ENGINEERING AND CFD

Results and Methods

Unsteady Computational Fluid Dynamics (CFD) simulations were performed using the finite volume method. The film cooling domain was discretized with up to 14 million hexahedral elements. Based on this, the conservation equations for mass, momentum, and energy, known as the Navier-Stokes Equations (NSE), was solved using the open-source CFD software OpenFOAM. Large Eddy Simulations (LES) were conducted, which resolved over 95% of the turbulent eddies within the film cooling flow domain. This allowed for a high accuracy in predicting film cooling performance. The simulations for each design under laminar and gas turbine-like conditions were run on approximately 1,000 cores of SuperMUC-NG per run for 2-8 weeks. Throughout this investigation, we were able to utilize up to 5,000 CPU cores simultaneously to simulate different film cooling designs, enabling us to produce the result [3] within a reasonable timeframe. First, we discuss the time-averaged results for discrete film cooling holes and for the two investigated trenched designs. In Figure 3, a side view of the temperature field in the center plane of a film cooling hole configuration is presented. The color blue indicates low temperature, while the hot gas temperature is represented by red color. The solid line (-) corresponds to gas turbine-like high turbulence level conditions, whereas the results of the laminar hot gas flow simulations are depicted by the dotted lines (- -). The objective of film cooling is to maintain a low temperature at the wall. The state-of-the-art, simple effusion hole design features a film cooling jet emerging from the cooling hole that lifts off the wall, which is typically encountered in effusion-cooled combustor walls. With normal effusion cooling (top), an increased turbulence level reduces the jet lift-off slightly, resulting in a positive cooling effect as the coolant is transported in closer proximity to the wall. By employing a trench, the jet makes direct contact with the trench wall. Subsequently, the coolant flow can more effectively adhere to the downstream wall, creating a better cooling film. However, this efficiency is compromised slightly by elevated levels of free-stream turbulence for both designs. The next set of results focuses on the film cooling coverage immediately downstream of the trenches. lnstantaneous images of the film coolant distribution near the wall for both investigated trench designs are shown in Figure 4. In snapshot #1, the axial and lateral distribution of the coolant at the wall showed a low temperature downstream of the trenches for both designs. In a subsequent instant (#2), the axial extent of the film cooling efficiency was reduced for both configurations. Additionally, a hot spot emerged just downstream of the segmented trench. The transverse trench exhibited less sensitivity in the central region of the wall due to the higher trajectory of the coolant, as shown in Figure 3. However, it displayed significantly lower coolant coverage at the sides compared to the segmented trench. In instant #3, the hot spot downstream of the segmented trench shifted even further downstream. Moreover, brief hot spots emerged at both lateral sides. These results demonstrate how Large-Eddy-Simulations allow a detailed evaluation of the unsteady features of such highly turbulent flow fields and their effect on the wall heat fluxes.