Methodology for Acquiring Reliability Data in Novel Nuclear Energy Systems Using a Non-Homogeneous Poisson Process

Abstract In nuclear power plant maintenance, Reliability-Centered Maintenance (RCM) increasingly focuses on preventing equipment failure. This is particularly important in novel reactor models where mature data for maintenance decisions is limited. Bayesian Networks (BNs) are employed to fill this gap, modeling variables and their dependencies to assess component impact on system failure rates. Innovative approaches like Bayesian updating with non-homogeneous Poisson processes (NHPP) for small samples can modeling failures more accurately. This article utilizes the Markov Chain Monte Carlo (MCMC) method for Bayesian updating of component reliability data and employs BNs to analyze the reliability parameters of first of a kind equipment. These methodologies enhance the safety and efficiency of nuclear power plants by ensuring maintenance decisions are informed by existing and comprehensive reliability data.