A directivity correction for accurate semi-empirical wind turbine noise prediction

Public acceptance of wind farms is a significant challenge in the development of wind energy. The acoustic impact generated by wind turbines is a common concern among local residents. The primary noise source in wind turbines is generated by aerodynamics. Atmospheric turbulence reaching the blade leading edge or turbulent boundary layer passing the trailing edge produce the main aeroacoustic sources. The noise generated by these mechanisms is commonly predicted by means of semi-empirical models, which do not demonstrate great reliability when compared to acoustic measurements. This paper presents a correction to the directivity of airfoil noise radiation, resulting in improved sound pressure levels on the ground plane surrounding a wind turbine. This improvement is achieved without requiring any additional computational effort. The sound pressure levels perceived on the ground plane are known to have asymmetrical shape. Maximum noise levels correspond to observers directly in the upwind and downwind locations, whereas the minimum levels belong to the positions close to the rotor plane. Said asymmetrical shape is not represented in the semi-empirical models. The proposed correction takes into consideration the airfoil thickness in the radiation directivity equations, resulting in the expected asymmetrical shape of noise footprints on the ground plane around a wind turbine. The correction was found to not affect the accuracy of the spectrum predicted by the semi-empirical models when compared to dedicated field measurements under the standard IEC 61400-11 procedure. When implementing the proposed correction, the virtual NREL 5 MW wind turbine's published noise footprints, which were originally calculated using computationally expensive methods, are accurately reproduced.