Predictive Maintenance of Critical Systems & Components in Canada's Nuclear Energy Infrastructure Under Seismic Risk

核电厂安全运行面临的两个主要挑战是：1）超出设计基础的地震危险; 2）核电厂内关键、安全相关系统和部件的相互依赖性导致的复杂性。 对于北美东部的岩石场地，加拿大标准协会和美国核管理委员会推荐的设计反应谱对于17 Hz以上的频率是不安全的，对于较低的频率则过于保守。这种情况对加拿大构成了重大威胁，因为其核基础设施集中在东部省份安大略、魁北克和新玩法，而目前的设计建议是基于西部地震记录;特别是考虑到加拿大在不久的将来很有可能经历地震事件。核电厂安全运行的第二个挑战涉及安全相关系统和组件的功能，这些系统和组件形成了一个由相互依赖的组件组成的错综复杂的系统，其运行方式通常仅在组件级别上才能理解，而其系统级别的行为通常对运营商是隐藏的。工厂管理人员目前无法提供（系统级）系统故障的可靠预测，其中少数组件的相对较小的损坏可能导致一系列后果，最终导致完全（通常是不安全的）NPP关闭。在这种情况下，电动阀（MOV）是一个主要问题，因为它们对核电厂的安全运行至关重要。更好地理解MOV相互依赖的网络是预测和预防未来灾难性事件的关键。为了减轻服务中断的风险，本研究将集中在智能工具的开发，旨在帮助NPP运营商在预测性维护的挑战性任务。将开发一个三层模型来预测在受到与加拿大核电站场地相对应的地面运动激励时MOV和MOV网络的行为。根据模型输出，工厂管理人员将能够优先考虑未来的维护和功能优化干预，以最大限度地减少停电的持续时间，频率和总成本。在重点领域I中，研究将限于原始条件下的MOV网络;在重点领域II中，将扩展模型以纳入MOV共因故障和地面震动的并发和复合效应，并通过振动台测试验证结果;在重点领域III中，将开发基于风险的监测工具，以促进拟议的MOV网络预测性维护工具。这项研究计划将通过降低与服务中断、维护和基础设施更新相关的成本，加强核工业的安全性并提高生产率。核工业每年创造60亿美元的收入，提供30，000个直接就业岗位，同时通过低温室气体排放为保护环境做出贡献。

Two of the main challenges facing safe nuclear power plants (NPPs) operations are: 1) Beyond design-basis seismic hazards; and 2) Interdependence-induced complexity of critical, safety related, systems and components within NPPs. For rock sites in eastern North America, design response spectra recommended by the Canadian Standard Association and the US Nuclear Regulatory Commission are unsafe for frequencies above 17 Hz and overly conservative for lower frequencies. This situation poses a significant threat for Canada, because its nuclear infrastructure is clustered in the eastern provinces of Ontario, Québec, and New Brunswick, while current design recommendations are based on western seismological records; especially considering the high likelihood that Canada will experience a seismic event(s) in the near future. The second NPPs safe operation challenge pertains to the functionality of safety related systems and components, which form an intricate complex system of interdependent components that operates in a manner often understood only at a component-level, while its system-level behaviour is often concealed to the operators. Plant managers are currently unable to provide robust predictions of (system-level) systemic failures, wherein relatively minor damage of a handful of components may lead to a cascade of consequences culminating in complete (and often unsafe) NPP shutdown. In this context, motor operated valves (MOVs) are a primary concern, given how central to the safe operation of NPPs they are. A greater understanding of MOV interdependent networks is key to predict and prevent future catastrophic events. In order to mitigate the risk of service interruption, this research will focus on the development of intelligent tools designed to aid NPP operators in the challenging task of predictive maintenance. A tri-level model will be developed to predict the behaviour of MOVs and MOV network when excited by ground motion corresponding to Canadian NPP sites. Based on the model output, plant managers will be able to prioritize future intervention of maintenance and functional optimization, to minimize the duration, frequency and overall cost of power outages. In Focus Area I, the study will be limited to MOV networks in original condition; in Focus Area II, the model will be extended to incorporate the concurrent and compounding effects of MOV common cause failures and ground shaking, and the results will be validated via shake table testing; in Focus Area III, risk-based monitoring tools will be developed to facilitate the proposed MOV network predictive maintenance tool. This research program will bolster safety and augment productivity in the nuclear industry-a sector that generates $6 billion in revenues every year and provides 30,000 direct jobs, while contributing to the tutelage of the environment via low GHG emissions-by reducing the costs associated with the interruption of service, maintenance, and infrastructure renewal.