Using Functional Knowledge in Lubricant Condition Monitoring for Diagnostics and Prognostics

Abstract

Lubricant condition monitoring has been widely applied for equipment diagnostics and prognostics under condition-based maintenance (CBM). The main purposes of using lubricant are to fulfil requirement of machinery performance and indicate health status of equipment, which imply that the condition of lubricant can not only be used to predict future performance but also diagnose mechanical failures. However, there is deficiency of functional knowledge in existing lubricant information analysis methods, hence the results are usually hard to be interpreted as effective maintenance decisions. This paper presents an approach based on multilevel flow modeling (MFM) to represent the lubricant functions and establish function structure of mechanical equipment that can clearly show the relationship between lubricant and the other components in a functional perspective. The causal feature of MFM allows to utilize a function model to diagnose the possible causes of abnormal lubricant condition and predict further influences. The approach is demonstrated by a hydrodynamic bearing system.