Research on the application of Bayesian Network in LNG marine transportation risk assessment
To scientifically and accurately assess the risks associated with LNG maritime transportation and identify key influencing factors,we propose a two-stage model for LNG maritime transportation risk assessment.In the initial stage,we identified 22 risk factors influencing LNG maritime transportation through a comprehensive review of relevant literature on LNG shipping risks and maritime transportation risk assessment.Subsequently,these risk factors are treated as basic events and utilized to construct a fault tree,incorporating their structure and interrelationships with 9 intermediate events.Expert scoring method is then applied to obtain the prior probability of basic events,with the introduction of fuzzy set theory to mitigate the effects of fuzziness and uncertainty in expert scoring.In the second stage,the fault tree can be converted into a Bayesian Network by establishing connections between the events and logical relations in the fault tree,and the nodes and conditional probability tables in the Bayesian Network.Building on this foundation,the model enables the prediction of probabilities associated with LNG maritime transportation risks.Following model validation,Bayesian Network are employed to calculate and prioritize the probability and critical importance of events.This aids in identifying events that are more likely to cause risk and occur more frequently,thus facilitating effective risk prevention strategies.Utilizing the bidirectional diagnostic reasoning function of Bayesian Network,we calculate the posterior probability of each event.This enables accurate identification of basic events with significant impact on the risk of LNG maritime transportation,facilitating effective risk diagnosis.Empirical analysis of LNG transportation on the China-Australia route reveals a 7.84％probability of risk occurrence for LNG maritime transportation on this specific route.Additionally,factors such as global epidemics,piracy,terrorist attacks,waves,ship traffic flows,and risks associated with LNG loading and unloading are significant contributors to the occurrence of risks in LNG maritime transportation on this route.These research findings serve as a valuable reference for shipping companies to conduct scientific risk assessments of LNG maritime transportation.
safety engineeringrisk assessmentmaritime transportfault tree analysisBayesian networkliquid natural gas
万方数据
维普
