Development of URANS and Fast Multipole Boundary Element Method Workflow for Installed Propeller Noise

Abstract

For the simulation of isolated propeller noise, a well-established computational aeroacoustics (CAA) approach relies on the simulation of the unsteady propeller blade loads with the unsteady Reynolds-Averaged Navier-Stokes (URANS) method for the aerodynamic part. The porous Ffowcs-Williams and Hawkings (FW-H) integral method is then applied for propagating sound waves to a far-field. Under realistic propeller installation conditions, one has to account for the propeller wake and airframe interaction, as well as for the shielding effects. The URANS / FWH approach needs to be extended to accomplish that, meaning that the FW-H surface must enclose significant parts, if not all of the aircraft. The problem arises where the URANS simulation lacks quality on distant surfaces due to excessive numerical dissipation and dispersion. If these data are treated as unknown the surface integral is best tackled by means of a fast multipole boundary element method (FM-BEM). One challenging part is to verify that the FW-H integral and the Kirchhoff's formulation used in the FM-BEM are compatible. In terms of the application, the DLR's own Dornier 228 STOL aircraft (registration D-CODE) that features five-bladed twin propellers has been chosen as a test bed. Several potent noise mechanisms are considered: propeller induced noise, vortex-surface interaction noise and hydrodynamic near-field interference with the airframe. The latter could be eliminated by improved propulsion system integration. Thus, a design study is carried out where the distance between the propeller plane and engine is increased to evaluate the noise reduction potential.