Sleep Disturbance and Industrial Wind: Update with Stephen Cooper

By Sherri Lange
March 18, 2021

“It would appear in our recent work that the major level of disturbance occurs during the change in the power output of the windfarm and that the percentage of the change is significantly less than that presented in Cape Bridgewater.”

“In Australia, residents can plot the output of an individual windfarm and identify the change in outputs that may be giving rise to the disturbance they are experiencing, which was shown in the Cape Bridgewater study.”

Master Resource has tracked the ongoing research of acoustical engineer Stephen Cooper into vibrations and infrasound (low frequency noise) from industrial wind turbines on nearby residents.* Cooper has been part of the wind power debate since his 2014 pathbreaking study of the Cape Bridgewater Wind Farm in southwest Victoria in Australia (also see here).

Mr. Cooper continues to explore his original hypothesis that the disturbance from wind farms relates to changes in the electrical power output of the wind farm, a correlation endorsed by fellow experts Paul Schomer and George Hessler.

Recently a community group in Australia “Windfarm Living” posted a methodology that residents can use to determine the level of sleep disturbance from the operation of a wind farm that was developed by Cooper and has been proven to be effective for identifying the extent of their experienced sleep disturbance.

Windfarm Living has encouraged residents in areas in which there are proposed wind farms to undertake the exercise to provide baseline data with respect to their sleep pattern for the existing acoustic environment of the area. We understand at the moment that for one area in Victoria that is proposed to be impacted by a windfarm there are more than 20 houses currently undertaking their own noise testing that can be coupled with the Cooper methodology.

Sherri Lange caught up with Steven Cooper to understand more about his methodology, given his national and international audience.

Q. Sherri Lange: Since your peer reviewed paper in Acoustics, discussed at MasterResource last year, you have presented three papers at the e–forum Acusticum 2000. Could you tell us how this latest investigation fits in with your original Cape Bridgewater study?

A. Steven Cooper: In Cape Bridgewater, the specific brief was to identify various sound levels and wind conditions that gave rise to disturbance to the six local residents involved in the study. I introduced the concept of sensation in addition to evaluating noise and vibration.

Sensation was described by the residents as something they detected, felt, or perceived, but not necessarily attributed to noise or vibration. Sensation became the major basis of disturbance.

From the analysis of the data and the diaries of the six residents, there was a pattern that showed a higher level of disturbance at four specific times: when wind turbines commenced to start operations, when the windfarm had reached its maximum output for high wind speeds to the point that they would be de-powering the blades, and when the windfarm was changing its power output by 10% in either direction.

Q. Yes, I see that in June 2017 Paul Schomer presented a paper to the wind turbine working group (of the Acoustical Society of America) that the change in electrical power output of a wind farm appeared to be a cause of disturbance to residents from his review of your Cape Bridgewater study and his assessment of the Shirley windfarm study. I recall that in effect you had undertaken an exercise which was suggested in the conclusion for the Shirley windfarm report that needed to be done.

A. Yes

Q. Then I see that in December 2017, you and Mary Morris reviewed the diaries of the residents involved in a study for the Waterloo windfarm and you presented correlation of the residents’ disturbance with the electrical power output of the windfarm, which in honor of Paul Schomer you called it the “Schomer effect”.

A. Correct. The EPA Waterloo study (which preceded my Cape Bridgewater study) only considered in their analysis complaints from residents that were related to noise disturbance. However, limiting the complaints to only noise disturbance eliminated the majority of the complaints that when plotted versus the power output tied in with my Cape Bridgewater hypothesis.

It is on that basis and the hypothesis that I have spent a number of years in analysing the acoustic signature of wind turbine noise and uncovered the difficulties in accurately reproducing the signal in an acoustic laboratory. There are technical issues in reproducing the signal accurately and also technical issues with respect to the acoustic analysis of that signal.

It is upon the basis of that work and finding the pulsations in the signal occur across the entire frequencies (and not just limited to the infra-sound region) that led to the pilot study of inaudible wind turbine noise which was presented to the ASA Wind Turbine Working Group in December 2017 as a separate paper separate to the Cooper Morris review.

Q. I remember that material which identified a concept previously expressed as dynamically pulsed amplitude modulation but then, as discussed in our previous Master Resource article, you have identified that in an electrical sense it is not “amplitude modulation” but should be expressed as “modulation of the amplitude”.

Now I know that the concept of using the words “amplitude modulation” incorrectly did present some reservations to acousticians at the ICA conference in Germany in late 2019.

A. Correct. But more importantly I also introduced the concept of Fluctuation that by necessity, required identification of the depth of the modulation. In explaining Fluctuation as a response to people that was sensed – not heard, by reference to the work of Prof Alex Salt. It is the outer hair cells of the inner ear that are sensing this very low frequency modulation.

Q. In our last Master Resource interview with respect to your paper in Acoustics, “Determination of Acoustic Compliance of Wind Farms” (June 2020), we touched on the issue of wake effects and that the dirty air from an up wind turbine can affect down wind turbines in terms of their operation and the generation of these pulsations. In the US, we tend to have smaller numbers of turbines in a windfarm compared to Australia.

In some cases, residents may be dominated by a single turbine (as in the case of the Shirley windfarm that therefore give rise to different patterns of acoustic signature to that of a wind farm). Is it the case that in some of the American instances the modulation of the amplitude is more profound than in Australia and particularly in light of residents being closer to individual turbines than that in Australia.

A. Yes. In a number of reports in the UK and America, it would appear that the modulation of the amplitude of the dB(A) signal can become readily apparent by reason of residents being experienced to a single turbine or in some instances, in a downward situation of a group of turbines such that one can get fluctuations in the A weighted level in the order of 5 to 10 dB(A). This would clearly give the regular thumping sound which occurs at the rate of the blade pass frequency which for three bladed turbine operating and 17 rpm would be approximately every 1.2 seconds.

Q. In your Cape Bridgewater study you suggested that your hypothesis was on the basis of the blades of the turbines not being efficiently aligned with the wind and therefore by the blades not being at the correct angle, this gave rise to significant increases in the level of the pulsations from the turbines.

A. Yes Sherri. The angle of the blade is called the “pitch angle” and can be altered by the control mechanisms in the nacelle of the turbine. The blades themselves, as you are aware, have a significant weight by reason of the physical dimensions of those blades.

If you imagine a 50 m long blade suspended up in the air on the on the end of the nacelle then the mechanism for changing the pitch of the blade would need to be substantial and cannot in the real world move backwards and forwards every few seconds. So therefore, the computer controls for the turbine requires a period of time in which the wind strength (and the wind direction) is found to be at a consistent level upon which the blade angle or the direction of the turbine is altered to suit that averaged wind condition.

Therefore, for each of the four configurations set out in the hypothesis, one can comprehend the basis of why the angles would not be aligned with or efficiently aligned with the wind strength all the time.

It would appear in our recent work that the major level of disturbance occurs during the change in the power output of the windfarm and that the percentage of the change is significantly less than that presented in Cape Bridgewater.

Q. Could you please clarify this change in power output? For example, how small is an increment?

A. If you are able to track the power output of an individual turbine or a wind farm you can see that over a period of time, let us say, for the exercise 30 minutes, the output of the windfarm can be gradually decreasing, gradually increasing, or abruptly increasing and then abruptly decreasing.

In a general sense the output of the windfarm is determined by the wind speed, then the power output graph indicates changes in the wind speed. If the windfarm is experiencing a noticeable change in its output it then follows as discussed above that you will have during those changes inefficient angles of the blade pitch.

If you have access to the turbine data turbine (SCADA) and examine that material you can have the actual power output of individual turbines versus the wind obtained on top of the nacelle, the rotor speed and the blade pitch angles. When all that material is plotted, then it becomes very easy to see the patterns and why there is a delay in the turbines being set up as appropriate for the operations.

Q. In the US we do not have public access to that turbine information and I assume in many cases the same applies to Australia. So, can you please explain how in Australia, residents can undertake the exercise you have identified?

A. In Australia all of our power utilities have to report on a continuous basis the output of the utilities to our Australian energy market authority. That is simply the output of a coal fired power station, a gas-fired power station hydro electricity,  solar or wind farms (we don’t have any nuclear power generators in Australia).

An electrical engineer Paul Miskelly and his son (Andrew) developed a program number of years ago to interrogate the power output data from which they convert that material into graphs which are published by them. Anybody when accessing the graphs can look through all the utility power outputs in Australia or focus on any individual power utility, which in in the case of concern being wind farms. Readers of STT (Stop These Things) will see many articles utilising those power output graphs which have been an immense assistance to the community in understanding instability on our power grid and by the provision of renewable energy.

In Australia, residents can plot the output of an individual windfarm and identify the change in outputs that may be giving rise to the disturbance they are experiencing, which was shown in the Cape Bridgewater study.

Q. Yes, I understand the advantage that Australians have over us in terms of tracking power output with respect to your hypothesis. But let us get onto the interesting part about how this has now been married with the output of an Apple IWatch.

A. We have a client that has been severely impacted by a windfarm. As a result of community complaints around existing windfarms our client undertook their own noise monitoring and prior to the construction of the windfarm and has continued during the windfarm operations. That represents a number of years of full-time acoustic data.

When that client having heard about the ability of Fitbit and Apple IWatch to track sleep purchased a number of Apple IWatches and was able to correlate herself the levels of disturbance by family members, both in terms of general noise levels and the wind farm power output.

That client undertook a full sleep study and where one person was wired up in the standard format for such studies and at the same time used an Apple IWatch.

The sleep study material was assessed and analysed by Sleep Australia and then reviewed by a sleep physician to find an excellent correlation between the IWatch output and the Sleep Australia results.

In our noise monitoring for that client (which occurred for five months), we utilised the resident’s diary and the IWatch output to assess that material with respect to the power output of the windfarm and also our noise monitoring.

Q. So, to be able to present the hypothesis to assist communities with the use of an IWatch you must have obtained good results to show what was happening.

A. The correlation in terms of the sleep disturbance, which shows the heart rate is plotted by the Sleepmatic program lines up in and in some cases is a mirror image of the change in the power output.

This actually ties in with the work that I presented in Germany on the brain Quantitative EEG results and the heart rate variability for the different (inaudible) noise sources that we studied in our laboratory.

Q. I understand that you are in the process of preparing a number of papers for publication in Acoustics that deal with the modulation of the amplitude and this sleep disturbance issue. But can you give us a sneak preview of some of that material as to what can be done by the process.

A. Well Sherri, we look at the variation in the power output of the windfarm by use of the SCADA information of the AEMO material and to simplify the matter we take the absolute change in the power output and produce that plot in addition to the power output of the windfarm. This absolute (delta) change plot makes the changes easier to see. We do this in 10 minute averages and depending upon the sensitivity of the individual, we actually can find that small changes in the power output can give rise to disturbance.

We then take the noise data obtained inside the buildings (and that is important, that the disturbances to the residents are inside so there is no point in using outside measurements). We then utilize the statistical modulation index approach to determine high levels in that modulation index.

The higher levels of modulation index correspond to the change in the power as described above, and then when you look at the sleep disturbance information from the resident’s the diary there is agreement.

As such we can actually take the power output and the modulation index and identify the likely times in which disturbance may occur. When we then check the predictions with the IWatch outputs then the results are amazing.

Sherri Lange: So even if the IWatch program and your power output was only 50% accurate, then, there is a huge benefit in terms of identifying a relationship with the windfarm and sleep disturbance. But when one gets into a higher level of investigation and tying that material in with the noise analysis then I understand how this tool can be a benefit to the communities impacted by wind turbine noise.

Thank you for your time.

*These posts include:

[Originally published at Master Resource]

Click here for computer translation

Click here for computer translation