At the preliminary design stage of the engine design process, the behaviour and efficiency of different engine designs are investigated and evaluated in order to find a best matching design for a set of engine objectives and requirements. The prediction of critical part temperatures as well as the reduction of the uncertainty of these predictions is decisive to bid a competitive technology in aerospace technology. Automated workflows and Design of Experiments (DOE) are widely used to investigate large number of designs and to find an optimized solution. Nowadays, technological progress in computational power as well as new strategies for data handling and management enables the implementation of large DOEs and multi-objective optimizations in less time, which also allows the consideration of more detailed investigations in early design stages.
This paper describes an approach for a preliminary-design workflow that implements adaptive modelling and evaluation methods for cavities in the secondary air system (SAS). The starting point for the workflow is a parametric geometry model defining the rotating and static components. The flow network within the SAS is automatically recognized and CFD and Thermal-FE models are automatically generated using a library of generic models. Adaptive evaluation algorithms are developed and used to predict values for structural, air system and thermal behaviour. Furthermore, these models and evaluation techniques can be implemented in a DOE to investigate the impact of design parameters on the predicted values. The findings from the automated studies can be used to enhance the boundary conditions of actual design models in later design stages.
A design investigation on a rotor-stator cavity with axial through flow has been undertaken using the proposed workflow to extract windage, flow field and heat transfer information from adiabatic CFD calculations for use in thermal modelling. A DOE has been set up to conduct a sensitivity analysis of the flow field properties and to identify the impact of the design parameters. Additionally, impacts on the distribution of the flow field parameters along the rotating surface are recognized, which offers a better prediction for local effects in the thermal FE model.