Component State Prediction Based on Field Data : Master Thesis in Energy System Engineering

Abstract: This master thesis is part of a big project at Siemens Industrial Turbomachinery (SIT) in Finspång aimed to use the operation experience available at SIT to predict the state of the gas turbines in general and some mechanical components in particular. The objective of the thesis is to continue the development of a prediction model based on experience data for estimations of a components lifetime. In a previous master thesis by Alessandro Olivi statistical analysis of environmental attributes effect on the expected lifetime of components in a gas turbine was performed. Olivi’s thesis constitutes the starting point on which to keep building to create a reliable prediction model. In this thesis extensive validation tests have been performed in order to further quantify the reliability of the model. Investigations aimed towards finding ways to further develop and improve the prediction model are carried out. The relevant new findings are applied to the model and analysis concerning improvements in the prediction accuracy is carried out. It was revealed that the model is able to make accurate predictions for most of the validation points for each failure mode, but more research is needed to obtain a completely reliable prediction model.