An integrated risk assessment model for safe Arctic navigation

Chi Zhang, Di Zhang*, Mingyang Zhang, Xiao Lang, Wengang Mao

*Corresponding author for this work

Research output: Contribution to journalArticleScientificpeer-review

41 Citations (Scopus)


Safety is always the first concern for a ship's navigation in the Arctic. Ships navigating in the Arctic may face two main accident scenarios, i.e., getting stuck in the ice and ship-ice collision. More specifically, excessive speed may cause severe hull damage, while a very low speed may lead to a high probability of getting stuck in the ice. Based on this multi-risk perspective, an integrated risk assessment model was proposed to obtain the overall risk using the Bayesian Network (BN), in which the probabilities of accident occurrence and the severities of the possible consequences for ships getting stuck in the ice and for ship-ice collision could be estimated. Then, the voyage data collected from Yong Sheng's Arctic sailing in 2013 were inputted into the integrated risk assessment model to perform a case study. A sensitivity analysis was performed to validate the proposed model and reveal the inherent mechanisms behind these two accidental scenarios. The proposed model can be applied to identify the safe speed for Arctic navigation under various ice conditions, a duty that is traditionally performed by well-trained crew members, but which entails too many uncertainties. The results can, to some extent, provide useful suggestions for navigators. They are imperative in supporting decision-making to shape the Arctic policy and to enhance the safety of Arctic shipping.

Original languageEnglish
Pages (from-to)101-114
Number of pages14
Early online date7 Nov 2020
Publication statusPublished - Dec 2020
MoE publication typeA1 Journal article-refereed


  • Bayesian Network
  • Risk assessment
  • Safe speed
  • Ship-ice collision
  • Stuck in the ice


Dive into the research topics of 'An integrated risk assessment model for safe Arctic navigation'. Together they form a unique fingerprint.

Cite this