Increasing the firing temperatures in gas turbines require better, and highly efficient means of heat removal of turbine blades so that metal temperatures stay within the limit of safe operation with respect to metal properties. This study focuses on the trailing edge region of a turbine blade. Ribs were added into a pin fin array in order to achieve better heat transfer compared to pin fin arrays without additional ribs as they are commonly used.

Heat transfer measurements are obtained using the thermochromic liquid crystal technique (TLC) in a trapezoidal duct with pin fins and rib turbulators representing endwall cooling. The blockages due to pins are 35%, 50% and 65%. There are a total of 15 rows of pins in the streamwise direction, and 5 columns in the spanwise direction. The non-dimensional rib heights are 0, 0.27, 0.7 and 0.1. The minor angle of the trapeze is 14 degrees, the hydraulic diameter of the duct is 21 mm. The Reynolds Numbers tested, based on free stream velocity and the hydraulic diameter of the experiment, are 40,000 60,000 and 106,000. The test matrix for this study contains all possible blockage and rib height combinations for all three Reynolds Numbers tested.

Streamwise averaged and spanwise averaged Nusselt number augmentations are compared to the Dittus-Boelter baseline case, and are presented for the endwalls together with heat transfer results for the pins. A pitot probe was traversed at the inlet and exit of the wind tunnel in order to measure the inlet and exit velocity profiles. For the endwall heat transfer, it was found for all configurations, that a local maxima occurs around one pin diameter downstream of the pin and a local heat transfer minima occurs near two pin diameters downstream of the pin. Nusselt number augmentation is generally higher closer to the longer side of the trapeze. The same trend is seen for the pin heat transfer which is in the columns closer to the long side of the duct larger than on the short edge of the duct. This claim can be supported with the results from the velocity profile measurements. Through the length of the duct, the flow shifts from the nose region to the larger opening on the opposite wall. This effect is weaker at higher flow rates, higher blockages, and larger ribs since more flow resistance exists, and this resistance hinders the flow to move sideward.

Also, it is observed that increasing the blockage ratio as well as increasing the rib height, has a positive impact on heat transfer. It is also observed that increasing the Reynolds number causes a reduction in Nusselt number augmentation. At higher flow rates, the flow has higher momentum, and tends to be less impacted by the inclusion of the ribs, which results in the ribs being more effective at lower flow velocities. However, for low flow rates, the ribs only act as an extended surface, for higher flowrates though, the ribs act as turbulators as well which causes better mixing and a more evenly distributed heat transfer on the endwall.

In order to interpret the presented measurements correctly, a comprehensive uncertainty analysis was conducted, and all heat transfer results are reported accurately within 12.3%. Repeatability tests show a maximum difference of 6%.

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