Modelling hazardous event scenarios for decision support

Abstract

Quantitative risk analysis has been a successful tool to support decision-making related to the design of technical (safety) barriers. A quantitative risk model may comprise a combination of event trees and fault trees, which is a basis for causal and frequency analysis of accident scenarios. Fault tree analysis is a well-documented technique, and it is commonly used for casual analysis of hazardous events. Fault tree analysis, however, is not fully suitable for modelling dynamic systems where the status of barriers may change, depending on various operational decisions. This paper introduces Petri nets as a formalism to consider this aspect. Petri nets are employed to model the sequence to a hazardous event, considering the effect of testing, repair, and daily activities on the availability of safety barriers. A representative case of an atmospheric tank overfilling is considered. Moreover, the effect of decisions on the risk level is addressed from a perspective of decision-making support.