The present study is a three-dimensional numerical investigation of the effectiveness of film cooling for a turbine blade leading-edge model with both a single and a three-hole cooling configuration. The model used has the same dimensions as those in the experimental investigation of Ou and Rivir (2006). It consists of a half cylinder with a flat after-body, and well represents the leading edge of a turbine blade. The single coolant hole is situated approximately at the spanwise center of the cylindrical model, and makes an angle of 21.5 degrees to the leading edge and 20 degrees to the spanwise direction. For the three-hole configuration, the center hole is positioned the same as the single hole in the single-hole configuration, with the adjacent holes located at a spanwise distance of 37.4 mm on either side of the center hole. Multi-block grids were generated using GridGen, and the flows were simulated using the flow solver Fluent. A highly clustered structured C-grid was developed around the leading edge of the model. The outer unstructured-grid domain represents the wind tunnel as used in the experimental study of Ou and Rivir (2006), and the leading-edge model is located at the center of the domain. Simulations were carried out for blowing ratios, M, ranging from 0.75 to 2.0. Turbulence was represented using the k-? shear-stress transport (SST) model, and the flow was assumed to have a free-stream turbulence intensity of 0.75%.
Two types of boundary conditions were used to represent the blade wall: an adiabatic surface, and a conductive surface. The adiabatic-wall results over-predicted the film-cooling effectiveness in the far downstream region for low blowing ratios. Also, in the vicinity of the cooling hole, an increase in blowing ratio resulted in higher film cooling effectiveness than observed in the experiments. It should be noted that the steady RANS-based turbulence model used under-predicts the interaction between the coolant and mainstream flow near the cooling-pipe exit. The conductive-wall results show a much closer agreement with experimental data for film effectiveness as compared to the adiabatic-wall predictions. Simulations were also performed with higher values of turbulence intensity at the cooling-hole inlet, and these predicted the coolant-mainstream interaction and the film-cooling effectiveness more accurately.
Finally, a novel concept of pulsing the coolant flow was implemented so as to achieve film-cooling effectiveness equivalent to that with constant cooling, but with reduced overall coolant air, thereby enhancing turbine efficiency. Pulsed cooling with pulsing frequency PF = 5 and 10Hz, and duty cycle DC = 50%, shows the greatest cooling effects. The three-hole cooling results indicate that the 49 mm spanwise distance used for computing the spanwise-averaged values for film-cooling effectiveness accounts for all of the film-coolant spreading provided by the single hole. Also, the neighboring cooling holes contribute little film cooling to the 49 mm spanwise distance. The most significant new finding in this work is that the inclusion of wall conductance is the main factor responsible for reproducing the experimental data.