An identification and grouping approach to analyze the output of a dynamic safety assessment

Abstract

Dynamic Probabilistic Safety Assessment embeds models of the system process (typically thermo-hydraulic models) and of human operator dynamics within stochastic simulation engines. These engines generate sequences of component and operator action events, representing success, failures and other sources of variability. A challenge is to retrieve and organize the scenarios hidden in the large amount of information produced. The present paper discusses an approach for identifying and grouping the produced scenarios, based on possibilistic clustering classification; the aim is to single out the principal patterns of system evolution with respect not only to the final system states but also to the time of events and to the process evolution.