Models simplified from real annular turbine cascades into linear cascades have been widely used in the experimental investigations on the cooling performances of turbine blade and endwall (including hub and shroud), but errors caused by the simplifications are seldom considered and discussed. This paper presents a numerical comparison of film cooled endwall between real annular and linear cascade. The model used in this work is derived from a real gas turbine stator, which consists of 46 vanes. Numerical simulations are carried out by the commercial software ANSYS CFX 15.0 and ICEM 15.0. Conjugate Heat Transfer (CHT) analysis is used to solve pressure field and temperature field through a fully implicit multi-grid coupling algorithm, and the two-equation turbulence model, Shear Stress Transport (SST) k-ω, is selected as closed equations. To validate the numerical strategy, a series of experiments is conducted by a flat endwall model, in the hot gas wind tunnel at University of Science and Technology of China (USTC). Using the validated numerical strategy, the fluid flow and heat transfer characteristics of the real annular and flat endwalls models are simulated and compared. The differences in static pressure, temperature and overall cooling effectiveness between the two models are exhibited, and the reliability and reasonability of the simplification from a real annular cascade to a linear cascade are discussed. The results show the following important conclusion: If the pressure problems of hub endwall are considered only, the flat endwall simplification is acceptable, but at experimental conditions, the cooling effectiveness of the flat endwall could be 10% higher than that of the real annular endwall, while at realistic condition, the difference decreases obviously.

This content is only available via PDF.
You do not currently have access to this content.