Probabilistic Approach for Assessing the Weather and External Forces Hazard Using Bayesian Belief Networks

Abstract

This paper presents a probabilistic approach for the assessment of the Weather and External Forces hazard in pipelines using a Bayesian Belief Network. Such model was constructed through structured expert judgment methods for those factors whose deterministic modelling is complex or unknown and existent models for those known factors or previously studied phenomena. For the hazard modelling, three fundamental groups of variables were considered: susceptibility, triggering events and vulnerability. Through this proposal, the hazard is described from a probabilistic perspective, including the uncertainty associated with instability processes, information gathering and analysis and risk perception, among others. A transportation system characterized by its high geotechnical complexity is used as a case study, and results were compared with the current assessment model. The results show that the new model has greater sensitivity to the hazard level while being consistent with both the condition of the Right of Way (ROW) and the hydrocarbon transportation infrastructure.