On the Use of Microworlds for an Error Seeding Method to Support Human Error Analysis

The field of human reliability analysis (HRA) has matured significantly since its original conception in nuclear safety. However, the field of HRA must be advanced in order to catch up with the advances in the probabilistic safety analysis field to support more accurate quantification of plant risk. This paper examines the Rancor microworld as an HRA data collection tool designed to quickly administer and gather human performance data to support quantitative models of human error. An error seeding method formed the basis of the simulation design to capture human error in response to routine and abnormal operations. The rationale behind a simplified simulator, designing experiments and developing scenarios is described. Furthermore, the results of a task analysis is presented, which establishes a basis of comparison against existing data collections performed with full scope simulator studies with data collected from the Rancor microworld.