Predictive Modeling for Health Condition and RUL Estimation in Critical Equipment of Nuclear Power Plants

Spanish nuclear power plants are surpassing 40 years of operation, requiring the highest standards of safety and reliability to be maintained across all systems. In this context, the development of tools capable of accurately assessing equipment health and predicting Remaining Useful Life (RUL) is highly valuable. By leveraging data from online condition monitoring, equipment degradation can be detected early and failures prevented, ensuring the safe and efficient operation of nuclear facilities. This work introduces a methodology for RUL prediction, structured in three main steps. First, the variables most closely associated with equipment degradation are identified and processed to extract meaningful information, resulting in a degradation indicator. Second, various probability distributions are fitted to construct a reliability model, enabling the estimation of failure probability over time. Third, the evolution of the degradation indicator across operating cycles is modelled. This, combined with the reliability model, allows for future failure probability projection and RUL estimation. Results from a case study demonstrate the model's predictive capability, providing a robust framework for proactive maintenance management of nuclear assets, contributing to the continued safety and sustainability of Spain's nuclear fleet.

Keywords: Condition monitoring, Remaining Useful Life (RUL), nuclear safety, proactive maintenance.