A DIAGNOSTIC MODEL FOR MACHINERY FAULT DETECTION

A fault diagnosis system for ship machinery plants is being developed in The Netherlands. The monitoring system should assist the operator to control and monitor the machinery plant by fault prediction, diagnosis and by giving recommendations for operation and maintenance. As a test-case a diagnostic model has been developed for a water cooling system of a diesel engine. In this diagnostic model a component/subsystem reference model is used to determine in each operational condition the healthy behaviour of a component or subsystem, which can be compared to the actual behaviour. The diagnostic process is of a model based type. A simulation model is used not only to produce the reference model, but is also used for fault simulation. The model has been developed based on the physics of the process. This simulation model is used not only to produce the reference model, but is also used for fault simulation. The model helps to determine the relationship between faults and symptoms and to determine the limits at which faults become unacceptable. The diagnostic model uses trend analysis to predict the moment at which a fault will exceed its limit and will become unacceptable. The presented research aims to develop flexible techniques which can be applied to other more complicated systems as well.