Abstract

Due to the intricate and interrelated nature of nuclear power plant systems, any anomaly can trigger a casc ade of alarms. Operators must skillfully assess the system's status amidst this deluge of alarm information and promptly implement appropriate mitigation measures. To address these challenges, an intelligent alarm analysis method based on a Multi-Level Flow Model (MFM) is employed, facilitating the analysis of causal relationships between alarms. Utilizing the secondary loop system and the Low-Pressure Feedwater Heater System (ABP) of a Pressurized Water Reactor (PWR) nuclear power plant as research subjects, a multi-layer flow model for the ABP system was constructed. This model, based on the steady-state operation at full power of the nuclear power plant, clarifies the causal relationships between functions and targets. The mul tilayer flow model was validated using an alarm scenario triggered by an insertion fault superimposed under steady-state operating conditions. The results indicate that the system can filter information, identify root causes, and provide operator s with effective comprehensive diagnostic information. This reduces the impact of a large volume of information on diagnostic tasks and enhances the safety and reliability of nuclear power plant operations.