Amplitude modulation of sound from wind turbines under various meteorological conditions 

Abstract

Wind turbine (WT) sound annoys some people even though the sound levels are relatively low. This could be because of the amplitude modulated “swishing” characteristic of the turbine sound, which is not taken into account by standard procedures for measuring average sound levels. Studies of sound immission from WTs were conducted continually between 19 August 2011 and 19 August 2012 at two sites in Sweden. A method for quantifying the degree and strength of amplitude modulation (AM) is introduced here. The method reveals that AM at the immission points occur under specific meteorological conditions. For WT sound immission, the wind direction and sound speed gradient are crucial for the occurrence of AM. Interference between two or more WTs could probably enhance AM. The mechanisms by which WT sound is amplitude modulated are not fully understood.

Conclusion

Studying AM is very complex due the many factors that govern sound propagation from WTs. For an ideal analysis of how AM is produced and transmitted, emitted sound power, wind direction, temperature gradients, wind gradients, and turbulence would need to be known three dimensionally in small time steps. This is impossible to measure in the field, so simplifications must be made. Furthermore, the interaction of sound from several WTs complicate the analysis.

Higher prevalence of AM is detected when the sun is close to or under the horizon, which corresponds well with when temperature inversions occurs on clear nights. A temperature inversion near the ground changes the angle of incidence of the sound waves and affects the ground attenuation. The reflected sound waves are normally less damped if the sound comes more from the zenith than parallel to the ground. At the Dragaliden site when AM was present, a typical pattern was approximately 15 s of distinct AM followed by a minute of steadier sound levels.

Analyzing approximately 30h of AM measurements recorded simultaneously at both an emission and an immission point shows that enhanced AM at an immission point could not be explained by enhanced AM at the emission point. It is instead an effect of interference between sound from several WTs or of different ray paths of the sound from one turbine. However, this last possibility requires further testing.

The AM detection method works well and does not react to passing cars, birds, or airplanes. During strong masking, the WT signal is lost using the detection method; the sound will of course not be experienced as amplitude modulated, but the signal may still be present in the background noise. We could conclude from our measurements that amplitude modulated sound from WTs is more common under certain meteorological conditions and is observable approximately 20%–30% of the operational time, depending on the distance from the turbines. In future studies, it would be interesting to investigate WT sound annoyance coupled to conditions with and without AM present.

Journal of the Acoustical Society of America 135(1):67–73, January 2014
doi:10.1121/1.4836135

Conny Larsson and Olof Öhlund
Department of Earth Sciences, Uppsala University

Download original document: “Amplitude modulation of sound from wind turbines under various meteorological conditions”