Robust lifecycle optimization of turbine components using simulation platforms

Paper i proceeding, 2012

n early phases of turbofan engine component design, simulation is favored since it reduces the need for expensive physical testing. However, deterministic simulations for model validation do not consider uncertainty at all. Uncertainties can be classified into three types: aleatory uncertainties, epistemic uncertainties, and error. In this paper, we investigate the potential of a multidisciplinary simulation platform to address these uncertainties and errors for a given test case. We place specific focus on the geometry assurance of a given turbofan component - the Turbine Rear Structure (TRS). Simulations are generally performed based on nominal geometries, materials and loads. However, when a product is mass-produced, each realization of the product design will deviate from the nominal geometry. By generating CAD models from scanned 3D-data of manufactured parts and running them through the simulation platform, the effect that geometric variation has on aerodynamic and structural performance can be investigated. Further, by moving the reference points in a virtual assembly process, we can, to some extent, suppress the effects that this variation has on aerodynamic and structural performance. From a technical point of view, the suggested approach means a significantly improved ability to numerically simulate and optimize robustness of component designs with functionality criteria from principally different disciplines. From an industrial application point of view, the suggested approach provides a tool for including part variation in the early design face, rather than being treated downstream in the development process.