The Effect of Reynolds Number, Compressibility and Free Stream Turbulence on Profile Entropy Generation Rate

Detailed aerodynamic data from the suction surface boundary layer on a turbine blade arranged in a linear subsonic cascade was acquired under high free stream turbulence conditions (∼ 5.2%) generated using a perforated plate placed upstream of the cascade. In addition, data was also obtained from a transonic turbine cascade utilizing the same blade profile but of much smaller chord at free stream turbulence levels of 3.5%. Velocity profiles from the laminar, transitional and turbulent boundary layers were measured at various locations along the airfoil suction surface for the incompressible regime at Re_C of 76,000. For the compressible test cases, boundary layer velocity profiles were measured at two locations towards the aft section of the blade at Re_C of 163,000 and M_Ex of 0.37 respectively. For both cases the boundary layer velocity profiles were acquired by traversing a single normal hot wire probe normal to the blade surface. In addition the extent of the transition region over the blade surface was determined for both compressible and incompressible regimes by the use of an array of heated thin film sensors over a range of Reynolds and exit Mach numbers. It was observed that an earlier transition ensued at high free stream turbulence conditions in comparison to a previous investigation at comparable Re_C and lower turbulence level (0.8% Tu). In addition comparisons were made to existing incompressible data at Re_C = 185,000 and 0.8% free stream turbulence intensity. One of the primary observations resulting from an earlier transition was a thicker turbulent boundary layer, but in addition it was also noted that shear strain rates in the laminar boundary layer were significantly higher than those obtained at the 0.8% turbulence intensity. Further analyses also elucidated the presence of fluctuating components of velocity in the laminar boundary layer and were attributed to the effects of the free stream turbulence. This leads to the notion of a hybrid boundary layer, possessing both laminar and turbulent characteristics. These findings have implications regarding the profile loss of the blade, that is the loss generated in blade boundary layers and wakes normally associated with phenomena such as viscous shear, Reynolds stress production, shock wave formation and heat transfer across temperature differences and can be quantified in terms of the amount of entropy generated. For the purposes of this study entropy creation is solely restricted to that arising due to fluid dynamic phenomena, thereby assuming an adiabatic and quasi-isothermal flow. The entropy generation rate per unit volume is obtained directly from the boundary layer velocity profile; further integration gives rise to the entropy generation rate over the boundary layer at a point or over the entire suction surface length. Even though the number of quantitative measurement points on the transonic cascade was limited due to the very thin boundary layer present, no effects attributable to compressibility were observed on the entropy generation rate at the Mach number in question. Increased free stream turbulence had a greater effect on the generated entropy due to increased viscous shear in the laminar boundary layer and increased Reynolds stress production. In contrast, free stream turbulence did not have any significant effect on the turbulent boundary layer in the context of this study, as it was observed that the amount of entropy generated in the turbulent boundary layer was approximately equivalent for both turbulence levels at comparable Reynolds number.