DNS and ILES of Wall Bounded Flows Using a Discontinuous Galerkin Method and Inlet Synthetic Turbulence

High pressure turbine (HPT) blades are cooled partly by guiding air from the compressor through internal ducts. The massive boundary layer separations in their bends are the main source of head losses and may have a significant impact on the local heat transfer distribution. To capture such physics accurately, scale-resolving simulations such as Direct Numerical Simulation (DNS) or Large Eddy Simulation (LES) are required. Nevertheless, due to the complex configurations of cooling channels in turbomachinery and the level of accuracy required by scale-resolving simulations, standard industrial or academic solvers may not be able to predict efficiently this type of flow. It was shown in recent literature that discontinuous Galerkin method (DGM) combines the accuracy of academic solvers and the flexibility of industrial codes. These methods have shown large potential to perform efficient scale-resolving simulations (see e.g. [5]). In this paper, the capability of a DGM to perform fluid flow analysis of HPT cooling channels will be investigated. For internal flow configurations, the flow forcing strategy to obtain fully developed turbulence occurring at industrial conditions is an issue of primary importance. The generation of synthetic fluctuations, as described in [3, 7] will be tested. The first validation case is the turbulent channel flow at 𝑅𝑒𝜏=590 . The second part of the study will address a problem relevant in the framework of turbines cooling passages: the turbulent flow in a serpentine duct at 𝑅𝑒𝜏=180 .