Full-Field Measurement Uncertainty Quantification for Luminescence Phosphor Thermography on Surfaces in 3D Space

Background The luminescent intensity ratio (LIR) method of measuring temperature with thermographic phosphors has long been reported to exhibit sensitivities to experimental geometry as a test article moves in 3D space. Objective This work aims to evaluate the sensitivity of combined thermographic phosphor digital image correlation (TP+DIC) to experimental geometry and test article motion. In doing so, two novel geometry compensated formulations of LIR are derived that ensure consistent phosphor emission volume is measured in both cameras and compared against the traditional LIR methods. Methods Full-field uncertainty quantification techniques were applied to compare traditional LIR methods to the novel LIR methods as the phosphor moved throughout a measurement volume representative of a thermomechanical experiment. Three experiments were performed: (Experiment A) systematic rotation at room temperature to evaluate the angular emission profile of the phosphor coating, (Experiment B) systematic translation where temperature calibrations were taken at each fixed location in space, and (Experiment C) compound translation/rotation at two steady-state temperatures. Results Experiment A showed that the phosphor coating was non-Lambertian, and compensation was achieved using the viewing angle measured from DIC. Experiment B showed that while mean filtered traditional LIR measurements may yield accurate temperature measurements at a single position in space, they often suffer from significant bias during material motion that can be reduced by 3x by moving to geometry compensated LIR techniques. Finally, Experiment C confirmed that applying mean filters to traditional LIR methods causes significant over-confidence in the measurement and showed that the temperature calibration of the non-Lambertian geometry compensated LIR method remained reliable at all positions within the working volume. Conclusion Through all three experiments, it is shown that the geometry compensated, non-Lambertian (GCNL) method remains consistently aligned with the temperature calibration. Therefore, the new GCNL LIR method presented herein can be applied to complex thermomechanical experiments while using the prediction interval from the temperature calibration as a reasonable full-field uncertainty estimate throughout deformation.