Hybrid RANS/LES study of complex turbulence characteristics and flow mechanisms on the highly-loaded turbine endwall

Abstract The strong secondary flow and the interaction between multiple vortices result in substantial aerodynamic losses in the endwall region of the highly-loaded turbine. Accurate predictions of complex turbulence characteristics propose severe challenges to the widely used Reynolds-Averaged Navier-Stokes (RANS) approach. This work adopts a high accuracy hybrid RANS/Large Eddy Simulation (LES) method on a highly-loaded turbine blade for which the Reynolds number is 611,000 and the exit Mach number is 0.8. Results verify that, in contrast with the RANS method, the hybrid method can provide more accurate predictions and detailed small-scale flow structures than the RANS simulation. Analysis of the Reynolds stresses predicted by the hybrid method reveals the complex turbulence characteristics in the endwall region and highlights shortcomings of Reynolds stresses predictions existing in the RANS method in the highly-loaded turbine blade environment. To analyze turbulence characteristics near the blade endwall, the Lumley triangle method is applied to quantify the turbulence anisotropy degree. Based on fine flow field structures, the proper orthogonal decomposition (POD) method is used to extract the dominant flow structures from the unsteady hybrid method results. The turbulence kinetic energy budget is employed to study dominant terms in turbulent flow. The total pressure loss coefficient based on the first law of thermodynamics and the entropy generation rate (EGR) based on the second law of thermodynamics are both used to study the loss generation mechanism in the endwall region. Predicted turbulence characteristics and flow mechanism can provide guide towards low loss design and flow control of the turbine blade.