The failure of a turbine airfoil is a local phenomena. However, to date, the design of these airfoils has been based on steady state heat transfer tests that are capable of yielding only locally averaged data. To overcome this limitation, a transient technique using active surface coatings has been developed and is capable of yielding local data. This technique has been used to determine the Nusselt number distributions within augmented passages typical of gas turbine airfoils. However, certain assumptions have been made in these analyses without verification. This paper will address this aspect of the problem, as well as an improved data reduction procedure, and an alternative error analysis.
The data reduction procedure has been improved by incorporating a higher order approximation to the convective boundary condition, and by introducing a means of calculating the fluid bulk temperature-time-space profile. An image analysis system which yields an unbiased means of determining the time required for the surface to reach a specified temperature is introduced. Furthermore, it was observed that for augmented surfaces, the one dimensional conduction assumption made in the heat transfer solution is not valid for all times. Finally, treating the experimentally obtained quantities as values that are randomly distributed about some true value is not correct for all experimentally measured quantities.
