Ship machinery condition monitoring using performance data through supervised learning

Research output: Chapter in Book/Report/Conference proceedingChapter

100 Downloads (Pure)

Abstract

This paper aims to present a methodology for intelligent monitoring of marine machinery using performance data. Monitoring of machinery condition is a crucial aspect of maintenance optimisation that is required for the vessel operation to remain sustainable and profitable. The proposed methodology will train models pertinent to specific machinery components using pre-classified performance data and then classify new data points using the models developed. For this, measurements are suitably analysed and processed to retain most of the information (variance) of the original dataset while minimising number of required dimensions. Finally, new data are compared against the models developed to evaluate their condition. The above will provide a flexible but robust framework for the early detection of emerging machinery faults. This will lead to minimisation of ship downtime and increase of the ship’s operability and income through operational enhancement. Case studies that show initial results obtained through main engine data are included.
Original languageEnglish
Title of host publicationProceedings of the 2017 Smart Ship Technology Conference
Place of PublicationLondon
PublisherRoyal Institution of Naval Architects
Pages105-111
Number of pages7
ISBN (Print)9781909024632
Publication statusPublished - 24 Jan 2017
EventSmart Ship Technology 2017 - RINA HQ, 8-9 Northumberland Street, London, WC2N 5DA, London, United Kingdom
Duration: 24 Jan 201725 Jan 2017
https://www.rina.org.uk/Smart_Ships2017.html

Conference

ConferenceSmart Ship Technology 2017
CountryUnited Kingdom
CityLondon
Period24/01/1725/01/17
Internet address

    Fingerprint

Keywords

  • performance measurements
  • predictive maintenance
  • machine learning
  • condition monitoring
  • SVM

Cite this

Gkerekos, C., Lazakis, I., & Theotokatos, G. (2017). Ship machinery condition monitoring using performance data through supervised learning. In Proceedings of the 2017 Smart Ship Technology Conference (pp. 105-111). London: Royal Institution of Naval Architects.