Secondary flow loss reduction in highly loaded turbines

Date of Completion

January 2007


Engineering, Aerospace|Engineering, Mechanical




Experimental results are presented for several airfoil leading edge modifications to assess their impact on turbine aerodynamic efficiency and airfoil loading. A baseline (Langston) configuration, a small and large leading edge bulb, and a leading edge fillet were tested in a large-scale, low aspect ratio, high turning linear cascade. Results show that while the fillet geometry reduced overall loss by approximately 7%, the small and large bulbs did not exhibit a loss reduction. For the fillet, overall turning was slightly reduced, while for the small bulb turning increased slightly. Thus, this bulb shows potential for increasing airfoil loading without an associated loss penalty. ^ In addition to these experimental efforts, turbulent computational simulations of each configuration were performed using Fluent computational fluid dynamics (CFD) software. The Shear-Stress-Transport version of the k - ω turbulence closure is used. The exact conditions at which experimental data were measured were replicated in the CFD solution, and the resulting loss and shear values for the baseline configuration are compared with the experiments. It was found that CFD in general overpredicts both aerodynamic loss, as well as surface shear stress. The overprediction is most pronounced in the shear region associated with the passage vortex. Surface shear vectors from the CFD indicate an inaccurate prediction of endwall cross flow from the pressure to suction surface, showing near wall velocities that are overly aligned with the freestream direction. This inability to predict the strong crossflow appears to be related to an excessive amount of momentum transport from the freestream to the boundary layer due to the inability of the turbulence representation to accurately model relaminarization. ^