Beschreibung
The consideration of uncertainties within the design procedures of jet engine components implies the possibility of reducing rework procedures in order to increase the robustness of preliminary designs. Here, automated routines for multidisciplinary system evaluations, analysis and optimization were developed within the past decade. These systems imply the ability of combining a various number of fluid and structure descriptive models and provide the design engineer with an enormous advantage during the decreasing time periods in modern development cycles. In order to reduce the rework time due to an uncertain system behavior, uncertainty needs to be considered within the design process in the first place. Following this, the presented thesis outlines a strategy to make the system's uncertainty available for an automated design process, like multidimensional design analysis or optimization. It is shown how the deterministic design problem of a high-pressure compressor blisk's vibration behavior can be transformed into a probabilistic design problem considering the geometric uncertainty within the manufacturing process. There it is demonstrated how it is possible to capture the geometric variation of a bladed compressor disk and to process these data sets of surface meshes in order to quantify the geometric uncertainty. These statistic geometry descriptions are used afterwards to produce probabilistically generated compressor airfoils and make them available for CAD-based, multidisciplinary design analysis or optimization routines. The presented algorithms, models and procedures are validated throughout the complete workflow to outline the magnitude of this thesis' contribution to future compressor design procedures.
