A data mining method for automatic identification and analysis of icebreaker assistance operation in ice-covered waters. (15th December 2022)

• Record Type: Journal Article
• Authors: Liu, Cong; Musharraf, Mashrura; Li, Fang; Kujala, Pentti
• Is Part Of: Ocean engineering. Volume 266(2023) Part 2
• Publication Date: 2022-12-15
• Subjects: Maritime safety -- Data mining -- Escort operations -- Convoy operations -- Ship performance in ice -- Ice-covered waters; Ocean engineering -- Periodicals; Ocean engineering; Periodicals; 620.4162
• DOI: 10.1016/j.oceaneng.2022.112914
• Languages: English
• ISSNs: 0029-8018

Abstract

Abstract: Icebreaker assistance is a common but complex operation in ice-infested regions. Currently, the operational decision-making and the decisions regarding the safety indicators are primarily based on expert knowledge, resulting in subjectivity and the ad hoc nature of icebreaker assistance. This can negatively impact both the navigational efficiency of icebreaker services and the productivity of port services. This paper proposes a data mining method to automatically identify icebreaker assistance cases from big data. The identified cases are then used to statistically analyze the safety indicators. The data used in the paper include navigational data obtained from the Automatic Identification System (AIS) and sea ice data in the Baltic Sea area. A multi-step clustering method is adopted to cluster similar trajectories of merchant vessels and icebreakers, identifying assistance events automatically. The results show that the proposed method can automatically identify icebreaker assistance cases with an accuracy of 99.6%, precision of 87%, and recall of 78.3%. The automatic identification along with the statistical analysis can assist in the development of an intelligent decision-making system for safe and efficient winter navigation. Highlights: A database is developed merging traffic flow information with ice conditions. A data mining method is proposed to automatically identify icebreaker assistances. Safety indicators are statistically analyzed using the cases identified from big data. The results can support the intelligent decision-making of winter navigation.