The characteristics of fluid flow and heat transfer within a smooth three-pass channel of a real low pressure (LP) turbine blade have been investigated through experimental and numerical approaches. The serpentine channel consists of two inlet passes, two dividing walls, two 180 degree bends, twenty-five exits at the trailing edge, and two exits at the blade tip. In the experiments, purified water was used as working medium, the secondary flow patterns at five cross-sections were captured by a particle image velocimetry (PIV) system, the inlet Reynolds number was controlled by a turbine flow meter, and the mass flow rate ejected from each exit was measured by rotameters. Using the commercial software ANSYS CFX 13.0, numerical investigations were carried out. The practicability of four turbulence models, the SSG RSM, SST k-ω, RNG k-ε and standard k-ε models, were estimated. Through qualitative and quantitative comparisons of the secondary flow patterns, local velocity variation trends and mass flow rates between the experimental data and numerical results, the SSG RSM was selected as the most appropriate model in the following numerical investigations. Using ideal gas as working medium, the impacts of Reynolds numbers and rotation numbers on the heat transfer performances were numerically investigated. The numerical results predicted three interesting phenomena: 1) The locally averaged Nusselt number increases generally with the inlet Reynolds numbers. However, the increasing amplitude is significantly different from the correlation suggested by Dittus-Boelter, Nuo = 0.023Re0.8Pr0.4. The effect of the Reynolds number on the Nusselt number is substantially enhanced due to the serpentine channel design with two 180 degree-bends. The enhancement amplitude is described by two fitted coefficients based on Dittus-Boelter correlation. 2) Under a rotation condition, in the 1st and 3rd passes, the enhancement amplitude of the average Nusselt number on the pressure side (PS) is more significant than that on the suction side (SS), whereas in the 2nd pass, the enhancement amplitude on the PS is lower than that on the SS. 3) In the 3rd pass, a higher rotation number leads to a more uniform distribution of the local Nusselt number along the streamwise direction on both the PS and SS.

This content is only available via PDF.
You do not currently have access to this content.