Probabilistic Petri-net addition enabling decision making depending on situational change

Abstract

Emergency response was usually not taken into account in the analysis of safety barriers to prevent fire-induced domino effects. When emergency response is considered, in addition to failure events, successful events may also have an impact on fire escalation, and there may be a dependency between pre- and post-events. To model such events and carry out probability analysis, a probabilistic Petri-net (PPN) approach is proposed. The definition of a PPN with the complementary arc and its execution mechanism are provided, and a probability reasoning algorithm reflecting the parallel processing of PPNs is developed. On this basis, a PPN model discussed in two parts for failure analysis of fire escalation prevention is established. The application of the methodology is demonstrated by fire prevention in a chemical storage tank farm, focusing on the influence of fire detection on follow-up actions. In addition to that the fire is not been found, the time at which the fire is found will also have an impact on the success of preventing the fire escalation. The influences of not considering/considering fire detection time are analyzed by the fault tree method and the PPN method, respectively, and there is a difference of about 11% between the results.