Risk Models Utilizing Historical Plant Data for Optimizing Maintenance Strategy at Nuclear Power Plants

The current maintenance practices at commercial nuclear power plants (NPPs) adhere to a fixed schedule designed to preclude any form of degradation or failure of structures, systems, and components (SSCs). This preventive maintenance paradigm, although effective in enhancing reliability, could result in considerable operational inefficiencies. This paper presents a methodology for addressing the current scheduled maintenance practices through a risk-informed approach that provides component health and condition assessment to inform condition-based maintenance practices. In this work component risk parameters are estimated and updated utilizing historical maintenance data at NPPs, and Markov chain–based models are developed to represent healthy, degraded, and failed states of SSCs. The transition rates of Markov models are estimated using Bayesian estimation that utilizes historical maintenance data of components at a commercial NPP. By integrating historical maintenance data of components, this approach offers a nuanced understanding of SSC condition, beyond the simplistic preventive maintenance schedules. The parameter estimation in this work facilitates the transition from the prevailing fixed-schedule preventive maintenance paradigms to a more efficient, component-specific, condition-based maintenance framework. Such an advanced maintenance paradigm would not only fill a significant gap in current practices but would also align with the industry’s objective to maintain high safety standards while improving operational cost-effectiveness and sustainability.