A Comparative Study of RANS, URANS and NLES Approaches for Flow Prediction in Pin Fin Array

Gas Turbine are nowadays largely used for aircraft propulsion and land-based power generation. The increased attention to environmental aspects has promoted research and development efforts both from manufacturers and research centres. The latest developments in turbinecooling technologies play a critical role in the attempt to increase the efficiency and the specific power of the most advanced designs. Pin fin arrays, in particular, are widely used in jet engine application because of their ability to enhance cooling by providing extended surfaces for conduction and convection. They are also known to be an effective means to create turbulence which naturally increases heat transfer. Pin fin turbulators are typically located inside the trailing edge of high pressure turbine blade where they also act as a structural support. The optimum shapes and spacing of such elements are usually determined experimentally, or more recently, by using Computational Fluid Dynamics (CFD). On the other hand, the comprehension of the real physics controlling the heat transfer enhancement process and the role played by the large scale vortical structures generated by the inserts, still represent a great challenge for fluid mechanic researchers. The problem has been intensively investigated by Ames et al. (2005) by means of an experimental campaign on pin fin matrix. From the numerical point of view, the principal bottleneck of the CFD approach as applied to this kind of massively unsteady flow is related to the high computational cost and to the reliability of the turbulence models. The main objective of this work is to offer a critical analysis of the performance of a cooling device consisting of a pin fin turbulators geometry, as predicted by different CFD models of various complexity, using similar computational technology to integrate the corresponding governing equations. Local velocity and turbulence distributions are presented and compared with available experimental data.