ABSTRACT
ADS-IDAC is a discrete dynamic PRA simulation platform in which the time-dependent changes in the functional state and parameters associated with the system elements are traced to generate scenarios by branching to new sequences at various time steps following a small set of general branching rules. These model-based branching rules have been developed to obtain a more realistic and complete solution space than the traditional static PRA methods, and avoid the sequence explosion phenomenon as the number of system states increases. This paper describes a new version of the ADS-IDAC simulation platform that includes: branching based on important human operator events – e.g., information processing, decision-making, procedure-following, or action-taking type, and full implementation of Human Error Probability (HEP) quantification rules that explicitly account for HEP dependencies based on shared performance shaping factors modeled using a dynamic Bayesian network.
