Best-Practice Guidelines for High-Fidelity Simulations Based on Detailed Analysis of a Highly-Loaded Low-Pressure Turbine Cascade

In previous studies, the highly-loaded, low-pressure turbine T106C has been used as a validation test case for direct numerical and large-eddy simulation codes. In contrast to the experimental test campaign, previous high-fidelity simulations simplify the cascades’ geometry by considering only a fraction of the spanwise extent with periodic boundary conditions to analyze the flow in the mid-span region. As a result of these simplifications, experimentally obtained data of the blade loading are notoriously difficult to accurately reproduce as most significant differences have been observed just after suction side flow reattachment. The discrepancies between high-fidelity simulations and experiments have mostly been assumed to originate from insufficient spanwise domain extent, from prescribing laminar inflow conditions in the numerical setup or were left unanswered. To further address these differences, the present paper reports on results obtained from a dedicated high-fidelity computational fluid dynamics solver as well as a multi-purpose open source solver. At the operating condition of isentropic exit Reynolds number of 80,000 and isentropic exit Mach number of 0.65, a series of highly-resolved large-eddy simulations and a reference direct numerical simulation have been performed. Firstly, grid resolution and domain extent were systematically varied to highlight their conjugate effect on the blade load along the suction side. Secondly, the influence of free-stream turbulence and acoustic pressure perturbations determining the turbulence levels in the computational domain are discussed. Finally, a rigorous comparison between two different flow solvers summarizes best-practice guidelines for conducting high-quality, scale resolving simulations of turbomachinery flows.