Wind turbine noise synthesis from numerical simulations

This paper proposes a methodology for physics-based auralization of wind turbine noise with potential application to complex urban and rural scenes. Time-dependent aerodynamic broadband noise levels are computed with cost-efficient semi-analytical models that account for the presence of ground and atmosphere. The algorithm for an accurate reconstruction of the time domain sound signal is provided, with particular attention to the calculation of the Doppler effect in the time domain. Thanks to its low computational cost, this methodology can be applied simultaneously to multiple wind turbines in a complex environment. The scenario can potentially include several ground types, buildings, and other noise sources such as road or railway traffic. The numerical workflow is first validated in free-field with Schlinker and Amiet's analytical theory. The accuracy of the predicted noise levels is assessed by comparing them with measurements available in the literature for a 2.3 MW wind turbine. The spectrograms and the audio signals show the "swishing" amplitude modulation for receivers close to the crosswind direction. Besides wind energy application, the workflow presented in this work can be applied to auralize the broadband noise generated by any rotating machine, such as a helicopter and fans.