A data-driven Ship Risk Profile model for Turkish Straits (TS-SRP) using Machine Learning

Cengiz Vefa Ekici*, Ülkü Öztürk, Yunus Emre Şenol

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

Marine accidents exhibit a higher frequency within ports, internal waters, and narrow waterways as compared to open seas, emphasizing the paramount concern of maritime authorities regarding the safe navigation of ships in congested waters. In response to this issue, coastal states, in addition to flag state and port state measures, institute preventive measures by conducting risk analyses in narrow waterways with geographical restrictions and high traffic density. To meet this need, this study aims to develop a Ship Risk Profile (SRP) model to monitor ships by evaluating the potential risks of the ships passing through narrow waterways and taking the necessary precautions proactively by Vessel Traffic Service (VTS) for mitigating the potential accident and environmental pollution risk. The Turkish Straits is determined as the application area for this paper. The proposed model employs Machine Learning (ML) methods and analyzes targeting factors from Port State Control (PSC), ship particulars obtained from Sailing Plan (SP)-1 reports specific to the Turkish Straits, as well as relevant environmental factors. The results demonstrate that the developed model holds promise for improving safety in Turkish Straits and can be adapted by VTSs in other sea regions by customizing it to suit their specific characteristics and conditions.

Original languageEnglish
Article number119002
JournalOcean Engineering
Volume311
DOIs
Publication statusPublished - 1 Nov 2024

Bibliographical note

Publisher Copyright:
© 2024 Elsevier Ltd

Keywords

  • Data analysis
  • Machine Learning
  • Maritime safety
  • Risk assessment
  • Ship Risk Profile
  • Turkish Straits

Fingerprint

Dive into the research topics of 'A data-driven Ship Risk Profile model for Turkish Straits (TS-SRP) using Machine Learning'. Together they form a unique fingerprint.

Cite this