High-fidelity simulations of turbomachinery components at full-scale Reynolds number (Re) characteristic of the application in flight are of high interest as they allow assessment of designs early in the development phase. However, when employing turbulence-resolving techniques such as large eddy simulation (LES), these calculations are computationally very expensive given the high Re of the application. Using the exascale supercomputer Frontier at the Oak Ridge Leadership Computing Facility (OLCF), wall-resolved LES of an open fan blade at full-scale Re and at a highly-loaded takeoff condition has been performed. In the configuration considered, the forward blade row is rotating while the downstream blade row is stationary. Based on the high-order LES data, we investigate the evolution and mixing of the flow downstream of the fan blade from the blade trailing edge to a location near the leading edge of the downstream stator.While the stator is not included in the present simulation, the study is directly relevant to interaction noise as it focuses on the evolution of the blade wakes and tip vorticity to understand both the tonal and broadband sources that lead to interaction noise. Notably, while there are differences in details of the wake shapes and mixing rates between the RANS models and LES, the resulting gust upwash signature at the stator leading edge for the wake portion of the fan gust is still predicted to be within~2dB for low frequency interaction tones as well as turbulent kinetic energy. Larger differences are observed in the tip region, consistent with literature on turbulence modeling for wing tip vortices. These modeling deficiencies of RANS are highlighted for future investigation.
No takes yet. Share an insight, caveat, or question.
Karve et al. (2024) studied this question.
Synapse has enriched 4 closely related papers on similar clinical questions. Consider them for comparative context: