Investigation of inlet and diffuser geometry modifications on film cooling performance of additively manufactured shaped holes in crossflow

Film cooling holes permit gas turbine firing temperatures to significantly exceed the melting point of the constituent materials by venting compressor bleed air along the surface of a component forming a buffer between the wall and surrounding gas. A film cooling hole is a passive geometric feature with performance entirely derived from the holes geometry and the operating conditions of the coolant and mainstream. Significant effort has been made to characterize a wide variety of hole geometries but no method has been put forth to determine the optimal hole geometry for a given local flow field and component. Even for traditional, subtractive machined holes this would be a daunting task, but the difficulty grows exponentially as additive manufacturing (AM) permits greater design freedom to the thermal engineer. Presented here is a validated method for determining the optimal film cooling hole geometry of both traditionally or additively manufactured components using computationally inexpensive RANS CFD. Additionally, beyond just validating existing designs, this method can generate novel designs which leverage additive manufacturings unique design space to significantly enhance performance beyond what is possible with traditionally machined holes. While this method has many limitations inherited from RANS, which we will explore in depth, it has proven robust and effective at calculating performance in any coolant/mainstream flowfield. This work stands unique in film cooling literature but will hopefully be superseded by improved methods still to come. Realizable K-epsilon RANS is validated and found to be robust in predicting the flow field of film cooling holes. This information is used to investigate the flow inside of holes where traditional experimental methods are severely restricted. Key separation regions at the inlet and diffuser are identified to be severely detrimental to film cooling performance. CFD was used to predict geometries that would improve hole performance leveraging the unique design freedoms of additive manufacturing. This resulted in large performance gains as predicted by the RANS. Furthermore, as the gross separation regions within the hole were improved, the RANS predictions of surface temperature were found to be increasingly reliably. CFD was employed to search for better performing traditional and AM diffuser designs, the best of which were verified experimentally to significantly improve performance as predicted. Finally, adjoint optimization was used to fully optimize the hole geometry yielding further improvements in performance which were again experimentally validated