Experimental heat transfer study in an intermediate turbine duct

Paper in proceeding, 2013

Due to demands from industry on improved efficiency, reduced CO2 and NOx and decreased noise levels, the trend of future aero engines show that turbofan engines are designed with higher by-pass ratio. Two-spool and three-spool turbofan engines are designed with an intermediate turbine duct that connects the high-pressure turbine to the low-pressure turbine in the two-spool engine and two intermediate turbine ducts from HPT to intermediate pressure turbine (IPT) and IPT to LPT in the three-spool engines. The design of agressive intermediate turbine ducts (high radial offset for a short axial length) for these engines enables the possibility to increase the energy efficiency of the aero engine. The flow and heat transfer in a turbine duct is very complex. The flow has large secondary structures and is sensitive to flow separation, which is difficult to predict with numerical methods. Limited information is available in open literature that can be used for validation of numerical methods. This paper presents an experimental study of the heat transfer in an aggressive intermediate turbine duct. The aim of this study is to measure the of a surface temperature distribution and convection heat transfer coefficient with very high resolution and precision on a loaded guide vane which is located inside the intermediate turbine duct. Furthermore, the experimental results are compared to CFD carried out with ANSYS CFX. This experiment was carried out in a state-of-the-art large-scale low-speed turbine facility at Chalmers University of Technology. The duct configuration investigated represents a modern counter rotating turbine design, with a flow turning structural vane. The facility includes a turbine stage which provides realistic inlet conditions into the duct and operates at realistic flow Reynolds number based on the ITD vane chord length. The measurements were performed by using an infrared camera. The results shows that the heat transfer coefficient predicted in the computations close to the shroud is not well predicted. There can also be seen areas where there is flow transition and boundary layer transition.