Bayesian modelling of navigational safety for tanker ships. An applicative case study in Venice channels

The complexity and potential environmental impact of tanker ship operations require effective risk assessment methods to ensure the safe navigation, particularly in sensitive and constrained waterways. In this study, a framework for tanker-specific navigational risk assessment, combining hazard identification, Bow-Tie analysis, and quantitative evaluation through fault and event trees, is proposed. Historical accident data are integrated with expert elicitation to derive baseline probabilities of hazardous events, providing a robust foundation for probabilistic risk modelling. A Dynamic Bayesian Network (DBN) extending beyond static approaches by incorporating its evolution over the ship journey, is designed and tailored across successive transit phases—channel entry, mid-channel, port approach, and docking—allowing for the phase-dependent updating of posterior probabilities. This facet enables the definition of Dynamic Safety Indicators (DSI), which capture the time-dependent evolution of navigational risks and support informed decision-making for safe transit scheduling and operational planning. A case study is developed to demonstrate how the DBN framework improves situational awareness and proactive safety management. The proposed methodology thus provides both a scientific contribution to maritime risk assessment and a practical decision-support tool for enhancing the safety and sustainability of tanker operations.