Application of laboratory data from small-scale simulators to human performance issues in the nuclear industry

Laboratory analogs of nuclear power plant tasks were simulated on personal computers in two experimental studies. Human performance data were collected during each experimental study. The goal of the first experiment was to validate a quantitative model of dependence among human errors during testing, calibration, and maintenance activities. This model, the Multiple Sequential Failure (MSF) model (NUREG/CR-2211) has been used to quantify dependent human error failure probabilities for human reliability analyses in Probabilistic Risk Assessments (PRAs). The goal of the second experiment was to examine the relationship among psychological and behavioral characteristics of individuals and their performance at controlling a simulated nuclear power plant. These studies demonstrated the usefulness of the experimental psychology approach for validating models of human performance at nuclear power plant tasks.