PREDICTION OF TIP LEAKAGE VORTEX DYNAMICS IN AN AXIAL COMPRESSOR CASCADE USING RANS ANALYSES

The tip leakage flow in turbine and compressor blade rows is responsible for a relevant fraction of the total loss. It contributes to unsteadiness, and have an important impact on the operability range of compressor stages. Experimental investigations and, more recently, scale-resolving CFD approaches have helped in clarifying the flow mechanism determining the dynamics of the tip leakage vortex. Due to their continuing fundamental role in design verifications, it is important to establish whether RANS/URANS approaches are able to reproduce the effects of such a flow feature, in order to correctly drive the design of the next generation of turbomachinery. Base studies are needed in order to accomplish this goal. In the present work the tip leakage flow in axial compressor rotor blade cascade have been studied. The cascade was tested experimentally in Virginia Tech Low Speed CascadeWind Tunnel in both stationary and moving endwall configurations. Numerical analyses were performed using the TRAF code, a state-of-the-art in-house-developed 3D RANS/URANS flow solver. The impact of the numerical framework was investigated selecting different advection schemes including a central scheme with artificial dissipation and a high-resolution upwind strategy. In addition, two turbulence models have been used, the Wilcox linear k model and a non-linear eddy viscosity model (Realizable Quadratic Eddy Viscosity Model), which accounts for turbulence anisotropy. The numerical results are scrutinized using the available measurements. A detailed discussion of the vortex evolution inside the blade passage and downstream of the blade trailing edge is presented in terms of streamwise velocity, streamwise vorticity, and turbulent kinetic energy contours. The purpose is to identify guidelines for obtaining the best representation of the vortex dynamics, with the methodologies usually employed in routine design iterations and, at the same time, evidence their weak aspects that need further modelling efforts.