Development of a Dynamic Cognitive Modeling Architecture of Human Reliability Simulation using the Rancor Microworld Simulator

Human error is regarded as the largest contributor to plant safety. Even with careful selection and training human performance can vary between operators and even within operators. Despite decades of advancement Human reliability analysis (HRA) primarily relies on experts to perform subjective analyses of tasks and subtasks to estimate performance shaping factors (PSFs). These PSFs are then used to estimate human error probabilities (HEP). The subjective analysis is also prone to error, bias, and requires the experts to accurately understand the task complexity and time constraints. During operations the conditions can be highly dynamic and the conditions may not match those envisioned during the HRA. Dynamic HRA aims at estimating HEPs on-the-fly based on dynamically estimated PSFs. Here we describe the development of a cognitive modeling architecture for dynamic HRA and the use of Rancor for model development and validation.