Evaluation of Adjoint-Optimized Holes—Part I Baseline Performance

Abstract

With the advent of the use of additive manufacturing to build gas turbine components, the design space for new hole geometries is essentially unlimited. Recently, a computational adjoint-based optimization method was used to design shaped film cooling holes fed by internal co-flow and cross-flow channels. The associated Reynolds-averaged Navier–Stokes computations predicted that the holes optimized for use with cross-flow (X-AOpt) and co-flow (Co-AOpt) would significantly increase adiabatic effectiveness. However, only the X-AOpt hole was tested experimentally in this previous study. Though the experimentally measured performance for this hole was much less than computationally predicted, it still had a 75% improved performance compared to the conventional 7-7-7-shaped hole. In the current study, the X-AOpt and Co-AOpt-shaped holes were experimentally evaluated using measurements of adiabatic effectiveness and overall cooling effectiveness. Coolant was fed to the holes with an internal co-flow channel. For reference, experiments were also conducted with the baseline 7-7-7-shaped hole, and a 15-15-1-shaped hole (shown in a previous study to be the optimum expansion angles for a shaped hole). Furthermore, overall cooling effectiveness measurements were made with engine-scale models to evaluate the performance of additively manufactured (AM) X-AOpt and Co-AOpt holes with a realistic metal build. Results from this study confirmed that the X-AOpt hole had a 75% increase in adiabatic effectiveness compared to the 7-7-7-shaped hole. However, the Co-AOpt hole had only a 30% increase in adiabatic effectiveness, substantially less than had been computationally predicted. Measurements of overall cooling effectiveness for the engine-scale models and the large-scale models followed similar trends.