Modeling crew behavior and diagnoses in the control room

This study investigated the feasibility of using task network simulation to predict task performance and operator diagnoses in complex situations. Crews participated in four process-disturbance scenarios, each of which was modeled using Micro Saint. Three scenarios investigated operator task performance; the other scenario investigated operator diagnoses. Comparisons of model and empirical results revealed good predictive ability of the models. For task performance, models accurately predicted latencies between a critical event and appropriate response initiation. The model of cognitive performance predicted many of the diagnoses the operators held during the scenario, as well as the diagnoses shifts over time. These results indicate potential for using task network modeling to supplement empirical research and contribute to control room human-machine interface design