Full surface endwall heat transfer distributions have been acquired in a staggered pin fin array with the use of an infrared camera. Values are presented at Reynolds numbers of 3000, 10,000 and 30,000 based on pin diameter and average velocity through adjacent pins. Average endwall Nusselt numbers agree closely with archival values at each Reynolds number. Locally averaged heat transfer levels show a substantial increase from the inlet through the first few rows and finally a nearly streamwise periodic condition in the second half of the eight row geometry. Increasing levels of heat transfer in the inlet region can be attributed to the leading edge vortex system, flow acceleration around pins, and the generation of turbulence. Distributions of turbulence intensity and turbulent scale are shown to help document the turbulent transport conditions through the array. Detailed endwall Nusselt number distributions are presented and compared at the three Reynolds numbers for the first four and last four rows. These detailed heat transfer distributions highlight the influence of the horseshoe vortex system in the entrance region and the wake generated turbulence throughout the pin fin array. Local velocity and turbulence distributions are presented together with local Stanton number and skin friction coefficient data to examine the aggressive nature of the turbulent mixing.

The objective of this research was to investigate pin fin midline heat transfer in terms of our understanding of stagnation region heat transfer for cylinders in cross flow and turbine airfoils. An experimental investigation was conducted in a staggered-pin fin array at Reynolds numbers of 3000, 10,000, and 30,000 based on the maximum velocity between cylinders. Midline distributions of static pressure and heat transfer were acquired for rows 1 through 8 at the three Reynolds numbers. Turbulence measurements and velocity distributions were acquired at the inlet and in between adjacent pins in rows using hot wire anemometry. One-dimensional power spectra were calculated to determine integral and energy scales. Midline heat transfer distributions are reported as the Nusselt number divided by the square root of the Reynolds number as a function of angle. In these terms, heat transfer was found to increase through row 3 for a Reynolds number of 30,000. After row 3, heat transfer diminished slightly. The Reynolds number for each row was recast in terms of an effective approach velocity, which was found to be highest in row 3 due to the upstream blockage of row 2. Based on this effective velocity the Nusselt number divided by the square root of the Reynolds number increased through row 4. These data indicate that heat transfer is highest in row 3 pins due to the highest effective velocity, while heat transfer augmentation due to turbulence is highest in row 4 and beyond. Hot wire measurements show higher turbulence intensity and dissipation rates upstream of row 4 compared to upstream of row 3. Generally, pressure, heat transfer, and turbulence measurements were taken at all rows, providing a better understanding of turbulent transport from pin fin arrays.

An experimental research program was undertaken to examine the influence of large scale high intensity turbulence on vane heat transfer. The experiment was conducted in a four vane linear cascade at exit Reynolds numbers of 500,000 and 800,000 based on chord length corresponding to exit Mach numbers of 0.17 and 0.27. Heat transfer measurements were made for four inlet turbulence conditions including a low turbulence case (Tu ≅ 1%), a grid turbulence case (Tu ≅ 7.5%), and two levels of large scale turbulence generated with a mock combustor at two upstream locations (Tu ≅ 12% & Tu ≅ 8%). The heat transfer data demonstrated that the length scale, Lu, has a significant effect on stagnation region and pressure surface heat transfer. The average heat transfer augmentation over the pressure surface was found to scale reasonably well on the relative level of dissipation. The stagnation region heat transfer correlated well on the {Tu ReD5/12 (Lu/D)−1/3} parameter of Ames and Moffat (1990). The dependence of heat transfer augmentation on Reynolds number was estimated to scale on the 1/3 power for the pressure surface. The absolute level of heat transfer augmentation was found to be highest near the stagnation region. The combustor closely coupled to the cascade produced an average augmentation on the pressure surface of 56 percent at a Reynolds number of 800,000.

An experimental research program was undertaken to examine the influence of large-scale high-intensity turbulence on vane exit losses, wake growth, and exit turbulence characteristics. The experiment was conducted in a four-vane linear cascade at an exit Reynolds number of 800,000 based on chord length and an exit Mach number of 0.27. Exit measurements were made for four inlet turbulence conditions including a low-turbulence case (Tu ≈ 1 percent), a grid-generated turbulence case (Tu ≈ 7.5. percent) and two levels of large-scale turbulence generated with a mock combustor (Tu ≈ 12 and 8 percent). Exit total pressure surveys were taken at two locations to quantify total pressure losses. The suction surface boundary layer was also traversed to determine losses due to boundary layer growth. Losses occurred in the core of the flow for the elevated turbulence cases. The elevated free-stream turbulence was found to have a significant effect on wake growth. Generally, the wakes subjected to elevated free-stream turbulence were broader and had smaller peak velocity deficits. Reynolds stress profiles exhibited asymmetry in peak amplitudes about the wake centerline, which are attributable to differences in the evolution of the boundary layers on the pressure and suction surfaces of the vanes. The overall level of turbulence and dissipation inside the wakes and in the free stream was determined to document the rotor inlet boundary conditions. This is useful information for assessing rotor heat transfer and aerodynamics. Eddy diffusivities and mixing lengths were estimated using X-wire measurements of turbulent shear stress. The free-stream turbulence was found to strongly affect eddy diffusivities, and thus wake mixing. At the last measuring position, the average eddy diffusivity in the wake of the high-turbulence close combustor configuration (Tu ≈ 12) was three times that of the low turbulence wake.

A four-vane subsonic cascade was used to investigate the influence of turbulence on vane film cooling distributions. The influence of film injection on vane heat transfer distributions in the presence of high turbulence was examined in part I of this paper. Vane effectiveness distributions were documented in the presence of a low level of turbulence (1 percent) and were used to contrast results taken at a high level (12 percent) of large-scale turbulence. All data were taken at a density ratio of about 1. The three geometries chosen to study included one row and two staggered rows of downstream film cooling on both the suction and pressure surfaces as well as a showerhead array. Turbulence was found to have a moderate influence on pressure surface film cooling, particularly at the lower velocity ratios. The strong pressure gradients on the pressure surface of the vane were also found to alter film cooling distributions substantially. At lower velocity ratios, effectiveness distributions for two staggered rows of holes could be predicted well using data from one row superposed. At higher velocity ratios the two staggered rows produced significantly higher levels of effectiveness than values estimated from single row data superposed. Turbulence was also found to reduce effectiveness levels produced by showerhead film cooling substantially.