Total simulation of operator team behavior in emergencies at nuclear power plants

In a large and complex system (i.e., a space aeronautics and nuclear power plant) it would be valuable to conduct operator training and support to demonstrate standard operators' behavior in coping with an anomaly caused by multiple malfunctions in which procedures would not have been stipulated previously. A system simulating operator team behavior including individual operator's cognitive behavior, his operations and physical behavior, and even verbal communication among team members, has been developed for a typical commercial nuclear power plant. This simulation model is not a scenario-based system but a complete knowledge-based system, based on the mental model that was envisaged by detailed analyses of experimental results obtained in the full-scope plant simulator. This mental model is composed of a set of knowledge bases and rules able to generate both diagnosis and prognosis depending on the observed situation even for multiple malfunctions. Simulation results of operator team behavior and plant dynamics were compared with corresponding experiments in several anomalies of multiple malfunctions. The comparison showed a reasonable agreement, so the simulation conditions were varied on cognitive task processing speed of individual operators, on team role sharing scheme, and on human machine interface (1st generation to 2nd generation control panel) to assess the sensitivity of this simulation model. Finally, it was shown that this simulation model has applications for the use of training standards and computer aided operator support systems.