Human factors in annunciator/alarm systems: annunciator experiment plan I

This report presents a plan for the experimental evaluation of alarm reduction techniques on a full scale nuclear power plant simulator. The alarm reduction techniques include suppression of irrelevant alarms based on plant mode, time delay of alarm groups, and separation of component status annunciators from alarms. Five hypotheses about the benefits of alarm reduction are stated. The independent variables whose effects are to be evaluated in the experiment are identified and their possible interactions are described. The performance measures are defined in terms of operator decision levels, operator task performance, operator perceived work load, and operator control of key process variables. The data obtained from measurements of the performance measures in response to changes in the independent variables will be analyzed by statistical analysis of variance in order to assess the extent to which the data support or contradict each of the hypotheses