Intelligent Alarm Analysis for Fault Diagnosis and Management in Nuclear Power Plants

The safety and efficiency of nuclear power plants (NPPs) are heavily reliant on the timely and accurate decisions made by operators in the control room, particularly during fault scenarios. Traditional alarm systems often fall short in identifying the root causes of faults, especially in complex, multi-fault situations. This paper presents an advanced fault diagnosis method based on the multilevel flow model (MFM), which addresses the limitations of conventional alarm systems by linking alarms to system goals and functions. The MFM framework organizes the system hierarchically, aiding in the clear understanding of causal relationships and the underlying causes of faults. Additionally, this paper proposes a method that combines the longest causal path and root cause analysis, addressing both single- and multiple-fault diagnosis needs. This approach improves diagnostic efficiency and comprehensiveness, providing operators with a systematic basis for analysis and decision-making in multi-alarm scenarios.