Abstract: To prevent accidents induced by leakage from offshore platforms,an integrated method based on DEMATEL-BN was worked out for performing a dynamic and accurate vulnerability assessment.Three-level indicators were established to identify factors,events and subsystems that may lead to leakage,fire,and explosion.Interactions among the factors influencing the development of accidents were analyzed using DEMATEL.A sequential model was built for describing the escalation from initial events to small-scale fire and explosion A,large-scale fire and explosion B,small-scale leakage C,and large-scale leakage D.The model was converted into a BN to dynamically calculate the posterior probabilities of accidents with the introduction of accident precursor data.The results show that the factors related to organizational management are apt to influence other factors,and the probabilities of severe accidents will increase with time,and that the proposed method overcomes the limitations of traditional methods that cannot utilize the operational data on platforms,and can be used to dynamically evaluate risks in offshore platforms.

Key words: offshore platform, vulnerability, Bayesian network (BN), dynamic risk, decision making trial and evaluation laboratory (DEMATEL)