Development of a robust numerical optimization methodology for turbine endwalls and effect of endwall contouring on turbine passage performance
Airfoil endwall contouring has been widely studied during the past two decades for the reduction of secondary losses in turbine passages. Although many endwall contouring methods have been suggested by researchers, an analytical tool based on the passage design parameters is still not available for designers. Hence, the best endwall contour shape is usually decided through an optimization study. Moreover, a general guideline for the endwall shape variation can be extrapolated from the existing literature. It has not been validated whether the optimum endwall shape for one passage can be fitted to other similar passage geometry to achieve, least of all a non-optimum but a definite, reduction in losses. Most published studies were conducted at low exit Mach numbers and only recently some studies on the effect of endwall contouring on aerodynamics performance of a turbine passage at high exit Mach numbers have been published. There is, however, no study available in the open literature for a very high turning blade with a transonic design exit Mach number and the effect of endwall contouring on the heat transfer performance of a turbine passage.
During the present study, a robust, aerodynamic performance based numerical optimization methodology for turbine endwall contouring has been developed. The methodology is also adaptable to a range of geometry optimization problems in turbomachinery. It is also possible to use the same methodology for multi-objective aero-thermal optimization. The methodology was applied to a high turning transonic turbine blade passage to achieve a geometry based on minimum total pressure loss criterion. The geometry was then compared with two other endwall geometries. The first geometry is based on minimum secondary kinetic energy value instead of minimum total pressure loss criterion. The second geometry is based on a curve combination based geometry generation method found in the literature. A normalized contoured surface topology was extracted from a previous study that has similar blade design parameters. This surface was then fitted to the turbine passage under study in order to investigate the effect of such trend based surface fitting. Aerodynamic response of these geometries has been compared in detail with the baseline case without any endwall contouring.
A new non-contoured baseline design and two contoured endwall designs were provided by Siemens Energy, Inc. The pitch length for these designs is about 25% higher than the turbine passage used for the endwall optimization study. The aerodynamic performance of these endwalls was studied through numerical simulations. Heat transfer performance of these endwall geometries was experimentally investigated in the transonic turbine cascade facility at Virginia Tech. One of the contoured geometries was based on optimum aerodynamic loss reduction criterion while the other was based on optimum heat transfer performance criterion. All the three geometries were experimentally tested at design and off-design Mach number conditions. The study revealed that endwall contouring results in significant performance benefit from the heat transfer performance point of view.