Aggregation of autocalculated human error probabilities from tasks to human failure events in a dynamic human reliability analysis implementation

Dynamic human reliability analysis (HRA) typically operates at the task or subtask level, allowing a tight coupling between plant evolutions and operator response. However, most HRA methods calculate human error probabilities (HEPs) at the human failure event (HFE) level. This results in a mismatch between the types of HEPs generated for static vs. dynamic HRA. There has been no clear guidance on how to aggregate the higher sampling frequency of dynamic HEPs to match the HEPs in static HRA. Applying available task dependence correction factors, for example, artificially inflates the HEPs, resulting in unrealistically conservative HEPs generated from dynamic HRA. In this paper, we review aggregation techniques that calibrates dynamic task-level HEPs to static HFE-level HEPs. This aggregation allows more direct validation of dynamic HRA results to reference static HRA results.