In this paper an innovative on-line condition monitoring system is introduced. It consists of a novel object-oriented database, a machine learning algorithm and a detailed model to predict machinery failure. The Condition Monitoring Database (CMD), which is an Object-Oriented Database (OOD) built entire in Python, has been developed in order to avoid any relational-object mismatch problems which are common when using relational databases in complex applications involving machine learning techniques. The database will intelligently store data from various sensors and then streamline the data into a pipeline to the diagnostic and prognostic system, offering a constant evaluation of the ship machinery at high speed and accuracy. The suggested Condition Based Maintenance (CBM) framework is based on detecting the change of the condition of the machinery in real time by utilizing the Local Outlier Factor (LOF) algorithm for novelty detection. The training, validation and testing of the algorithm have been performed by using data for the degradation of a naval propulsion system. The results show that after tuning the algorithm it identified with high accuracy when a machinery system changed condition for both cases of a turbine and a compressor and for different operational profiles of the ship.
An Innovative Machine Learning System for Real Time Condition Monitoring of Ship Machinery
Lecture Notes in Civil Engineering
Practical Design of Ships and Other Floating Structures ; 2019 ; Yokohama, Japan September 22, 2019 - September 26, 2019
2020-10-04
16 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
Real time condition monitoring , Object-Oriented Database , Machine Learning , Artificial intelligence , Local Outlier Factor , Ship machinery Engineering , Offshore Engineering , Structural Materials , Solid Mechanics , Engineering Fluid Dynamics , Building Construction and Design , Fire Science, Hazard Control, Building Safety
Intelligent condition monitoring of ship machinery systems
British Library Conference Proceedings | 1994
|Real time ship monitoring system and real time ship monitoring method
Europäisches Patentamt | 2022
Real time ship monitoring system and real time ship monitoring method
Europäisches Patentamt | 2021
|