Analysis of lubrication oil contamination by fuel dilution with application of cluster analysis

Abstract

Contamination via fuel dilution can result in degradation of the engine oil which could lead to increased engine wear and/or engine failure. Exposing lubricant contamination by fuel dilution can involve a multi-parametric evaluation while undertaking lubricant condition monitoring also known as used oil analysis (UOA). We utilize hierarchical cluster analysis to extricate the contamination related parameters caused by fuel dilution in the medium speed engine lubricant. Cluster analysis classifies several objects into some cluster according to the similarities between them but the results are uncertain due to sampling error of the data. We address this by evaluating the probability values (pvalues) for each cluster, while Cluster validity is ensured by using a module to offer the best clustering scheme among different results. The result of the cluster analysis is analysed using expert assessment and literature to pick out the effects of fuel dilution contamination in a lubricant used in a medium speed engine. This methodology is applied to thermal power plant UOA case study. The novelty of the study is firstly, exposing the used oil parameters associated with fuel dilution using hierarchical clustering, secondly, assessment of the uncertainty of the hierarchical cluster developed, thirdly, use of expert assessment to confirm the cluster relationship developed as well as the effects of this contamination to the performance and integrity of the engine, and lastly provide insights for maintenance decision making and moreover, highlighting critical used oil analysis parameters that are correlated which are indicative of degradation via fuel contamination. By addressing such related parameters in UOA, organizations can better enhance the reliability of critical operable equipment like engines