Influence of blade compactness and segmentation strategy on tonal noise prediction of an automotive engine cooling fan

This paper discusses a blade segmentation strategy for tonal noise computation of automotive engine cooling fans. The noise sources are extracted from an Unsteady Reynolds-averaged Naviers Stocks simulation (URANS), and propagated into the far field using dipole noise equations derived from Ffowcs-Williams acoustic analogy. On the aerodynamic side, a mesh independence study is carried out. On the acoustic side, three levels of acoustic meshes are compared to evaluate the assumption of blade compactness and its influence on the final results. The blade is modeled with one dipole at the lowest level of the acoustic mesh: it is assumed acoustically compact regardless of the frequency. At the finest level, the blade is meshed with 26,726 dipoles which are the surface elements of the CFD mesh. An intermediate level of mesh is also studied where the blade is cut into three acoustically compact strips. A comparison is also established between two approaches for summing the contributions of the dipoles in the cases of the intermediate and fine meshes. The first approach is to neglect the axial, tangential and radial phase lags between the dipoles, and assume uncorrelated sources. The second approach takes into account the phase lags between the dipoles due to the blade’s geometry. It is based on the equations derived by Hanson Parzych [4]. The results are additionally compared to values extracted from experiments carried out in an anechoic chamber.