High-fidelity numerical investigation of a real gas annular cascade with experimental validation

Abstract

This study aims at investigating real gas flow in the complex geometry of the Cambridge University annular turbine cascade using numerical simulations. The objectives include validating the numerical approach and understanding the loss mechanisms in this configuration. The numerical results are compared to experimental measurements obtained at various locations in the domain. Two turbulence modeling techniques, large Eddy simulation (LES) and Reynolds-averaged Navier–Stokes (RANS), are employed to assess the influence of turbulence models and inlet turbulence levels. The results show good agreement between numerical simulations and experimental measurements in regions upstream of the trailing edge. However, discrepancies arise in the transition region of the suction side boundary layer, and RANS results are influenced by the choice of turbulence injection. In the wake of the blade, both RANS and LES accurately predict the stagnation pressure ratio, with some slight differences in shock positions and total pressure levels. The analysis reveals that large vortical structures at the hub contribute significantly to the overall losses in this annular configuration. The study quantifies losses due to boundary layers, the wake, and vortical structures using a loss coefficient, with RANS and LES producing slightly different results. These differences, while calling for further experimental measurements, also hint at the possible inaccuracy of the present turbulence models in the context of real gas flows for which a dedicated modeling effort is required.