Condition Monitoring and Diagnostic of a Diesel Engine through Detecting Changes in Estimated Dynamic Models

Abstract

In this paper an approach to conchtion monitoring and diagnosis by comparing identified dynamic models is considered. Dynamic models are estimated and a statistical test is used to test for significant differences. As an example a diesel engine model was simulated and analysed with various pertubations to five of its physical parameters. The patterns of changes in dynamic model coefficients caused by changes in those physical parameters were found and used to diagnose the changes. Most of the changes introduced were detected and diagnosed correctly.