4Feb/100
Wind Turbine Noise Explained
From:
MA.DR SWINBANKS
To:
MPSCEDOCKETS;
CC:
MA.DR SWINBANKS;
Subject:
Re: Case No U-15899
Date:
Wednesday, December 09, 2009 10:09:47 AM
Attachments: MAS Research Ltd Michigan Windmill Letter.doc
Case No U-15899
Dear Sir,
I am enclosing a Word document relating to comments on Windmill
Setbacks & Noise, which represents a letter that I have sent to you. I
posted this letter in Port Hope yesterday, and hope that it arrives in
Lansing before your Friday 5pm deadline. But I am also enclosing an
electronic copy with this email. I hope this is satisfactory.
Sincerely,
M.A.Swinbanks
MAS Research Ltd
Mathematical & Scientific Research
Tel: 011-44-1-223-512250
Company Number 1586916
8 Pentlands Court
Incorporated 1981
Pentlands Close,
Cambridge CB4 1JN,
Executive Secretary, MPSC,
United Kingdom
December 8th, 2009
P.O.Box 30221,
Lansing,
Michigan 48909
Re: Case No U-15899
Dear
I am a professional consultant engineer, and my company is based in the United
Kingdom, but fourteen years ago I was asked to come to the US to lead an advanced
research project for the Office of Naval Research. My American wife & I now live at
7087 Kinde Road, Port Hope, Michigan. During the course of my career, I became a
consultant to many different companies and research organizations on a wide variety of
problems related to unsteady dynamics, noise, vibration, shock and acoustics.
I have worked personally with both Professor J.E.Ffowcs-Williams, and Dr
H.G.Leventhall, two of the foremost UK acousticians. 20-30 years ago, I worked
directly in collaboration with both on several low-frequency noise installations, thus
gaining first-hand experience of the problems associated with low-frequency noise and
infrasound. My actual time-on-site addressing low-frequency noise probably well-
exceeds either. [ 1], [2 ].
This letter addresses three separate issues relating to wind-turbines. First, an unresolved
issue relating to low-frequency sound generation by wind-turbines. Second, further
well-established characteristics of low-frequency noise. Third, the present status of
permitted noise levels and setbacks.
(1) Low-Frequency Sound Generation by Wind-Turbines
The opinions of the two UK acousticians relating to wind-turbine noise differ.
Professor Ffowcs-Williams has stated “It is known that modern, very tall turbines,
do cause problems, and many think the current guidelines fail adequately to
protect the public.”
while Dr Geoff Leventhall has commented "I can state quite categorically that
there is no significant infrasound from current designs of wind turbines.
• Infrasound is not a problem, • Low frequency noise may be audible under
certain conditions, • The regular 'swish' is not low frequency noise.”
1
In practice, the transition from infrasound to low-frequency sound may be blurred.
Based on my own experience, the consistent reports of physical discomfort resulting from
wind-turbine noise reinforce my perception that low-frequency noise can indeed be a
problem. The reported effects are entirely consistent with those that I have experienced
at first hand, 20-30 years ago.
Low frequency noise can induce feelings of discomfort and nausea, not unlike
seasickness. Like seasickness, the sensitivity of different individuals varies enormously,
some being immediately sensitive, while others can barely detect anything. I have stood
beside two people on a site where low-frequency noise was present. One person said “I
can’t really hear anything”. The other said “I feel ill – I should like to leave”. Both
were reporting accurately; there can often be more than 12dB difference ( a factor of 4)
in the sensitivity of individuals to low-frequency noise. Given that for very low
frequencies, 12dB represents the difference between just audible, and uncomfortably
loud, it is clear that very real problems are experienced by some individuals, while others
remain largely unaffected.
It is important to emphasize that there does not yet appear to be a full understanding of
how to assess low-frequency wind-turbine noise. As recently as April 2008, A Danish
researcher, T.H.Pedersen demonstrates clearly in [3] how different conventions for
measuring the noise field of a turbine can lead to diametrically opposite conclusions. He
summarizes by writing “The above mentioned issue has been discussed with a number of
researchers (Henrik Moller, Aaborg University, Torsten Dau, Ranish Technical
University, Hugo Fastl and Geoff Leventhall) and solutions have been sought for without
result.” He goes on to describe a procedure involving weighting the spectra with the
inverse hearing-threshold (HT-weighting) but while clarifying the problem, this does
nothing to resolve the issue.
So it is difficult to understand how it can be argued emphatically that there is no problem,
when it is clearly reported that significant ambiguity still remains in assessing these
effects.
The present author has considered this aspect, and believes that the misunderstanding
may lie in a failure to take into account correctly the impulsive nature of the turbine
noise, as each blade passes the tower, and interaction takes place between the blade, the
wake, and the tower. Although it is now widely recognized that this can give rise to low-
frequency modulation of higher frequency aerodynamic noise, resulting in a “swishing
sound” (aerodynamic modulation), it remains the case that the low-frequency effects of
the impulse are often incorrectly analyzed. This latter effect has been described as a
distinct repetitive “thumping sound” audible at distances of 500 to 1000 meters (~ 1600
to 3300 ft.)
The feature of impulsive noise is that there is a large signal present for a short period of
time. Consequently, the mean, or root-mean-square (rms) level of the signal may be
very low, apparently well below the threshold of hearing, but the peak level is much
higher and can be perceived. This ratio of peak-to- mean level is the Crest-Factor.
2
The present convention of combining frequency-weighted spectral or octave levels only
measures the rms level – it does not take any account of the crest-factor.
The hearing threshold has been determined experimentally using individual sinusoidal
sound waves. But sinusoidal waves have the lowest of all crest factors. C.S.Pedersen
[4] has reported that band-limited 2Hz-20Hz, and 2Hz-40Hz white noise is audible 7-
10dB below the threshold defined for sinusoidal signals. This observation is consistent
with the increased crest-factor of such noise. But low-frequency, repetitive impulsive
sounds possessing a multiplicity of harmonic components, have an even more
recognizable characteristic, and are likely to be audible at even lower levels. Preliminary
calculations indicate that periodic 1Hz impulses may be audible even when the individual
components of spectral lines lie 25dB below the threshold of hearing. So simply
examining low-frequency spectra and observing that individual spectral lines lie well
below the threshold of hearing does not begin to summarize this situation accurately.
A further comment relates to this impulsive component of noise. If an observer stands
near to the wind-turbine, the distance from him to different portions of the tower and
blade varies significantly. Consequently, the time taken for sound to propagate to this
observer differs for each portion of the blade segment. As a result, the arrival times of
the impulsive effects are “smeared-out”, and much less audible, despite the close-up
distance. But for an observer positioned several hundred feet away, along the line of the
axis of the turbine, the impulsive components all tend to arrive at the same time, giving a
much enhanced effect.
(2) Additional Well-Established Effects of Low-Frequency Noise
Two further effects relate directly to the annoyance of low-frequency noise. The
hearing threshold of individuals does not remain fixed at a constant level, but rises or
falls according to the background sound level. In addition, there is an acquired learning
process, where a person can become much more sensitive to a specific low-frequency
sound after repeated exposure. Unfortunately, the people most likely to become ultra-
sensitive to low-frequency wind-turbine sound are precisely those people who live closest
to the unwanted source.
The variation of hearing threshold according to background noise has an important
consequence. People are often invited to visit wind-turbine sites during daytime, when
ambient levels are high, and they conclude that the turbine noise levels are not excessive.
But under these circumstances, their own threshold of hearing is raised, so the extent to
which the turbine noise protrudes above their threshold is minimal. But at night, in the
quiet of an interior living room or bedroom, the ambient level is lower, their hearing
threshold drops accordingly, and the wind-turbine noise can rapidly become intrusive or
intolerable.
Indeed, subsequent attempts to shut out the sound, by closing doors and hiding under
pillows and bedclothes, have exactly the opposite effect. The higher-frequency
3
background ambient levels are reduced still further, while the remaining component, the
penetrating low-frequency turbine noise, can become even more dominant.
Several additional physical effects can cause the low-frequency sound levels of wind-
turbines to rise above conventional expectation. G.P.Van den Berg [5] has reported that
variations in wind-gradient at night can cause wind levels at the turbine hub height to be
considerably greater than wind speeds near ground level, thus giving rise to a more
rapidly changing wind profile and underestimates of true wind speed. He reported
increased sound levels of 15dB as a consequence.
In addition, the present author is familiar with reduced-temperature night-time conditions
where low-frequency sound from a gas-turbine installation could be audible at distances
of 1-mile (5280 ft), given appropriate atmospheric conditions, possibly associated with a
temperature inversion. Calculation would have predicted that the gas-turbine noise
should have been inaudible at approximately 400 yards (1200ft). By implication, the
attenuation with distance was very much less than expected, apparently by an amount
corresponding to over 12dB.
Finally, it should be noted that operation of several wind-turbines together, near-
synchronized, gives rise to additional modulation of sound intensity which itself can be
very disturbing, and yields higher than predicted sound levels . Van den Berg has
reported this effect, and measured rising and falling intensities corresponding to the
effects of the turbine noise sources moving into and out of phase.
Few, if any, of these directly relevant effects are taken into account in the present
assessments of the low-frequency noise associated with wind-turbine farms. Yet they
directly impact the quality of life for individuals and families living close to wind-farms.
(3) Setbacks & Noise Criteria for Wind-Turbines.
It should be noted that UK criteria have been guided by a 1997 recommendation, ETSU-
R-97 which has advocated night-time levels not exceeding 43dBA or 5dBA above
background levels for external noise-levels at habitations. Specific setbacks are
calculated according to the individual performance data and geometry of proposed wind-
turbine configurations, but in general, have tended to underestimate the actual sound-
levels that subsequently are manifest in practice.
These criteria have been consistently questioned for 12 years since 1997, and there have
been repeated requests to revise the criteria in the light of actual experience. Professor
J.E.Ffowcs-Williams has stated
"Van den Berg's paper adds weight to the criticisms frequently offered of UK
regulations covering wind turbine noise, ETSU-R-97. The regulations are
dated and in other ways inadequate. It is known that modern, very tall
turbines, do cause problems, and many think the current guidelines fail
adequately to protect the public……. It really is time for the DTI (Dept of
4
Trade & Industry) to clear the air on this one, and institute a comprehensive
and fully transparent study, obtaining data from the United States and
Europe, as well as the United Kingdom."
Given that the UK night-time levels of 43dBA are now proving to be inadequate in
practice, it is clear that proposed Michigan levels of 55dBA, corresponding to sound
pressure levels 4 times higher, and 1000 ft setbacks would likely represent intolerable
levels for many members of the community.
Moreover, the convention of using the A-weighted decibel scale has itself been
questioned, since this specifically filters out and minimizes the effects of low-frequency
noise. The flatter C-weighting scale has been suggested as a more appropriate
alternative. This issue relates back to the author’s earlier comments about the lack of
rigour in defining the low-frequency impulsive effect.
In respect of actual measured levels for a windfarm, the paper by Van den Berg is very
relevant. He measured sound levels adjacent to a windfarm consisting of seventeen
1.8MW wind-turbines. In particular, he derived a continuous record of dBA levels taken
at 50 millisecond intervals, which showed modulating peak levels of 51-53dBA recorded
on the terrace of a house 750m ( ~ 2500ft) from the windfarm. Projecting these levels
back to 1000ft, would imply peak levels 8dB higher at 59-61dBA.
Van den Berg [5] described the situation as follows: “However, on quiet nights the
wind park can be heard at distances of up to several kilometres when the
turbines rotate at high speed. On these nights, certainly at distances
between 500 and 1000m ( ~ 1600 and 3300 ft) from the wind park, one can
hear a low pitched thumping sound with a repetition rate of about once a
second (coinciding with the frequency of blades passing a turbine mast), not
unlike distant pile driving, superimposed on a constant broadband ‘noisy’
sound. A resident living at 1.5km (~ 4900 ft) from the wind park describes
the sound as ‘an endless train’. “
In conclusion, it is well-reported that the close proximity of wind-turbines to residences
can cause very real annoyance and distress. Experience in practice has consistently
shown that present guidelines for setbacks are proving to be inadequate. There is no
fully agreed method of defining accurately the low-frequency noise effects, largely
because these can vary markedly according to circumstance, wind gradients, atmospheric
conditions, and personal susceptibility. Consequently, it is important to be guided by
lessons learned from experience.
Yours Sincerely,
Malcolm A. Swinbanks, M.A., PhD
References /(over)
5
References
[1] Swinbanks M.A. The Active Control of Low Frequency Sound in a Gas
Turbine Compressor Installation, Inter-Noise '82, May 17-19, 1982
[2] Swinbanks, M. A. The Active Control of Noise and Vibration and some
Applications in Industry. Proc. IMechE 198A, No. 13, pp. 281–288, 1984.
[3] Pedersen T.H. Low Frequency Noise from Large Wind Turbines
A Procedure for Evaluation of the Audibility for low Frequency Sound, and a
Literature Study. Delta Acoustics & Electronics. AV 1098/08 30 April 2008
Client: Danish Energy Authority
[4] Moller H & Pedersen C.S. Hearing at Low & Infrasonic Frequencies, Noise &
Health, Volume 6, Issue 23, April-June 2004
[5] G.P. van den Berg Effects of the wind profile at night on wind turbine sound.
Journal of Sound and Vibration, Volume 277, Issue 4-5, p. 955-970, 2004
