Quantifying uncertainties during the early design stage of a gas turbine disc by utilizing a bayesian framework

Quantifying uncertainties regarding the grain size of the turbine disk has been identified as a crucial aspect for the preliminary design stage. The reason for that is because the grain size is correlated to the life of the component which should preferably be maximized or at least quantified to the best of the designer’s abilities. In the grand scheme of things, this ultimately translates into a potential competitive advantage for the aero engine company. The prime focus of this paper is the investigation of material properties which was done by combining simulation and experimental data within a Bayesian framework in order to enhance the decision making process during the preliminary design stage. The aim of the case study presented here was to show how the physical processes can be modelled using a Bayesian network which updates prior probability distributions with real data in order to obtain more accurate predictors of reality. The first part of the paper explains the theory behind the framework, while the latter half shows some results as well as some conclusions which can be drawn.