Impact of a Skewed Inlet Boundary Layer on the Aerodynamic Performance of a Stator-hub Equivalent High-turning Compressor Cascade

H3 - Fan Performance (ii)

The paper reports on numerical investigations into the effects of inlet boundary layer skew on the aerodynamic performance of high turning, 2D compressor cascades. Cascades of this type are representative of the stator hub sections in highly loaded single-stage axial-flow low-speed compressors. All simulations were done with the commercially available, steady three-dimensional RANS solver CFX 14.5 of ANSYS. Preliminary simulations were performed with different turbulence and transition models, along with a grid sensitivity analysis. The shear stress transport (SST) k-w turbulence model of Menter turned out to be best suited for the present investigation with a focus on turbulent endwall and secondary flows. The stator-hub equivalent high-turning compressor cascade consists of 2D blades with 8% standard NACA 65 thickness distribution on 64.5° circular arc camber lines. The blade aspect ratio is h/l=1.0, the space/chord ratio t/l=0.5 and the stagger angle lambda=25°. The midspan inlet angle was varied between beta1=50° and 58°. Reynolds number and endwall boundary layer thickness at inlet were hold constant, Re1=5*10^5 and d/l=0.1 respectively.
In axial turbomachines, for example axial flow compressors, the relative motion between adjacent blade rows (rotor/stator, stator/rotor) causes the endwall boundary layers to be skewed and reenergized. This phenomenon has been picked up in the present paper for the particular case of a compressor hub flow at rotor exit and stator inlet. A boundary layer leaving a rotor hub with large velocity deficiencies is reenergized and has a new start as it enters the stator hub. This phenomenon has been investigated in a simplified manner using a linear cascade model with skewed boundary layers at the inlet of the cascade. The more important results of the investigation may be summarized as follows: i) no leading edge separations were found in spite of very high incidence angles next to the endwalls, ii) the overturning of the endwall flow was much less with inlet skew than without thus indicating a passage vortex of considerably reduced strength, iii) due to ii) the interaction between the endwall flow and the blade suction surface flow was less with inlet skew than without resulting in lower net total pressure losses (without inlet losses), iv) the performance improvements caused by a skewed inlet boundary layer decreased with increasing inlet angle.