Developing a cost-effective multispectral imaging system for real-time nuclear fuel pellet inspection

The production of sintered uranium dioxide (UO2) nuclear fuel pellets is a crucial process in the nuclear power industry. Given the intended use of these pellets - to provide fuel in nuclear reactors - ensuring the integrity and quality of UO2 pellets during- and post- manufacture is paramount. Currently, quality assurance (QA) methodologies tend to rely upon sampling and time-consuming offline lab analysis as well as subjective human visual inspection on the production line. In this work, we propose that QA for nuclear fuel manufacturing can be improved by Hyperspectral imaging (HSI) - a powerful technique which can be used to image a scene to capture combined spatial and spectral data simultaneously. In recent work, we have shown that HSI has been successfully used to characterise UO2 pellets and we have identified useful spectral regions in the short-wave infrared (SWIR) range [1000 nm – 2500 nm]. In particular, we have identified prominent absorption bands at around 1100 nm and 2300 nm, as well as a series of smaller spectral features between 1300 nm and 1750 nm. A key challenge for deploying HSI on the production line is associated with the cost of the HSI system itself. To address this, we have designed, built and tested a relatively low-cost multispectral imaging (MSI) sensor (10 wavelengths rather than 100s) to perform QA of UO2 fuel pellets during manufacture. In this paper, we compare the results of our novel MSI sensor against the full HSI sensor to determine its utility in practice.