Experimental test of a 3D parameterized vane cascade with non-axisymmetric endwall

Abstract

Experimental test of a 3D parameterized vane cascade with non-axisymmetric endwall was conducted by using the particle image velocimetry (PIV) and five-hole probe. Both the straight annular cascade with axisymmetric endwall (denoted by Ref) and the negatively bowed cascade with non-axisymmetric endwall (denoted by Opt) were tested. Here, the Opt design was obtained through optimization based on Computational Fluid Dynamics (CFD), by combining the techniques of section profiling, compound lean and non-axisymmetric endwall [1]. The results showed that, the pressure distributions of blade surface at 10%, 50% and 90% spans of CFD results and experimental data are in good agreement. The trends of total pressure coefficient at cascade outlet (denoted by Cpt) and outlet flow angle (denoted by α1) were well matched; the Cpt of Opt design was shown increased by 0.47%, and 0.41% in experiment and CFD simulations, respectively. The contours of Cpt showed that the profile loss of Opt design is significantly reduced and the secondary loss region gets closer to hub and shroud. By tracking the secondary vortices near the hub of cascades, the affected area of secondary flow of Opt design was confirmed to be smaller when compared to that of the Ref design; but the size of passage vortex near the cascade exit was increased. The results of PIV and five-hole probe were consistent; they both confirmed that the combined parameterization of both blade and endwalls is effective in reducing the affected area of secondary flow and the profile loss of the low-aspect-ratio vane cascade.