LEVERAGING EXISTING TOOLS FOR SIMULATING OPERATOR PERFORMANCE IN DISCRETE DYNAMIC EVENT TREES

摘要

The conventional probabilistic risk assessment (PRA) methodology uses a static event-tree approach, hence it can only account for the time element in system evolution through predetermined sequencing of events. As an advanced reliability and risk assessment technique, discrete dynamic event trees (DDETs) are developed to determine event sequencing and quantify branching probabilities dynamically. Compared to the conventional static methodology, the dynamic methodology has many potential advantages, including reducing modeling simplifications, subjective judgment in event sequencing, and uncertainty in the timing of key events. In addition, the integration of an operator response model in the DDET generation process can improve treatment of human interaction and mitigation as it can provide extensive contextual information for the specific decision-making that will take place during accident management. This context can be particularly helpful in post-core damage accident management where the severe accident management guidelines (SAMGs) guide plant personnel response in a less prescriptive manner. Although a variety of tools and techniques (e.g., Analysis of Dynamic Accident Progression Trees (ADAPT) and the Information, Decision, and Actions in a Crew context (IDAC) cognitive model) have been proposed by different organizations to address the dynamic response of both plant systems and/or operators during an accident, they have limitations in simulating crew cognitive activities and actions. With the goal of leveraging existing tools to the extent possible to evaluate the dynamics of operator actions on postulated severe accident scenarios, this article first discusses the challenges in integrating existing tools for generating DDETs, and then proposes a framework to fully implement a cognitive model capability with ADAPT, the Accident Dynamics Simulator (ADS)-IDAC, and MELCOR, which is the NRC's severe accident simulation tool.