Traceable thermal imaging in harsh environments

TL;DR

This paper develops a traceable calibration method for thermal imagers to accurately measure surface temperatures of nuclear material containers in harsh environments, addressing a key challenge in metrology.

Contribution

It introduces a calibration approach that makes thermal imagers traceable to international temperature standards, improving measurement accuracy in nuclear safety applications.

Findings

Calibration uncertainties were less than 3.20°C for uncooled imagers.

Temperature differences on coated surfaces were reduced to 1.8°C.

The method enables deployment of thermal imaging in nuclear storage monitoring.

Abstract

Despite being regarded as a well-established field, temperature measurement continues to pose significant challenges for many professionals in the metrology industry. Thermal imagers enable fast, non-contact and a full field measurement, however there is a lack of metrological development to support their use. Here, thermal imagers have been examined for the monitoring of special nuclear material containers; the surface temperature is an important parameter for store management decisions. Throughout this research: a selection of thermal imagers were calibrated and made traceable to the International Temperature Scale of 1990; laboratory observations of a proxy steel plate were made; initial measurement of nuclear material storage containers were made; then a deployment to an inactive store was demonstrated. For this technique to be feasible, uncertainties less than 10$^\circ$C would be required. During the laboratory calibration of an uncooled and cooled thermal imager against blackbody reference sources, across the measured temperature range of 10$^\circ$C to 100$^\circ$C the uncertainties were less than 3.20$^\circ$C ($k=2$) and 0.50$^\circ$C ($k=2$) respectively. Here $k$ is the uncertainty coverage factor. When these calibrations were applied to the plate, regions of steel and higher emissivity coating were evaluated. These uncoated regions were measured with a thermal imager to demonstrate temperature differences compared to surface mounted thermocouples of 8.3$^\circ$C and uncertainties up to 30.1$^\circ$C ($k=2$). For the coated regions this temperature difference was reduced to 1.8$^\circ$C with uncertainties up to 6.8$^\circ$C ($k=2$).