Evaluating the Safety of Port-to-Ship Liquefied Hydrogen Bunkering Through Fault Tree Analysis and Fuzzy Bayesian Networks

Abstract

As marine fuels transition toward zero carbon dioxide emissions, hydrogen emerges as a promising alternative due to its near-zero emissions and high energy density. However, hydrogen’s high flammability presents a significant risk of fire-related accidents onboard ships. Among the various operations involving hydrogen in the maritime sector, bunkering is one of the most critical. This process carries risks such as leakage, pressure failures, and human errors, which can lead to severe consequences. This study analyzes the safety of the liquefied hydrogen bunkering process. Fault Tree Analysis (FTA) was used to identify failures within the bunkering system, employing a top-down approach that traces failures from the top event to the basic events. Fuzzy Set Theory was applied to quantify risk by aggregating linguistic variables from five experts to estimate the probabilities of basic events. Bayesian Networks (BN) were then used to establish causal relationships between factors and identify the minimum cut sets leading to the top event of bunkering failure. Rate of Variation (RoV) analysis of the posterior probabilities of basic events revealed that human error, particularly due to task execution problems and human-machine interface issues, is the most significant factor influencing bunkering failures. The findings highlight critical areas that require attention to enhance the safety of liquefied hydrogen bunkering operations in the maritime sector, especially for relevant stakeholders.