Documenting Wind Turbine Sound



Do you know why welders use a heavy mask? This picture doesn't show you. It is because the welding arc generates an intense amount of ultraviolet (UV) light. Even if the welder is just a little slow in covering his eyes with the mask, the UV light may be enough to cause welder's flash, which is sore, red, painful eyes. (I know, because it caught me once and it is a lesson you never forget).

But you would never know that from this picture. Because the camera is insensitive to UV light and pictures cannot show what the real problem is.

To identify the real problem you need a camera or other measurement system that is sensitive to UV light.



Similarly, most systems used to record wind turbine noise are not sensitive to the “unheard” (equivalent to invisible) infrasound component. So when camcorder videos are posted on You-tube, or when news crews turn up to record a segment about wind turbine noise for the local news, their equipment is incapable of demonstrating the problem. Many types of microphone (such as moving coil types) are insensitive to infrasound. And if the microphone does detect it, the infrasound is often filtered out electrically so that it doesn't “saturate” the recording system. These devices are designed to reproduce the sounds you normally hear. They are not designed to reproduce the unusual infrasound component of wind turbine noise that you cannot hear.

Even if the wind turbine sound was recorded with a condenser microphone and amplifier system capable of accurately reproducing the infrasound, it would still be hard to demonstrate this sound to people away from the wind turbines. Most speaker systems or headphones cannot generate infrasound adequately. It is possible that with headphones carefully sealed to the external ear canal, the sound could be replicated but at present there are no commercial systems available to do this.

This means that it is very difficult to show people away from wind turbines (politicians, doctors, etc.) what the effects of the noise are. So it is important to be aware of this important limitation of present technology. The sounds you hear from recorded media do not represent the “real”, “complete” sound of a wind turbine. As in the picture above, such recordings are missing what may be the most important component with regard to human health, which is the infrasound portion.

Detailed documentation of wind turbine noise characteristics (such as unweighted sound spectra) under different wind and load conditions is, surprisingly, virtually absent from the scientific literature. There are a few publications in “obscure” journals that we cited in our publication. It is notable that there are no recent publications in the Acoustical Society of America (of which I am a member) that document the noise characteristics of wind turbines. One would think that no-one was interested in the topic ! Since many of these measurements are “proprietary data”, commissioned by wind turbine operators, the fact that they are not made public itself speaks volumes about the nature of this noise.

Finally, visits to windfarms by politicians, neighbors-to-be and the like is easily subject to manipulation. It has been suggested that turbines may be set to “freewheel” during such visits (with minimal generator loading and blades orientated at a shallow angle to the wind) so that their noise generation is minimized. Any assessment or demonstration of wind turbine noise needs to be performed under normal functional conditions. The most appropriate monitoring of such noise should be through “dosimeters” recording noise levels within people's homes, 24 hours per day. This should include both A-weighted sound (representing the sound you can hear) and G-weighted sound levels (which represents the infrasound component). It is estimated that you can hear infrasound at about a 95 dB G levels, but the cells of the inner ear are stimulated by about 60 dB G. So dB G measurements of over 60 dB must be considered significant. Alternatively, wide band spectra (sensitive down to at least 1 Hz) could be measured.