Robotics and Artificial Intelligence for Nuclear (RAIN)

核工业机器人和人工智能 (RAIN)

• 批准号: EP/R026084/1
• 负责人: Barry Lennox
• 金额: $ 1631.98万
• 依托单位: University of Manchester
• 依托单位国家: 英国
• 项目类别: Research Grant
• 财政年份: 2017
• 资助国家: 英国
• 起止时间: 2017 至 无数据
• 项目状态: 已结题
• 来源: https://gtr.ukri.org/projects?ref=EP%2FR026084%2F1
• 关键词: Robotics; Artificial; Intelligence; Nuclear; RAIN

The nuclear industry has some of the most extreme environments in the world, with radiation levels and other hazards frequently restricting human access to facilities. Even when human entry is possible, the risks can be significant and very low levels of productivity. To date, robotic systems have had limited impact on the nuclear industry, but it is clear that they offer considerable opportunities for improved productivity and significantly reduced human risk. The nuclear industry has a vast array of highly complex and diverse challenges that span the entire industry: decommissioning and waste management, Plant Life Extension (PLEX), Nuclear New Build (NNB), small modular reactors (SMRs) and fusion.Whilst the challenges across the nuclear industry are varied, they share many similarities that relate to the extreme conditions that are present. Vitally these similarities also translate across into other environments, such as space, oil and gas and mining, all of which, for example, have challenges associated with radiation (high energy cosmic rays in space and the presence of naturally occurring radioactive materials (NORM) in mining and oil and gas). Major hazards associated with the nuclear industry include radiation; storage media (for example water, air, vacuum); lack of utilities (such as lighting, power or communications); restricted access; unstructured environments.These hazards mean that some challenges are currently intractable in the absence of solutions that will rely on future capabilities in Robotics and Artificial Intelligence (RAI). Reliable robotic systems are not just essential for future operations in the nuclear industry, but they also offer the potential to transform the industry globally. In decommissioning, robots will be required to characterise facilities (e.g. map dose rates, generate topographical maps and identify materials), inspect vessels and infrastructure, move, manipulate, cut, sort and segregate waste and assist operations staff. To support the life extension of existing nuclear power plants, robotic systems will be required to inspect and assess the integrity and condition of equipment and facilities and might even be used to implement urgent repairs in hard to reach areas of the plant. Similar systems will be required in NNB, fusion reactors and SMRs. Furthermore, it is essential that past mistakes in the design of nuclear facilities, which makes the deployment of robotic systems highly challenging, do not perpetuate into future builds. Even newly constructed facilities such as CERN, which now has many areas that are inaccessible to humans because of high radioactive dose rates, has been designed for human, rather than robotic intervention. Another major challenge that RAIN will grapple with is the use of digital technologies within the nuclear sector. Virtual and Augmented Reality, AI and machine learning have arrived but the nuclear sector is poorly positioned to understand and use these rapidly emerging technologies. RAIN will deliver the necessary step changes in fundamental robotics science and establish the pathways to impact that will enable the creation of a research and innovation ecosystem with the capability to lead the world in nuclear robotics. While our centre of gravity is around nuclear we have a keen focus on applications and exploitation in a much wider range of challenging environments.

核工业拥有世界上最极端的环境，辐射水平和其他危险经常限制人们进入设施。即使在人类有可能进入的情况下，风险也可能是巨大的，而且生产率水平非常低。到目前为止，机器人系统对核工业的影响有限，但很明显，它们为提高生产率和显著降低人类风险提供了相当大的机会。核工业面临着一系列高度复杂和多样化的挑战，这些挑战横跨整个行业：退役和废物管理、工厂寿命延长(PLEX)、核新建(NNB)、小型模块反应堆(SMR)和聚变。尽管整个核工业面临的挑战各不相同，但它们有许多与目前极端条件有关的相似之处。至关重要的是，这些相似之处还影响到其他环境，如空间、石油和天然气以及采矿，例如，所有这些环境都面临与辐射有关的挑战(空间中的高能宇宙射线以及采矿和石油和天然气中存在自然产生的放射性物质(NORAME))。与核工业相关的主要危险包括辐射；存储介质(例如水、空气、真空)；缺乏公用事业(如照明、电力或通信)；访问受限；非结构化环境。这些危险意味着，在缺乏依赖于未来机器人和人工智能(RAI)能力的解决方案的情况下，目前一些挑战是难以解决的。可靠的机器人系统不仅对核工业未来的运营至关重要，而且还提供了在全球范围内改变该行业的潜力。在退役过程中，将需要机器人确定设施的特征(例如绘制剂量率图、生成地形图和识别材料)、检查船只和基础设施、移动、操作、切割、分类和分离废物，并协助作业人员。为了支持现有核电站的寿命延长，将需要机器人系统来检查和评估设备和设施的完整性和状况，甚至可能被用于在核电站难以到达的地区实施紧急维修。核反应堆、聚变反应堆和SMR将需要类似的系统。此外，至关重要的是，核设施设计中过去的错误使机器人系统的部署具有极大的挑战性，不能在未来的建造中永久存在。即使是像CERN这样的新建设施，也是为人类设计的，而不是机器人干预的。CERN现在有许多区域，由于高放射性剂量率，人类无法进入。RAIN将努力应对的另一个主要挑战是在核能领域使用数字技术。虚拟现实和增强现实、人工智能和机器学习已经到来，但核行业在理解和使用这些快速新兴技术方面处于不利地位。RAIN将在基础机器人科学方面带来必要的步骤变化，并建立影响的途径，使研究和创新生态系统能够在核机器人领域引领世界。虽然我们的重心是围绕着核能，但我们非常关注在更广泛的具有挑战性的环境中的应用和开发。

项目成果

Input Shaping Predictive Functional Control for Different Types of Challenging Dynamics Processes

针对不同类型的挑战性动力学过程的输入整形预测功能控制

• DOI: 10.3390/pr6080118
• 发表时间: 2018
• 期刊: Processes
• 影响因子: 3.5
• 作者: Abdullah M

The effect of model structure on the noise and disturbance sensitivity of Predictive Functional Control

模型结构对预测函数控制噪声和扰动灵敏度的影响

• 发表时间: 2018
• 作者: Abdullah, M.

Alternative Method for Predictive Functional Control to Handle an Integrating Process

处理积分过程的预测功能控制的替代方法

• 发表时间: 2017
• 作者: Abdullah, M.

Autonomous Nuclear Waste Management

• DOI: 10.1109/mis.2018.111144814
• 发表时间: 2018-11-01
• 期刊: IEEE INTELLIGENT SYSTEMS
• 影响因子: 6.4
• 作者: Aitken, Jonathan M.;Veres, Sandor M.;Mort, Paul E.
• 通讯作者: Mort, Paul E.

A formal sensitivity analysis for Laguerre based Predictive Functional Control

基于拉盖尔的预测功能控制的形式敏感性分析

• 发表时间: 2018
• 作者: Abdullah, M.