Top Acoustic Engineer – Malcolm Swinbanks – Experiences Wind Farm Infrasound Impacts, First Hand
Top Acoustic Engineer, Dr Malcolm Swinbanks has been at the forefront of investigating the impacts of infrasound and low-frequency noise for over 40 years; and has been on the wind industry’s stinky trail in Michigan since 2009.
Last month, he delivered this technically brilliant paper: “Direct Experience of Low Frequency Noise and Infrasound within a Windfarm Community” at the 6th International Meeting on Wind Turbine Noise – the conference poster is available here: M.A.Swinbanks Poster
The results and observations as to the character and nature of incessant turbine generated low-frequency noise and infrasound backs up the groundbreaking work done by Steven Cooper at Pac Hydro’s Cape Bridgewater disaster (see our post here).
In that respect, the work sits amongst fine company. However, it’s Malcolm’s own experience with turbine noise and vibration that makes his paper all the more remarkable. Here’s a few extracts that tend to knock the wind industry’s ‘nocebo’ story for six.
The author first became aware of the adverse health problems associated with infrasound many years ago in 1974, when an aero-engine manufacturer approached him to consider the problems that office personnel were experiencing close to engine test facilities. He had been conducting research into the active control of sound, and the question was posed as to whether active sound control could be used to address this problem. At that time, this research was in its infancy, and the scale of the problem clearly lay outside practical implementation. Five years later, however, the author was asked to address a related problem associated with the low-frequency noise of a 15,000SHP ground-based gas-turbine compressor installation, having a 40 foot high, 10 foot diameter exhaust stack.
This problem was of a more tractable scale, and the author and his colleagues successfully reduced the low-frequency noise of the installation by over 12dB. He subsequently was requested to address a similar installation of significantly greater size and power, again with accurately predicted results.
As a consequence of this and subsequent work, the author has gained considerable experience of the disturbing effects of low-frequency noise and infrasound. So when he first became aware of the nature of adverse health reports from windfarm residents, they were immediately recognisable as effects with which he had been familiar for as many as 35 years.
Since late 2009, the author has lived part-time within a Michigan community where windturbines have been increasingly deployed. Consequently he has had significant interaction with residents whose lives and well-being have been damaged, and moreover has experienced the associated very severe effects directly, at first hand. His resultant perspective is thus based on both detailed theoretical analysis, and extensive personal, practical experience.
In the latter part of 2009, the intention was announced to install up to 2,800 wind turbines in Huron County, Michigan, together with adjacent regions of the Thumb of Michigan. The agricultural areas of the county are made up of 1 square mile sections, bounded by a grid of roads running north-south and east-west. The proposed wind-turbine density would amount to approximately 2-3 turbines per square mile, but in each square mile there can be typically 4 to 6 residences, usually located around the perimeter. Consequently, the requirement for adequate turbine separation would very substantially restrict the possible setbacks from residences. At that time, there existed two recently commissioned windfarms in Huron county, at Elkton (32 Vestas 80m diameter V80 turbines) and Ubly (46 GE 1.5MW 77m diameter turbines). The Elkton windfarm is in unobstructed open country, but the Ubly windfarm is in an area with significant clusters of trees, which in certain wind directions could obstruct and disrupt the low-level airflow to the turbines.
Following this announcement, the author attended an Open Meeting of the Michigan Public Services Commission, at which a number of residents spoke of the problems that they were already encountering from the windfarms, in particular the windfarm at Ubly.
This author immediately recognized these problems as relating to the characteristics of low-frequency noise and infrasound, with which he had been familiar for many years. But on subsequently visiting the windfarms, it became clear that the higher frequency audible noise levels were also unacceptable, at Ubly in particular, with up to 50dBA L10 being permitted by the ordinances. The author was astonished that any professional acoustician could possibly regard the levels as acceptable.
Following the county’s early experience the ordinances were reconsidered, so that the existing setbacks of 1000 feet, and levels of 50dBA L10, were changed for non-participating landowners to 1320 feet and 45dBA L10. But problems at Ubly were still apparent even at 1500 feet and 45dBA.
The author obtained data from one such residence, which was immediately downwind of 6 turbines located approximately in a line at distances of 1500 feet to 1.25 miles, and found that there could be significant impulsive infrasound present, even though these turbines were of modern, upwind rotor design. Under some circumstances this infrasound took the form of single pulses per blade passing interval, presumably from the nearest turbine, but sometimes up to 6 separate impulses could be detected from the turbine array.
The commissioning of further wind-turbine developments was initially hampered by the lack of high capacity transmission lines, but more recently a 5GW high voltage transmission line has been routed through the county, permitting more than adequate capacity for any intended number of windfarms and turbines. Several further windfarms, with larger 100m and even 114m diameter turbines up to 500 feet in height have now been constructed, resulting in a total of more than 320 wind-turbines installed to date.
Recently, the county has turned to reconsidering the ordinances, but as of the present date has not finalized any changes. Currently permitted wind turbine sound levels and setbacks appear to be dictated primarily by an over-riding incentive to install the requisite number of turbines per square mile.
The author has attended and commented at many public meetings, but has found that the reluctance to acknowledge adverse effects associated with low frequency and infrasound, has resulted in a situation where little traction can be gained.
Several aspects deriving from his first-hand experience will now be described in the following sections.
During the early 1980’s while working on an industrial gas turbine compressor, the author became very aware that the very low-frequency sound can quickly become imperceptible when outside in any moderate breeze. More recently, while attempting to sleep in a house 3 miles from the nearest wind-turbine of a new wind farm consisting of 35 GE 1.6 100m diameter wind turbines, the author and his wife have sometimes been kept awake by the lowfrequency rumble or infrasonic “silent thump” of the turbines.
This situation can occur when the wind has veered from a cold north wind from Canada, to a warm wind from the south blowing over cold ground. Such conditions give rise to a classic temperature inversion, and the resultant wind turbine infrasound can readily propagate for 3 miles or more.
On such occasions, the author has more than once donned outdoor clothes at 1am and gone out onto the road outside the house, clear of trees and obstructions, but in the airflow of an outside wind has been consistently unable to detect any similar subjective disturbance.
It is often argued that infrasound is more readily detectable within a residence simply because the building structure greatly attenuates the higher frequencies, but has little effect on the lower frequencies. There is an additional effect, however, that tends to be overlooked. Outside, individual ears effectively represent unshrouded pointwise microphones, equally sensitive to the full effects of airflow and true infrasound. In contrast, the conditions within a building are very different.
Pressure due to wind turbulence tends to be only locally correlated over the outside surface of the building, whereas true infrasound acts coherently over the entire structure. This gives rise to an additional spatial filtering effect, whereby the wind induced pressure distribution tends to cancel itself out, but the fully coherent very low frequency wind-turbine infrasound acts to fully reinforce itself over the entire structure.
This characteristic has been exploited for many years in the design of conformal sonar arrays – distributed pressure sensing surfaces which preferentially detect acoustic signals that are fully coherent over the surface, yet “average-out” the uncorrelated pressures due to hydrodynamic flow, yielding a significant improvement in signal-to-noise ratio.
A direct consequence of this difference between inside and outside observation is that observers visiting windfarms in the open air may quite correctly comment that they cannot hear any significant low-frequency sound. Put simply, they are not observing under the appropriate conditions. Perception within a residence, particularly in a quiet bedroom, can be entirely different.
This difference is significantly enhanced by the fact that the threshold of hearing is not a constant threshold, but is automatically raised or lowered according to the background ambient sound conditions. It is for this reason that people in urban areas, with typical ambient sound levels around 55dBA, have a naturally raised threshold and are able to tolerate additional noise of comparable level, yet this same level of noise would be completely intolerable in rural areas where ambient levels can be very much lower, not infrequently in the region of 25-30dBA.
This is one of the most important effects with respect to perception of low-frequency noise and infrasound, yet the widely cited AWEA/CANWEA Expert Health Report of 2009 (3), completely failed to indicate the consequences of this process of automatic threshold adjustment.
First Hand Experience of the Severe Adverse Effects of Infrasound.
Approximately 18 months ago, the author was asked by a family living near the Ubly windturbines to help set up instrumentation and assess acoustic conditions within their basement, which is partially underground, where they hoped to encounter more tolerable sleeping conditions.
In the early evening, the author arrived at the site. It was a beautiful evening, with very little wind at ground level, but the turbines were operating. Within the house, however, it was impossible to hear any noise from the turbines and it became necessary to go outside from time-to-time to confirm that they were indeed running.
The author did not expect to obtain any significant measurements under these conditions, but nevertheless proceeded to help set up instrumentation in the form of a B&K 4193-L-004 infrasonic microphone and several Infiltek microbarometers. Calibration of the microbarometers had previously been confirmed by performing background infrasonic measurements directly side-by-side with the precision B&K microphone. The intention was to define measurement locations, to establish instrumentation gains having appropriate headroom, and to agree and go through practice procedures so that the occupants could conduct further measurements themselves.
After a period of about one hour, which time had been spent setting up instrumentation in the basement and using a laptop computer in the kitchen, the author began to feel a significant sense of lethargy. As further time passed this progressed to difficulty in concentration accompanied by nausea, so that around the 3 hour mark, he was feeling distinctly unwell.
He thought back over the day, to remember what food he had eaten and whether he might have undertaken any other action that might bring about this effect. He had light meals of cereal for breakfast and salad for lunch, so it seemed unlikely that either could have been responsible. Meanwhile, the sun was going down leaving a beautiful orange-pink glow in the sky, while ground windspeed levels remained almost zero and the evening conditions could not have been more tranquil and pleasant.
It was only after about 3.5 hours that it suddenly struck home that these symptoms were being brought about by the wind-turbines. Since there was no audible sound, and the infrasound levels appeared to be sufficiently low that the author considered them to be of little consequence, he had not hitherto given any thought to this possibility.
As further time passed, the effects increasingly worsened, so that by 5 hours he felt extremely ill. It was quite uncanny to be trying to concentrate on a computer in a very solid, completely stationary kitchen, surrounded by solid oak cabinets, with granite counter tops and a cast-iron sink, while feeling almost exactly the same symptoms as being seasick in a rough sea.
Finally, after 5 hours it was considered that enough trial runs had been taken and analysed that it was decided to set up for a long overnight run, leaving the instrumentation under the control of the home owners. The author was immensely relieved finally to be leaving the premises and able to make his way home clear of the wind turbines.
But it was by no means over. Upon getting into the car and driving out of the gateway, the author found that his balance and co-ordination were completely compromised, so that he was consistently oversteering, and the front of the car seemed to sway around like a boat at sea. It became very difficult to judge speed and distance, so that it was necessary to drive extremely slowly and with great caution.
Arriving home 40 minutes later, his wife observed immediately that he was unwell – apparently his face was completely ashen. It was a total of 5 hours after leaving the site before the symptoms finally abated.
It is often argued that such effects associated with wind turbines are due to stress or annoyance brought about by the relentless noise, but on this occasion there was no audible noise at all within the house. Moreover, it was a remarkably tranquil evening with a very impressive sunset, so any thought that problems could arise from the turbines was completely absent.
It was only once the symptoms became increasingly severe that the author finally made the connection, having first considered and ruled out any other possibilities. So explanations of “nocebo effect” would hardly appear to be appropriate, when such awareness occurred only well into the event.
In the following two figures, the typical measured infrasound levels in the basement are shown, as measured with one of the Infiltek microbarometers.
Figure 8 shows the power spectrum, measured with a nominal 0.1Hz FFT bandwidth. As can be seen, the peak of the fundamental blade rate component, at 55dB, might not normally be considered to represent a particularly obtrusive level of infrasound. Several higher harmonics of progressively reducing amplitude are visible, but this characteristic is very much as one would expect for an upwind-rotor turbine operating in comparatively smooth airflow.
The corresponding time-trace is shown in Figure 9. It can be seen that there is a single comparatively sharply defined pulse per blade-passage, so it would appear that only the closest wind-turbine is contributing significantly.
Nevertheless, it should be noted that while the fundamental harmonic of blade-passage is at only 55dB, the cumulative effect of the higher harmonics can raise the peak level of the waveform on occasion to 69-72dB. Most of the author’s prior work has concentrated on time-history analysis of the waveform, consistent with the 2004 observation by Moller & Pedersen (4) that at the very lowest frequencies it is the time-history of infrasound which is most relevant to perception. Simply observing separate spectral levels at discrete frequencies and regarding these as independent components can lead to considerable underestimate of the true levels of repetitive infrasound.
The fact that balance and coordination were found to be adversely compromised during the night drive home would suggest interference with the vestibular organs, as proposed by Pierpont (5) and subsequently by Schomer (6).
An important additional observation, however, is that the effects persisted for 5 hours afterwards, when the immediate excitation was no longer present. In contrast, for sea-sickness, effects tend to dissipate rapidly once sea conditions moderate. It is of interest that a 1984 investigation (7), in which test subjects experienced 30 minutes exposure to 8Hz excitation at very much higher levels of 130dB, reported that some adverse effects could persist for several hours later.
It has been shown that upwind-rotor turbines can indeed sometimes give rise to impulsive low-frequency infrasound – a characteristic commonly attributed only to old-fashioned downwind rotor configurations. But perception of wind turbine low frequency noise and infrasound can be quickly suppressed by the effects of wind-induced airflow over the ears, with the result that incorrect conclusions can easily result from observations made when exposed to outside breezy conditions.
The effects within a residence are much more readily perceptible, and cannot be ignored. An account has been given of an occurrence of severe direct health effects experienced by the author, and considered to be due entirely to wind-turbine infrasound, yet manifest under superficially benign conditions where no such adverse effects were anticipated.
23 April 2015