SmartDecay – A next-generation nuclear waste segregation and sorting system

英国目前正在进行的核退役活动每年耗资30亿英镑，据估计，未来100年的退役工作仍将产生近500万吨新废物。超过90%的废物将被归类为低放射性废物，并将其与放射性含量较高的废物有效分离（中间水平废物- ILW）将有助于降低未来的废物管理成本，并为增加废物回收利用开辟新的机会。利用机器人、传感器技术和先进人工智能的创新组合，SmartDecay有可能从根本上提高不同类型的中低放射性废物的分类和隔离效率，从而显著降低核废物处理成本。（半）自主机器人系统的部署将减少人类与废物和相关处理设备的直接互动。这有助于最大限度地减少有害辐射的暴露，降低与工作有关的伤害的可能性，并降低对操作员健康的总体风险。安装在机器人上的辐射传感器将测量并记录每件废物的放射性含量，从而将其分类为ILW或LLW。然后将利用自动3D扫描仪和X射线荧光设备进一步分析废物-通过生成数字化3D模型确定其大小/形状并按材料类型对废物进行分类。经过详细分析后，废物样本将被转移到预先标记有RF-ID标签的临时隔离区，使机器人能够识别特定位置，以放置按放射性含量和材料类型分类的废物。分离后，将开发机器学习算法，并部署高性能边缘计算资源，为每种分类的废物建立最佳包装顺序，放入适当的桶和托盘箱。废物容器的最终包装将由机器人进行，该机器人将利用最先进的同步定位和地图（SLAM）在隔离区域周围进行自主导航。根据放射性含量和材料类型对退役废物进行有效分离也将有助于增加回收活动的机会。

On-going nuclear decommissioning activities in the UK currently cost £3 billion per year and it is estimated that nearly 5 million tonnes of new waste are still to be generated from future decommissioning work spanning the next 100 years. Over 90% of this waste will be categorised as low-level waste (LLW) and its efficient separation from waste with a higher radioactive content (intermediate level waste - ILW) will help to reduce future waste management costs and open up new opportunities for increased recycling of waste.Utilising an innovative combination of robotics, sensor technology and advanced AI, SmartDecay has the potential to radically improve the efficiency with which different types of intermediate and low-level radioactive waste are sorted and segregated, significantly lowering nuclear waste processing costs. The deployment of (semi-)autonomous robotic systems will result in a reduction in the number of direct human interactions with waste and associated handling equipment. This helps to minimise exposure to harmful radiation, lower the probability of work-related injuries and reduce the overall risk to operator health. Radiation sensors attached to the robot will measure and record the radioactive content of every piece of waste enabling classification as either ILW or LLW. Automated 3D scanners and X-ray fluorescence equipment will then be utilised to further analyse waste - determining its size/shape through generation of a digitised 3D model and classifying the waste by material type. After detailed analysis, waste samples will be moved into a temporary segregation area pre-marked with RF-ID tags that will enable the robot to identify specific locations to place waste classified by radioactive content and material type. After segregation, machine learning algorithms will be developed, and high-performance edge computing resources will be deployed to establish optimal packing order for each classification of waste items into appropriate drums and pallet boxes. Final packing of waste containers will be performed by a robot that will utilise state-of-the-art simultaneous localisation and mapping (SLAM) for autonomous navigation around the segregation area. Effective separation of decommissioning waste according to radioactive content and material type will also help to boost opportunities for increasing recycling activities.