A Detailed Uncertainty Analysis of Adiabatic Film Cooling Effectiveness Measurements Using Pressure-Sensitive Paint

Abstract

Pressure-sensitive paint (PSP) can be a powerful tool in measuring the adiabatic film cooling effectiveness. There are two distinct sources of error for this measurement technique: the ability to experimentally obtain the data and the validity of the heat and mass transfer analogy for the problem being studied. This paper will assess the experimental aspect of this PSP measurement specifically for film cooling applications. Experiments are conducted in an effort to quantifiably bound expected errors associated with temperature nonuniformities in testing and photodegradation effects. Results show that if careful experimental procedures are put in place, both of these effects can be maintained to have less than 0.022 impact on effectiveness. Through accurate semi in situ calibration down to 4% atmospheric pressure, the near-hole distribution of effectiveness is measured with high accuracy. PSP calibrations are performed for multiple coupons, over multiple days. In addition, to reach a partial pressure of zero the calibration vessel was purged of all air by flowing CO2. The primary contribution of this paper lies in the uncertainty analysis performed on the PSP measurement technique. A thorough uncertainty analysis is conducted and described, in order to completely understand the presented measurements and any shortcomings of the PSP technique. This quantification results in larger, albeit more realistic, values of uncertainty and helps provide a better understanding of film cooling effectiveness measurements taken using the PSP technique. The presented uncertainty analysis takes into account all random error sources associated with sampling and calibration, from intensities to effectiveness. Adiabatic film cooling effectiveness measurements are obtained for a single row of film cooling holes inclined at 20 deg, with CO2 used as coolant. Data are obtained for six blowing ratios. Maps of uncertainty corresponding to each effectiveness profile are available for each test case. These maps show that the uncertainty varies spatially over the test surface and high effectiveness corresponds to low uncertainty. The noise floors can be as high as 0.04 at effectiveness levels of 0. Day-to-day repeatability is presented for each blowing ratio and shows that laterally averaged effectiveness data are repeatable within 0.02 effectiveness.