| Popis: |
An experimental investigation was performed to assess the ability of infrared (IR) thermography methods to measure heat fluxes into a stainless steel flat plate in a medium-duration blowdown facility. Experimental work has shown that valuable insight can be gained from experiments performed with full-scale rotating rigs but IR thermography has generally been limited to experiments with plastic-like test articles and steady state or long-duration measurements. The goal of this experiment is to further develop IR thermography methods for eventual use in short-duration experiments investigating full-scale rotating turbines. This requires the ability to determine the heat flux on the surface of a metal blade exposed to hot flow for a very brief time.In order to develop the methodology that will be necessary for these rotating experiments, a set of stationary experiments was undertaken in a blowdown wind tunnel. The heat fluxes measured with IR thermography were compared side by side to heat fluxes measured by double-sided Kapton heat-flux gauges, which have been used extensively in full-scale rotating rigs. Two numerical models were used to post-process the IR-measured surface temperatures to calculate heat fluxes. The first was a one-dimensional finite volume method (FVM) model and both semi-infinite and symmetric boundary conditions were applied. The second was a two-dimensional finite element method (FEM) model in ANSYS with symmetric boundary conditions. The blowdown wind tunnel provides a challenging test case for these methods since the freestream total temperature changes significantly over the duration of the experiment.Results initially showed poor correlation between the heat fluxes measured with the IR methods and the heat fluxes measured by the heat-flux gauges. This discrepancy is attributed to the effects of a paint layer that was applied to the surface of the test plate to increase its emissivity for IR measurements. It was found that incorporating the effects of the paint layer into the model brought the two methods into agreement with average absolute deviations of 10% of the maximum heat flux value over the majority of the run. Additionally, it was shown that the slow response time of the IR camera compared to the heat-flux gauges caused it to miss the transient peak of the heat-flux. These experiments represent the first step towards accurately measuring the heat transfer to a blade in a turbine operating at design corrected conditions. By incorporating a faster cooled camera and further refining data reduction techniques to account for film cooling, the results can be expanded to enable a new type of experimental data set. |
