A Dynamic Bayesian Network model for ship-ice collision risk in the Arctic waters

Extreme weather conditions of the Arctic and its icy waters pose high-risk potential for a range of marine accidents in the region. Ship-ice collision is the focus of this paper. A large number of vessels operating in the Arctic waters are at risk of ice damage due to ship-ice collisions. The damage may vary from a minor hull deformation to ruptures that could put the lives, assets, and environment at risk. To minimize the risk of ship-ice collision in Arctic waters, a simple yet robust model to make routine safety-driven operational decisions could help. The present study proposes a Dynamic Bayesian Network (DBN) model to fill this gap. The model assesses the operational risk of ship-ice collision in an ice prone region using the hypothetical form of observations. Low temperatures, Weather, Ice, Fog, Darkness, Blowing snow, Poor visibility, Ice strength, Ice drift, Types of ice, Ice concentration and Speed of the vessel are considered as the primary risk factors in the region. The estimated collision risk would provide an easy to use indicator for decisions concerning safe operations in ice such as maneuvering, route selection, and safe speed. A case study of an oil tanker navigating across the Barents Sea is presented to explain the proposed model.