In the direct-computation approach of aerodynamically generated noise an extended computational domain is employed, which heavily increases the computational cost. Reynolds numbers of practical applications come into reach only by using large-eddy simulations (LES) combined with high-order accurate schemes, optimized time-integration methods and effective subgrid-scale (SGS) models. These simulations are still very costly and therefore parameter studies (physical or numerical) are usually not performed. In particular, grid-independence studies are seldom found in the literature. Rare exceptions, for comparatively low Reynolds numbers, are certain simulations for which results from direct numerical simulations (DNS) are available. In the present work, the azimuthal resolution effect on the directly computed noise of subsonic jet flow at a Reynolds number (based on the jet diameter) of Re = 4.5·10~5 and a Mach number of Ma = 0.9 is investigated. The results for this study are more complicated to compare because of the unavoidable lack of reference DNS data at this Reynolds number. Therefore, conclusions about SGS model effects on the simulation results cannot be drawn. Nevertheless, as it is well-known that jet flows are very sensitive to changes of the initial conditions we want to assess the predictive quality of our simulation results and thus an investigation of resolution effects seems most appropriate.
展开▼
