Development of temperature measurement method for gas turbine cooling application

Temperature measurement are one of the essential part in gas turbine cooling research. The resulting heat transfer coefficient and adiabatic wall temperature are two of the important information analysed from the temperature data. One dimensional semi-infinite heat transfer solution is widely used to solve for the heat transfer coefficient and adiabatic wall temperature. However, the experimental time for this solution was limited resulting in less temperature data for analysis. There is an issue regarding longer experimental time is needed to accurately calculate the heat transfer coefficient and the adiabatic wall temperature. A temperature measurement method was investigated to solve this issue. A test rig was designed to have similar test area to the wheel space area for a representative single stage gas turbine rig. Crank Nicolson finite difference method was proposed to solve for the internal temperatures of the test plate. In this work, the solution was designed to have two different back face boundary condition. First, an adiabatic back face boundary condition to simulate the one dimensional semi-infinite heat transfer condition. Second, a conductionconvection back face boundary condition to solve the time limitation issue. The resultant heat transfer coefficient from adiabatic back face boundary condition had an average of 2.5% difference and the adiabatic wall temperature had an average of 2% difference when compared to reference values. Duration for heat transfer experiments were longer for the conduction-convection back face boundary condition, at Fo = 0.7 rather than Fo = 0.1. This results in an increase of 40% more temperature data range for the heat transfer analysis. For these experiments, the conduction-convection back face boundary condition had an average of 5% difference in heat transfer coefficient and 3.5% difference in adiabatic wall temperature. Meanwhile, the adiabatic back face boundary condition had an average of 11.3% difference in heat transfer coefficient and 4.9% difference in adiabatic wall temperature when compared to reference values. Crank Nicolson solution method with conduction-convection back face boundary condition allowed more temperature data for analysis and provide more accurate heat transfer coefficient and adiabatic wall temperature values.

Text FK 2021 65 UPMIR.pdf Download (1MB)

Actions (login required)

View Item