Efficient modeling of manufacturing uncertainty is critical in the design of turbine engine components. In this research, a multilevel validation framework is developed to efficiently account for the geometric and material uncertainty associated with the manufacturing process to accurately predict the performance of engine components. This framework is created to handle both epistemic and aleatory uncertainty. Specifically, the spatial variability of the uncertain geometric parameters obtained from coordinate measuring machine data of manufactured parts is represented as aleatory uncertainty. Porosity and defects in the manufactured parts based on 3D X-ray CT scanned images are represented as epistemic uncertainty. Multiple efficient statistic tools are integrated into the proposed framework. Karhunen-Loeve expansion is utilized to create a set of correlated random variables from the obtained uncertainty data and a fine scale finite element model of the component is created that accounts for the uncertainties quantified by these correlated random variables. A stochastic upscaling method is then developed to form a simplified model that can represent this detailed model with high accuracy under uncertainties. In addition, a validation method for multi-variate responses is developed and used to validate the simulation results with the experimental results. The modal frequency analysis of a turbine blade example is used to demonstrate the efficacy of the proposed framework. The application results show that the proposed method effectively captures the geometric uncertainties introduced by manufacturing while providing accurate predictions under uncertainties.
Cite this work
Researchers should cite this work as follows:
MATIN Development Team