Investigation of rotational effects on a gas turbine internal cooling system

The objective of the present thesis, which lies within a national research project, is the analysis of a "cold bridge" leading edge cooling system for a high pressure gas turbine blade, with the aim to study the combined effects of jet impingement and mass flow extraction on heat transfer phenomena. Both experimental and numerical investigations have been carried out, with the main aim to assess the effects of rotation on the heat transfer distribution in a realistic leading edge internal cooling system. Experiments were performed in static and rotating conditions replicating the typical realistic range of jet Reynolds number (Rej) and Rotation number (Roj) up to 0.05 for three cross-flow cases representative of the working condition that can be found at blade tip, midspan and hub, respectively (CR of 10%, 40% and 70%). Experimental results show a significant dependency of Nusselt number value on Reynolds jet number which does not appreciably affect its distribution on the LE. On the contrary a mild relation has been found with respect to cross-flow conditions, which does not remarkably affect the averaged Nu value but sensibly impact on its pattern. Concerning rotating conditions, the aero-thermal field turned out to be rather complex, but a good agreement between heat transfer coeffcient and flow field measurement has been found. In particular, jet bending due to rotation strongly depends on cross-flow intensity, while Ro has a weak effect on both jet velocity core and area-averaged Nusselt number. A parallel computational analysis of the rig has also been performed, aimed at identifying a suitable numerical model to investigate such phenomenon in order to deepen its physical comprehension. Stationary tests with uniform extraction conditions have been simulated on a symmetric model by means of both a RANS approach (using a k-w SST turbulence model) and two hybrid RANS-LES models: Scale Adaptive Simulation (SAS) and Detached Eddy Simulation (DES). RANS, but also SAS and DDES approaches, present a slightly underestimation of heat transfer experimental values, but the heat exchange distribution shape is accurately reproduced by the models such as the HTC distributions and values trend with the investigated parameters. DDES has been exploited for the numerical investigation in rotating conditions. A fairly good agreement with experimental measurements is observed, which represent a further validation of the adopted computational model.