The engine-cycle performance of jet engines can be improved by more efficient cooling systems, either by reducing the required cooling air or by intensifying the cooling efficiency with the same amount of cooling mass flow. However, the multitude of geometrical design parameters and the strong multidisciplinary aspect of cooling mass flow consumption optimization make designing the cooling systems extremely challenging. Integrating probabilistic methods into the thermal design process enables the automated evaluation of multiple design variants which contributes to the development of more efficient systems.
In the present study, the sensitivity of a multi-pass cooling system to geometric variations is investigated. The cooling air flow, solved using a 1D, correlation based flow solver, is iteratively coupled with the 3D-FE thermo-mechanical analysis of the blade. The geometry of the cooling system is varied using the Harmonic-Spline-Deformation parametric, which has been extended to modify the wall thickness enabling to perform a geometrical-holistic analysis. Furthermore, the Elementary-Effects-Method (EEM) and the Monte-Carlo-Simulation (MCS) are compared to identify the most influential parameters and analyze their complex interactions. It is shown that the cooling system’s performance is mostly affected by the shape and position of the first web. Furthermore, MCS proves to be robust towards changes in design space while simultaneously enabling a more detailed analysis of the system behavior compared to EEM.