Experimental Investigation of a High-Speed Turbine With Rainbow Rotor and Rim Seal Purge Flow

Abstract

The present paper addresses the experimental and numerical study of the unsteady flow established in a high-pressure turbine stage with rim seal purge. The HPT test section, operated at engine-relevant flow conditions in the high-speed turbine rig of the von Karman Institute, is heavily instrumented for high-resolution, high-bandwidth aerothermal measurements. A rainbow rotor setup allows the simultaneous testing of six different sectors, each hosting a specific tip and platform geometry optimized for enhanced aerodynamic performance. This paper focuses on the flow over the baseline sector, equipped with an axisymmetric hub platform and a squealer tip, with a purge flowrate matching 1.74% and 1% of the stage mass flow. The numerical study relies on Reynolds-averaged Navier–Stokes (RANS) computations with test-calibrated boundary conditions. Unsteady pressure and heat transfer measurements are performed at the rotor shroud. The maximum heat transfer is achieved along the front pressure side rim, whereas the squealer cavity generates a region of uniformly low static pressure and heat flux. The experimental adiabatic wall temperature is derived to quantify the thermal contribution to the global heat flux, demonstrating an increase in over-tip gas temperature up to 1.2 T01 above the pressure side rim. A Euler-based model is proposed to evaluate the temperature-driving work-exchange mechanism in the tip gap. The peak in casing heat transfer coefficient (750 W/m2K) is found at the tip leading edge. Time-resolved measurements of outlet total pressure, Mach number, and flow angle confirm the predicted phase and intensity of the tip leakage and upper passage vortex in the near-casing region. At 20% hr, the total pressure minimum is highly underpredicted by the RANS, indicating an inaccurate modeling of the interaction between rim seal purge and main flow. The measured impact of the purge flow variation is more significant than predicted by the RANS computations, with a negative offset of about 0.5% P01 in the lower 50% hr at lower purge flowrate.