Performance monitoring on a gas admission valve in a dual fuel engine - Implementation of condition based maintenance by applying machine learning and statistical models

Abstract

An analysis of the performance data from a dual fuel engine is performed, in order to build several prognostic models to predict when a failure occurs on the gas admission valve. The model who performs and provides the best results was selected for further investigations. Based on the findings, a condition based maintenance strategy for the gas admission valve is not recommended to be implemented, because of poor performance and results from the failure prognostic models. For more information, see thesis.