SPN as a tool for risk modeling of fires in process industries

Abstract

The modeling of risk to safety of personnel in process industries is very often carried out by the application of event trees. Event tree method is an inductive analysis that starts with a specified initiating event and ends up with the possible consequences of this event. Risk to safety of personnel is then defined as a product of event frequency and its consequences. However, event tree is a steady-state method. Given dynamic nature of industry processes, substitution of event trees for better suited modeling tool is necessary for accurate risk estimation. One such tool was found in a form of modeling language Petri Nets and its extensions, all capable of modeling dynamic processes. This article presents modeling the risk to safety of personnel on an offshore hydrocarbon production facility following hydrocarbon leak. As survival of personnel during hydrocarbon leak depends on numerous time-dependent events like fire escalation and gas cloud explosion, an extension of Petri Nets, Stochastic Petri Nets, was chosen as a suitable modeling tool. Event descriptions and model construction were based on realistic data from an offshore industry. Resulting probabilities of fatality following hydrocarbon leak were computed using Petri Nets module of GRIF software. Obtained probabilities were then compared with event tree and Direct Monte Carlo method results.