The majority of sound measurements presented to the public by the wind turbine industry are so-called “A-weighted” or dB A measurements.
According to “RenewableUK”, the website of the British Wind Energy Association (http://www.bwea.com/ref/noise.html, 11/2/2010) “the noise from wind turbines is very low. Outside the nearest houses, which are at least 300 metres away, and more often further, the sound of a wind turbine generating electricity is likely to be about the same level as noise from a flowing stream about 50-100 metres away or the noise of leaves rustling in a gentle breeze. This is similar to the sound level inside a typical living room with a gas fire switched on, or the reading room of a library or in an unoccupied, quiet, air-conditioned office.” They then state that the noise level from a wind farm at 350m is typically 35-45 dB A.
The problem with such measurements is that they totally ignore the large infrasound component of wind turbine noise, which is the component that “bothers the heck out of people”.
The A-weighting curve corrects measurements according to the sensitivity of human hearing, de-emphasizing low frequency components for which the ear is insensitive. It is valid to use this correction if hearing the sound is the prime concern, but is not appropriate to use this approach when processes unrelated to hearing (such as whether a low frequency sound affects your ear) are being considered.
The A-weighting curve is not by any means a “small correction”. At 1 Hz (the dominant frequency of wind turbine noise) the correction amounts to over 140 dB, which is equivalent to dividing the measured voltage by over 25 million. If you take the quietest sound you can hear, and compare it with the sound of a 747 Jumbo jet flying 10 feet above your head under full power, then you may be approaching 140 dB. The point is that this is a massive correction.
This figure shows the influence of A-weighting the wind turbine noise spectrum obtained by Van den Berg 2006. A-weighting completely deemphasizes the low frequency components as if they didn't exist.
Because one of the main influences of wind turbine noise may be caused by the infrasound component of the noise, this manipulation has the effect of totally ignoring this potential source of the problem.
The A-weighted spectrum does not take into account that other structures in the ear are more sensitive to infrasound and are affected to a greater degree than hearing.
A-weighting wind turbine noise is equivalent to taking sunlight and considering only the visible portion of the spectrum, then concluding that sunlight is completely safe and there will be no adverse effects if you spend all day laying in the sun.
We know this logic is not valid for sunlight and the same logic is just as invalid in justifying A-weighted measurements of wind turbine noise.
Once the sound has been A-weighted (taking out all the infrasound component, the remaining peak is then about 40 dB, representing the 40 dB A measure.
Now the level is comparable with “the rustling of leaves”. Comparing wind turbine noise with the “rustling of leaves” and the sound of a “flowing stream” is disingenuous at best. It may be what is heard BUT it is totally misleading because the noise from rustling leaves and flowing streams is NOT DOMINATED BY INFRASOUND.
It is well established that infrasound does have effects on people. Indeed it bothers them intensely. So, to ignore this component of the noise is to ignore the major problem that this noise causes to people, by disturbing sleep and with subsequent damage to their health.
How different frequency-weighting functions affect wind turbine noise measurements
The common weighting functions related to low frequency noise are shown in the panel at the left, below. This includes the A-weighting (red) , C-weighting (purple) and G-weighting (green) curves. To see how these functions affect sound measurements, consider the panel at the right which shows an unweighted wind turbine spectrum (blue) and the same spectrum subjected to A-, C-, or G-weighting. As shown above, A-weighting cuts most of the low frequency sound out, leaving the peak of the A-weighted spectrum at around 42 dBA. (Ah – Did I hear a leaf rustle over there? No, it was just the refrigerator in the next room turning on.)
C-weighting (purple) does not cut out low frequencies so much, so the peak of the C-weighted spectrum of the same noise is higher, in this case around 56 dBC
G-weighting (green) emphasises the higher infrasound frequencies (10-20 Hz) and cut out frequency components lower and higher than this. The peak of the G-weighted spectrum is higher (68 dBG) which compares well with the G-weighted measurements by Jakobsen, J., 2005 Infrasound emission from wind turbines Journal of Low Frequency Noise Vibration and Active Control 24, 145-155.
These comparisons show that the measurements of wind turbine noise depends highly on exactly how the different frequency components are considered in, or excluded from, the measurement.
A-weighting cuts out most low frequency sound and gives the lowest reading.
C-weighted allows more low frequency sound and gives a measure intermediate between A- and G-weighting.
G- weighting gives a measure of the infrasound (10-20 Hz) components
Unweighted measures give the highest peak readings
Question to the group: Why do you think the wind turbine industry typically characterizes wind turbine noise with only A-weighted measurements?