Abstract: The Arctic sea ice environment and climate change in Arctic waters cause challenges to the safety and security of ship navigation. In order to reduce the risk of ships becoming stuck in ice and protect the safety of crews and ships, it is important to investigate relevant risk early warning systems for ships sailing in Arctic waters. This paper identifies the risk factors for ships being stuck in ice when navigating in Arctic waters from the perspectives of weather, ocean and sea ice conditions, ship, and sailors. A multi-state fault tree risk warning model was constructed based on navigational warning classifications, and the probability prediction was carried out to determine the risk of ships becoming stuck in ice under the combined scenarios of sea ice conditions and ship operation state. Sixteen nodes (including twelve basic events, three intermediate events, and one top event) were identified in the risk warning model for ships sailing in the Arctic, and the nodes were assigned to four state levels: alarm, warning, reminder, and safe. The model provides a dynamic risk warning system based on in situ data from ships sailing in Arctic waters.

Key words: ships stuck in ice,  multi-state fault tree,  risk early warning,  Arctic shipping