Bayesian network and game theory risk assessment model for third-party damage to oil and gas pipelines

Abstract

Tremendous amounts of oil and gas products are transported in pipelines worldwide resulting in increasing interest to identify the hazards and evaluate the associated risks associated with this critical infrastructure. Third-party intrusion is one of the least quantifiable factors being considered during the pipeline hazard assessment stage despite the substantial contributing to the total number of oil and gas pipeline incidents. This is because a probabilistic risk assessment cannot reliably model human actions and be applied to intentional acts. Due to the distinctive motivations of third-party damage, an unintentional third-party damage Bayesian Network model and a game-theoretic model on malicious intrusion will therefore be built, to examine the mechanism of pipeline failure caused by this mode. This study is conducted aiming at investigating pipeline risk resulting from third-party damage, and will formulate risk assessment models to identify threats, prioritize risks and determine which integrity plan should apply to different pipeline segments given the condition of third-party interference (both the accidental damage and malicious acts).