Integration and Validation of S2-Flow Computations for Aero-Thermodynamic Pre-Design of Turbines

Kurzfassung

In order to speed up the design of turbomachines, the Turbine Department of the DLR Institute of Propulsion Technology is developing a turbine pre-design chain that should organize and accelerate all design steps prior to initial 3D-CFD calculations. Thus, DLR is developing generic tools for preliminary design projects of turbine design. This paper reports the activities led by DLR for turbine pre-design, focusing on advanced 2D-calculations and optimization. The preliminary projects for the aero-thermodynamic design of turbines are supposed to provide reliable results in a short time. Nowadays the 3D-CFD is recognized for the precise results it can provide, but the significant computation times are yet a limitation for its use in preliminary projects. For the turbine pre-design, methods of fluid computation are chosen according to their computation rapidity more than the precision of the provided results. Hence 0-2D computation methods are used for the pre-design. The advantage of advanced 2D-computations for turbine pre-design is quickly introduced, before presenting the turbine pre-design chain developed by DLR. The main tools are then described separately. The integration of S2-computations (i.e. SLC4T program code) in the turbine pre-design process should contribute to accelerate significantly the switch from one step to the next. With this work, some routines were implemented, so that 1D pre-design can be followed automatically by SLC4T computations for providing turbine map data files. Second big task for this work was to implement a preconfigured and general optimization process using AutoOpti for annulus geometry of turbines that could be quickly set up coming from PrEDiCT 1D predesign. Many optimization criteria, such as the regions of interest of parameter values, are strongly problem-oriented. As regards S1-stream surfaces, human know-how is deeply involved when designing blade profiles. The complete automation of blade design itself is therefore not foreseen. However, it would be of interest to implement an automatic update of S2-flow computations input files, once the blades have been designed precisely. By this way the optimization process of the annulus contour could be performed on the turbine with updated blade parameters. Furthermore as HPT and LPT are optimized together, it could be of interest to optimize the inter turbine duct shape as well.