Wind turbine noise modeling based on Amiet's theory: Effects of wind shear and atmospheric turbulence

Broadband noise generated aerodynamically is the dominant source for a modern wind turbine. In this paper, trailing edge noise and turbulent inflow noise are modeled using Amiet's theory to predict wind turbine noise spectra, directivity and amplitude modulation. First, by comparing model predictions with wind tunnel experiments from the literature, we show that a wall pressure spectral model that includes the effect of an adverse pressure gradient is needed to correctly predict trailing edge noise spectra. Then, we adapt the model to rotating blades and compare sound power level spectra of trailing edge noise with field measurements, assuming a constant wind speed profile. A good agreement is found at frequencies higher than approximately 1000 Hz, but the levels are underestimated at lower frequencies. Finally, we account for wind shear and atmospheric turbulence effects using the Monin-Obukhov similarity theory. On the one hand, we show that angle of attack variations due to wind shear can produce a significant change in the wall pressure spectra of some blade sections, especially in stable atmospheric conditions, even though this effect is not clearly seen on the trailing edge noise spectra at the receiver. On the other hand, turbulent inflow noise does vary with atmospheric conditions, and contributes significantly to the noise radiated by a wind turbine at low frequencies. When both mechanisms are considered, the predicted sound power level spectra are in good agreement with measurements.