The following work is an in-depth investigation of the heat transfer characteristics and cooling effectiveness of a full-scale fully cooled modern high-pressure turbine (HPT) vane as a result of genetic algorithm (GA) optimization, relative to a modern baseline film cooling configuration. Individual designs were evaluated using 3D Reynolds-averaged Navier–Stokes (RANS) computational fluid dynamics (CFD) that modeled film cooling injection using a transpiration boundary condition and evaluated 10 cells from the wall with an isothermal surface condition. 1800 different cooling arrays were assessed for fitness within the optimization where film cooling parameters such as axial and radial hole location, hole size, injection angle, compound angle, and custom-designed row patterns were varied in the design space. The GA optimization terminated with a unique pressure side (PS) cooling array after only 13 generations. The fitness functions prescribed for the problem lowered the PS average near-wall surface temperature, lowered the near-wall maximum temperature, and maintained the level of near-wall average overall effectiveness. Results show how the optimization resulted in redistributed flow from overcooled areas on the vane PS to undercooled areas near the shroud. The optimized cooling array yielded a reduction of average near-wall gas temperature of 2 K, a reduction in the maximum near-wall gas temperature of 3 K, a reduction in maximum heat flux of 2 kW/m2 and a reduction in pressure loss over the vane, all while maintaining a constant level of surface-averaged overall effectiveness. Methods used to improve pressure side film cooling performance here are promising in terms of eliminating hot spots on individual HPT components in their proper operating environments as well as increasing the potential to use less air for cooling purposes in a gas turbine engine.
Accurate predictions of unsteady forcing on turbine blades are essential for the avoidance of high-cycle-fatigue issues during turbine engine development. Further, if one can demonstrate that predictions of unsteady interaction in a turbine are accurate, then it becomes possible to anticipate resonant-stress problems and mitigate them through aerodynamic design changes during the development cycle. A successful reduction in unsteady forcing for a transonic turbine with significant shock interactions due to downstream components is presented here. A pair of methods to reduce the unsteadiness was considered and rigorously analyzed using a three-dimensional (3D), time-resolved Reynolds-Averaged Navier-Stokes (RANS) solver. The first method relied on the physics of shock reflections itself and involved altering the stacking of downstream components to achieve a bowed airfoil. The second method considered was circumferentially asymmetric vane spacing which is well known to spread the unsteadiness due to vane-blade interaction over a range of frequencies. Both methods of forcing reduction were analyzed separately and predicted to reduce unsteady pressures on the blade as intended. Then, both design changes were implemented together in a transonic turbine experiment and successfully shown to manipulate the blade unsteadiness in keeping with the design-level predictions. This demonstration was accomplished through comparisons of measured time-resolved pressures on the turbine blade to others obtained in a baseline experiment that included neither asymmetric spacing nor bowing of the downstream vane. The measured data were further compared to rigorous post-test simulations of the complete turbine annulus including a bowed downstream vane of nonuniform pitch.
The focus of the study presented here was to investigate the interaction between the blade and downstream vane of the stage-and-one-half transonic turbine via CFD analysis and experimental data. A Reynolds-Averaged Navier-Stokes (RANS) flow solver with the two-equation Wilcox 1998 k-ω turbulence model was used as the numerical analysis tool for comparison for all of the experiments conducted. The rigor and fidelity of both the experimental tests and numerical analysis methods were built through two- and three-dimensional steady-state comparisons, leading to three-dimensional time-accurate comparisons. This was accomplished by first testing the midspan and quarter-tip two-dimensional geometries of the blade in a linear transonic cascade. The effects of varying the incidence angle and pressure ratio on the pressure distribution were captured both numerically and experimentally. This was used during the stage-and-one-half post-test analysis to confirm that the target corrected speed and pressure ratio were achieved. Then, in a full annulus facility, the first vane itself was tested in order to characterize the flowfield exiting the vane that would be provided to the blade row during the rotating experiments. Finally, the full stage-and-one-half transonic turbine was tested in the full annulus cascade with a data resolution not seen in any studies to date. A rigorous convergence study was conducted in order to sufficiently model the flow physics of the transonic turbine. The surface pressure traces and the Discrete Fourier Transforms thereof were compared to the numerical analysis. Shock trajectories were tracked through the use of two-point space-time correlation coefficients. Very good agreement was seen when comparing the numerical analysis to the experimental data. The unsteady interaction between the blade and downstream vane was well captured in the numerical analysis.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.