Abstract
The global nuclear industry has a growing volume of nuclear waste which needs to be scanned, sorted
according to its activity and material type, then processed into the correct waste packages for long term
storage and disposal. It is vital that there is a detailed understanding of the waste inventory stored in
long term waste containers, as knowledge of their contents could predict or prevent any adverse effects
in storage. The numerous “scan and sort” tables which are currently used at many different facilities
around the world to sort waste into their correct containers are human operated and require very slow
gamma scanning procedures combined with educated guesswork to manually sort the waste. This often
leads to excessive conservatisms, with placement of lower activity wastes in higher activity containers,
which in turn costs significantly more to store. In the United Kingdom it costs £46k per cubic meter to
store intermediate level waste compared to just £2.9k per cubic meter to store low level waste according
to a 2008 Department of Energy and Climate Change report in the UK. A proposed solution to this
problem, is the use of a robotic manipulator to automatically inspect the “scan and sort table” in order
to produce an accurate 3D model of the table’s waste contents and attach an overlaid radiation map.
The radiation map contains spectrometry data and can in consequence be used to distinguish and locate
specific radioisotopes. The 3D model should be as accurate as possible in order to allow for a second
robot arm with an attached gripper to grasp the objects and place them into their designated long-term
storage container. Various scanning procedures are explored in this study including basic raster
scanning, adaptive raster scanning and point sampling. The optimal solution will in practice be defined
by the required application and activity level of the wastes being inspected.
KUKA iiwa Robot Arm
Micro-gamma Spectrometer
The results presented in this study indicate that it is possible to produce a centimeter accurate 3D model
of a mixed assortment of components on a nuclear waste “scan and sort” table. In addition, it was shown
that the waste objects emitting radiation could be accurately identified and located, with an overlaid
radiation map
This study is applicable across the nuclear waste management sector. Many of the ideas and concepts
developed in this study are applicable in other decommissioning settings for example, dismantling of
legacy gloveboxes or routine inspection of nuclear waste packages in storage.