學研究所 === 85 === The purpose of present work is to investigate the convection transport phenomena over a smooth turbine blade or a shower-head film cooling blade under the influence of mainstream turbulence intensity or the upstream wake. The film cooling technique has been widely used in the aerospace industry to protect the hot components of gas turbines. However, the flow field and heat transfer characteristics produced by the interaction of main-stream and coolant injection flows still remain unknown. Both the measured local and spanwise-averaged Sherwood number distributions can be served as the data base for explanation these characteristics. By applying the heat-mass transfer analogy, the corresponding heat transfer coefficient can be evaluated from the measured mass transfer coefficient.
The present study also sets up an automated measuring system which is equipped with a quartz crystal microbalance, to measure the diffusion coefficient of naphthalene vapor into air at various temperatures. An empirical correlation between the diffusion coefficient of naphthalene vapor with temperature is evaluated for further data reduction.
All measurements were conducted in a blowing-type wind tuimel with a planar blade cascade and the naphthalene sublimation technique was employed. In the present study, several important parameters such as main-stream turbulence level, blowing ratio, and upstream passing wake which significantly affects the heat (mass) transfer coefficient were varied for investigation their effect on the turbine blade surface heat (mass) transfer distribution while the exit Reynolds number was kept at Re2＝ 397,000. The main-stream turbulence level is promoted by installing a bar-grid at different locations in the test section of wind tunnel. The upstream passing wake is generated by placing a cylinder with the same diameter of the trailing edge of blade at four different locations upstream of the blade cascade.
Without the endwall effect, measured results indicate that a separated-flow transition occurs on the convex surface near the trailing edge and the Taylor-Gortler vortices are observed over the concave surface as the mainstream turbulence level is low and no wake for a smooth blade. When the mainstream turbulence intensity is enhanced or a wake is generated right ahead of the leading edge of the test blade, both the A vortices over the convex surface or the Taylor-G o riter vortices over the concave surface are found to be absent. The mass transfer rate is augmented and the onset of transition point is moved more upstream as increasing the mainstream turbulence intensity or by generating a wake right ahead of the test blade.
The impact of vortices initiated near the endwall extends three dimensional effects to about 0.3 chord lengths on the convex surface and to about 0.02 chord lengths on the concave surface away from the endwall. Some small but intense corner vortices result in an obvious mass transfer augmentation near the endwall region.
For a shower-head film cooling blade without the influence of endwall, the existence of secondary injection flows can obviously enhance the mass transfer rate over both sides of blade and the onset of transition point is moved upstream. On the nearby region of blade leading edge, it was found a strongly periodic spanwise distribution of the local mass transfer coefficient at the tested range of mainstream turbulence intensity. The upstream passing wake has the similar influence on the blade mass transfer rate. Generally, the blowing ratio is proportional to the mass transfer rate for both sides of blade. At low blowing ratios (M≦1.0), the enhancement of mainstream turbulence intensity is not always accompanied with an increased of mass transfer rate over both surfaces of show-head film cooling blade. This is due to two different mechanisms, local turbulence intensity effect and the film cooling effect result in the enhancement of mass transfer rate. The twin-peak phenomena in the spanwisely local mass transfer distribution can be considered as the evidence of the existence of kidney vortices produced by jets. As the injection holes located in the front stagnation line is closed to only I hole pitch away from the endwall, the passage vortex is no longer attached on the convex side of blade.
|
|---|
