Conditional Residual Time Modelling Using Oil Analysis: A Mixed Condition Information Using Accumulated Metal Concentration and Lubricant Measurements

• Main Authors: Wang, W., Hussin, B.
• Format: Conference or Workshop Item
• Language: English
• Published: 2006
• Subjects: Q Science (General)
• Online Access: http://eprints.utem.edu.my/id/eprint/173/1/ICME2006_Proceedings_Mixed_Information_for_Oil_Analysis.pdf; http://eprints.utem.edu.my/id/eprint/173/
• Institution: Universiti Teknikal Malaysia Melaka

This paper reports on modelling decision problems associated with condition monitoring by providing some scientific support to maintenance decision making. In this study, a Spectrometric Oil Analysis Programme (SOAP) is used as monitoring measures for diesel engines used in ships. Using the residual time at a monitoring check as the condition of the diesel engines, we seek to establish the relationship between the residual time and the total wear concentrations which are available from SOAP data. These metal concentrations are treated as concomittent variables which are influenced by the residual time. We also used lubricant measurements data acting as covariates that may increase/decrease the residual time. The formula for finding a residual time distribution is presented using a filtering technique and a method for estimating the model parameters is discussed. Once the distribution of the residual time is known, a decision model can be established to recommend the optimal maintenance actions in terms of cost, availability or any criterion of interest.