Tuesday  |  Salon 9  |  02:40 PM–03:00 PM

#15973, Evaluating Uncertainties in TSA using Images Obtained from Low-cost Thermal Core Cameras

Recently ultra low cost, so called thermal core cameras (e.g. FLIR Lepton 3.5), based on microbolometer sensor arrays have become available. Several researchers have investigated their use for thermographic non-destructive evaluation (TNDE) and thermoelastic stress analysis (TSA). While qualitatively these cameras have been shown to be useful in TNDE and for qualitative TSA, currently there is no means of assessing measurement uncertainties. Therefore, it is unclear whether accurate, reliable, quantitative data can be obtained with such cameras e.g. as is required for TSA. Typically thermal core cameras operate with both inbuilt temporal and spatial smoothing applied by default. Previous work using higher grade FLIR AC655sc cameras has shown that the low-pass filtering combined with temporal smoothing has a serious deleterious effect on measurement of the transient temperature change developed in TSA. Similar work with a Lepton 3.5 camera has shown that the inbuilt software includes not only temporal but spatial smoothing making quantitative TSA practically impossible. To develop the full potential of the thermal core cameras in TNDE and TSA it is essential that their performance is evaluated in the context of measurement of transient signals. Hence, a custom software has been devised that can systematically disable the inbuilt smoothing of the cameras and enable fully quantitative TSA. The approach is demonstrated in a series of test on different materials subject to know transient temperature changes at different rates. A full performance evaluation is provided that determines the accuracy and measurement uncertainty (precision) of three nominally identical Lepton cameras.

Geir Ólafsson   University of Bristol
Roy Bullock   University of Bristol
Janice Dulieu-Barton   University of Bristol