This paper investigates, both experimentally and computationally, the heat transfer performance on an engine representative varying aspect ratio two-pass internal cooling channel, in both stationary and rotating conditions. The test geometry and design parameters were suggested by SNECMA as a representative HPT blade two-pass internal cooling channel. The cooling channel has radially outward flow in the first passage with an aspect ratio of 1:2.25 and after a 180 degree sharp turn, a radially inward flow in the second passage with an aspect ratio of 1:1.85. One side of the two passages is equipped with 45 degree angled rib turbulators with a rib spacing P/e=7 and blockage ratio e/Dh =0.116. The other side is smooth in order to have optical access for experiment.
The experiment was performed at three Reynolds numbers: 15,000, 25,000, and 35,000. Both forward and backward rotating directions were tested in order to study the heat transfer performance of the ribbed surface as trailing wall or leading wall individually. The tested Rotation numbers were Ro=±0.3 at Re=15,000 and Re=25,000, whereas the Rotation number was reduced to ±0.22 at Re=35,000, due to restrictions of the test facility. Infrared thermography technology is used to capture the temperature field for further evaluation of heat transfer performance.
Numerical simulations for all experimental cases were conducted using the same geometry including the air feeding system, applying the experimental wall temperature distribution in order to properly capture inlet and buoyancy effects, with the k–ω–SST turbulence model. Numerical results show overall agreement and similar trends than the experimental data. Numerical results also show that the rotation effects alter the internal flow significantly, resulting in different surface heat transfer distributions. Particularly, it is shown that heat transfer performance of the pressure side is not enhanced by the rotation in this study, which is a surprising result. This behavior was captured both in the experiments and the numerical predictions.