An operational risk analysis tool to analyze marine transportation in Arctic waters

Highlights: • An Object-Oriented Bayesian Network model risk model for marine operation. • The risk model is built upon multivariate joint probability conditions. • The risk model considers evolving environmental conditions, navigational errors and also human factors. • Model applicability is tested on a sea navigation through Northern Sea Route. - Abstract: The Arctic Ocean has drawn major attention in recent years due to its rich natural resources and shorter navigational routes. Arctic development and transportation involve significant risk caused by the unique features of this region, such as ice, severe operating conditions, unpredictable climatic changes, and remoteness. Considering the high degree of uncertainty in the performance of vessel operating systems and humans, robust risk analysis and management tools are required to provide decision-support to prevent accidents and ensure safety at sea. This paper proposes an Object-Oriented Bayesian Network model to dynamically predict ship-ice collision probability based on navigational and operational system states, weather and ice conditions, and human error. The model, when integrated with potential consequences, may help estimate risk. A case study related to oil tanker navigation on the Northern Sea Route (NSR) is used to show the application of the proposed model to predict oil tanker collision with sea ice.