An intelligent fault diagnosis system on ship machinery systems based on support vector machine principles

Ozturk U., Çiçek K., Çelik M.

26th European Safety and Reliability Conference, ESREL 2016, Glasgow, England, 25 - 29 September 2016, pp.318 identifier

  • Publication Type: Conference Paper / Full Text
  • City: Glasgow
  • Country: England
  • Page Numbers: pp.318
  • Istanbul Technical University Affiliated: Yes

Abstract

This study proposes an intelligent system to perform fault diagnosis actions in ship machinery systems. Considering the cost limitations, the main goal is to optimize the machinery system availability. The model takes the advantage of a classification tool based on support vector machines (SVM) principles. Statistical assumptions are considered for validity of the analysis. The test and statistical demonstration phases are also supported with the data, gathered from the specifically created operational scenarios in ship engine room simulator. Different faulty conditions other than the observed malfunction were inserted to the system in order to provided more realistic approach to simulate real world problem.