Chapter 2

Noise and Noise Measurement

2.1        The sound generated by wind turbines is caused by the conversion of wind energy to rotational and acoustic energy.  The rotational energy produces electricity while the acoustic energy produces sound.[1] According to the Clean Energy Council: 

Noise is often the most important factor in determining the separation distance between wind turbines and sensitive receivers like houses. The assessment of noise therefore plays a significant role in determining the viability of and the size of wind farms.[2]

2.2        The committee received evidence from many of the same acousticians as the Community Affairs References Committee in its 2011 inquiry into the social and economic impact of rural wind farms.  The Clean Energy Council, and the acoustic consultants, Sonus Pty Ltd provided the committee with the same technical information about the nature of sound as they did to the Community Affairs Committee.

2.3        The Sonus paper, prepared in 2010, discusses two principal types of noise that a wind farm may generate.  These are mechanical noise from the turbine itself, and aerodynamic noise from the operation of the blades.  Within the category of aerodynamic Noise the Sonus paper includes different types of noise:

• Amplitude Modulation – Commonly described as the "swish" noise that comes from the blades as they rise and fall;
• Low Frequency Noise – has a frequency range between 20 and 200 (Hz) often described as a "rumble"; and
• Infrasound – has a frequency range under 20 Hz and often described as inaudible.[3]

Much of the controversy in this inquiry concerns the subset of noise categories within aerodynamic noise.

Low Frequency Noise and Infrasound

2.4        Mr Steven Cooper from the Acoustic Group submitted that there are 'low frequency, infrasound components' in wind turbine noise that have:

...a unique signature associated with turbines and you can measure them near the turbines and measure them up to seven kilometres away...and seven kilometres away I can see this signature and the pattern is there. You cannot hear it because it is lower than the threshold of hearing, both in frequency and in level, but it is there.[4] 

2.5        Professor Hansen added that low frequency noise is particularly difficult to avoid, as the techniques used to mitigate higher frequency are significantly less effective:

The problem with wind farm noise is that it is dominated by low-frequency noise by the time it gets to people's residences. Many residences, especially if windows are open, are sort of transparent to that noise.

The noise level at low frequencies is not much less than what it is outside, whereas the higher-frequency noise—if there is a little bit left—gets attenuated through the walls of the house and the roof. What you are left with when you are inside is a dominant low-frequency noise, and there is no higher-frequency noise to mask it. There is nothing to mix with it. It is just this low-frequency, annoying noise.[5]

2.6        Sonus discussed the nature of infrasound from wind farms in their paper for the Clean Energy Council. It described the low frequency noise, which includes infrasound, as being:

...easily measured and can also be heard and compared against other noise sources in the environment. Low frequency sound produced by wind farms is not unique in overall level or content and it can be easily measured and heard at a range of locations well in excess of that in the vicinity of a wind farm.[6]

2.7        Dr Leventhall's paper in the journal Canadian Acoustics cited showing that wind turbines do produce infrasound but not at perceptible or harmful levels:  

Modern up-wind turbines produce pulses which also analyse as infrasound, but at low levels, typically 50 to 70dB, well below the hearing threshold. Infrasound can be neglected in the assessment of the noise of modern wind turbines (Jakobsen 2004).[7]

2.8        The current NSW guidelines, which are probably some of the most stringent in the world, also discount low frequency or infrasound as a significant component of wind turbine noise emissions. 

Noise Measuring Methodology

2.9        Mr Cooper described the difficulties in measuring noise in his experience over 35 years:

In some cases I have been to houses and I could not hear a thing and I could not measure anything. That is the nature of the beast. Sometimes the wind blows in different directions. That is the variability that you get. It happens in all sorts of noise studies. Take noise from a hotel. Sometimes there is a noise problem; sometimes there is not.[8] 

2.10      Professor Hansen discussed the technical difficulties in measuring background noise:

It is also important to define how background noise is measured. When you are trying to make a statement that you cannot exceed background noise by a certain amount, you need to be able to define how you measure it. For obvious reasons, there should not be a single number representing an average over many weeks or a single number as a function of wind turbines. Background noise is much lower late at night, in the early hours of the morning and also in cases when you have significant wind shear and there is no wind at the residence where the noise is being experienced. So there really should be different values of background noise at different hours of the night with different wind conditions for the measurement.[9]

2.11      Pacific Hydro Australia submitted that they have conducted testing at two of their wind farms and compared that to other natural and manmade sources and found that:

...[the] levels of infrasound at the wind farms to be well below the World Health Organisation hearing threshold and significantly lower than at the beach.[10]

2.12      The current method of measuring noise is to measure dB(A).  According to evidence in the Community Affairs Report in 2011, this measure is appropriate because:

...it simulates human hearing. Dr Warwick Williams, a Senior Research Engineer at the National Acoustic Laboratories, explained that the A-weighting heavily discounts the low frequencies and the very high frequencies. A-weighting discounts infrasound as it is below the level of human hearing.[11]

2.13      Professor Salt, in constrast, was not convinced that the A-weighted measure was adequate to detect potentially harmful noise such as low frequency and infrasound:

I do believe that the sound from wind turbines is a problem. I also think that the current method of using A-weighted sounds to characterise wind turbine noise is as big a problem, because that is missing the low-frequency content that these machines generate. I agree that the legislation, at the moment, is not considering different measurements, but at some point it needs to be considered that measuring infrasound levels from these machines could be extremely important to understanding how they affect people.[12]

2.14      Professor Hansen also agreed that there is a deficiency inherent in regulating noise using A-weighting:

...all of the current regulations are written in terms of A-weighted sound level and A-weighting does not properly account for low frequency components. Some regulations apply a five dB penalty if a noise is dominated by low frequency components, but in many cases this is insufficient to properly account for the true effect of low frequency noise. [13]

2.15      The Sonus report of 2010 explained that G-weighting is the most appropriate mechanism for picking up infrasound, and also what levels of dB(G) could have adverse impacts:

Weighting networks are applied to measured sound pressure levels to adjust for certain characteristics. The A-weighting network (dB(A)) is the most common, and it is applied to simulate the human response for sound in the most common frequency range. The G-weighting has been standardised to determine the human perception and annoyance due to noise that lies within the infrasound frequency range (ISO 7196, 1995).

A common audibility threshold from the range of studies is an infrasound noise level of 85 dB(G) or greater. This is used by the Queensland Department of Environment and Resource Management's (DERM's) draft Guideline for the assessment of low frequency noise as the acceptable level of infrasound in the environment from a noise source to protect against the potential onset of annoyance and is consistent with other approaches, including the UK Department for Environment, Food and Rural Affairs (DEFRA., Leventhall, 2003).[14]

2.16      Mr Cooper said that the underestimation of noise is compounded by the fact that the attenuation rate of low frequency noise is less than that of mid or high frequency noise:

The noise for general noise in the dBA drops off at six dB per doubling of distance. Every time you double the distance, it goes down six dB. But, when you deal with low frequencies and you deal with line sources, it goes off at a lower rate. It is identified in one of Dr Chapman's reference documents. It shows that the rate of low frequency is a much lower rate than normal noise. So what happens is that the low frequency and particularly the infrasound are underestimated as you go further away from the wind farm.[15]

2.17      Mr Cooper also critiqued other conclusions of the various reports by Sonus.  Specifically Mr Cooper questioned the report carried out by Sonus into the Cape Bridgewater wind farm.  He suggested that there were various technical discrepancies or omissions in the report that related to the noise levels inside and outside of the dwelling, and that the report omitted relevant data, including that of the wind speed.[16] 

Committee View

2.18      In the committee's view the technical issues raised by Mr Cooper are best answered by Sonus.  The committee is also of the view that even if the report were found to be flawed in the terms that Mr Cooper suggests, there was still no evidence provided that would suggest that infrasound is present at harmful levels.

2.19      The committee notes that, even amongst supporters of the bill, there appeared to be no agreement about what noise measurement should be included in the bill, nor what noise thresholds should be adopted.

Recommendation 1

2.20      The committee recommends that specific noise measures, thresholds and measuring locations not be included in legislation, as there is insufficient consensus on these elements of the proposed bill.

Changes to wind turbine technology

2.21      Wind Farm technology has changed since the introduction of turbines in Australia. These changes have made a significant difference to how sound is produced and the types of noise generated.  Mr Jonathan Upson from Infigen Energy also explained that new turbines have resulted in lower levels of infrasound that those that were manufactured 15 to 20 years ago:

...downwind turbines—that is, turbines with the rotor downwind of the tower—were known for producing higher levels of infrasound. Those types of turbines have not been made in probably 15 or 20 years, so it is possible for that turbine design to have higher low-frequency and infrasound levels than the large turbines of today.[17]

2.22      Alstom Wind gave evidence to the committee that while there are noise reduction benefits this was not the primary driver for the technical innovations: 

In the early eighties turbines did not have variable speed control; they operated at fixed speed because of the size of the turbine. As the industry has developed, to reduce the cost of energy globally, one of the big differences introduced about 10 years ago is what we call variable speed pitch control. Variable speed pitch control in a turbine—and all modern turbines now employ this—means that the tip speed of the turbine can be varied through variation of the generator speed.

There has been significant technical development in the industry—nothing to do with noise generation but to increase the efficiency and reduce the cost of energy of the wind industry. As a side effect of that what we can do is we can manually reduce the tip speed during normal power production to reduce the sound power level, purely because sound power is generated by the tip speed of the turbine. So we can control that. It was a secondary effect of the technical innovation in the industry. It was not developed for noise purposes at the start.[18]

Committee View

2.23      The committee is aware that wind turbines, like most industrial sources, create noise. The argument that was presented to the committee by Dr Sarah Laurie and others was that this noise is hazardous because of its low frequency and infrasound component. 

2.24      It was also suggested that this potential to do harm is compounded because noise is routinely measured in dB(A) and not dB(G) which picks up very low frequency noise and infrasound.  If the noise is not being measured then it cannot be regulated.  The committee believes that, as part of transparency and openness, low frequency and infrasound should be measured and endorses Recommendation 1 of the Community Affairs Committee's 2011 report that 'noise standards...should include appropriate measures to calculate the impact of low frequency noise...[19]

2.25      The committee heard evidence from a number of acousticians that infrasound is produced at various levels by a variety of different natural and industrial sources.  This includes wind turbines.  The question that concerns the committee is whether wind turbines emit noise, regardless of the frequency, at levels that are likely to cause harm.  In light of the evidence received through this inquiry the committee is of the view that while infrasound is produced it is not at levels that are likely to cause harm. This is considered further in the next chapter. 

Navigation: Previous Page | Contents | Next Page