A Computational Study on Turbulent Flow and Heat Transfer in a Strongly Curved RC/D = 0.65 Turbine Blade Cooling Passage

Abstract

It has been a common practice that serpentine cooling passages are used in gas turbine blade to enhance the cooling performance. Insufficient cooled blades are subject to oxidation, to cause creep rupture, and even to cause melting of the material. To control and improve temperature of the blade, we have to have a better understanding of flow behavior and heat transfer inside strongly curved U-bends. The interactions between secondary flows and separation lead to very complex flow patterns. To accurately simulate these flows and heat transfer, both refined turbulence models and higher-order numerical schemes are indispensable for turbine designers to improve the cooling performance. Previous studies have shown that the flow and heat transfer features through curved bends, even with moderate curvature, cannot be accurately simulated. It is the conventional belief and practice that the usage of a proper turbulence model and a reliable numerical method for achieving accurate computations. The three-dimensional turbulent flows and heat transfer inside a sharp U-bend are numerically studied by using a non-linear low-Reynolds number (low-Re) k-ω model in which the cubic terms are included to represent the effects of extra strain-rates such as streamline curvature and three-dimensionality on both turbulence normal and shear stresses. The finite volume difference method incorporated with the higher-order bounded interpolation scheme has been employed in the present study. For the purpose of comparison, the predictions with the linear low-Reynolds number k-ω model were also performed. The success of the present prediction indicates that the model can be applied to the flow and heat transfer through a coolant passage in an actual gas turbine blade. It is shown that the present non-linear model produces satisfactory predictions of the flow development inside the sharp U-bend comparing with linear Launder-Sharma model. In the present study, three turbulence models are used to predict Nysselt number distribution as well.