Network based approach for predictive accident modelling

Abstract

Accident modeling methodologies in literature such as the System Hazard Identification, Prediction and Prevention (SHIPP) consider accident precursors and the associated five engineering safety barriers to assess the likelihood of accident occurrence with the help of fault and event trees to model the cause-consequence relationship between the failure of safety barriers and potential adverse events and design preventive, controlling and mitigating measures for improving the industrial process safety. In the SHIPP method, a restrictive assumption is used that the severity of the adverse events progresses only through sequential failures of the five safety barriers considered.In the proposed methodology, shortcomings of the existing accident model are improved in the following ways. Firstly, the above mentioned restrictive sequential progression assumption is mitigated in the SHIPP methodology by allowing non-sequential failure of safety barriers to cause adverse events in any order. Secondly, in the prediction of posterior probabilities of adverse events for real time industrial data, an important mechanical safety barrier ‘Damage Control Emergency Management Barrier’ has been included. Further, posterior probabilities of the occurrence of the adverse events are calculated using a Bayesian network (BN) approach. The utility of this approach is tested and demonstrated with the data from a liquefied natural gas (LNG) process facility. The method allows for continual updating of occurrence probabilities for adverse events and failure probabilities of safety barriers for successive real time data from industry.