Etsu r 97 why it is wrong

BS page 2 sect 5. The original report as presented was a reasonable attempt better to understand the problems of wind induced noise but basically it was inconclusive and recommended further research. It does not provide a specification for a proven wind screen design. Until last year the report was generally unavailable so as a result until quite recently no one has questioned it. It does not. These are probably the best commercially available wind screens.

Wind speeds shown in the slide at ten metres height have been adjusted to compensate for the lower wind speeds at 1.

This graph shows that the performance of wind screens available from several manufacturers are quite similar and all produce similar levels of wind-induced noise. This adds a degree of uncertainty to the assessment process that even recent comparison studies have failed to quantify. In fact, the only studies conducted recently have been comparisons.

None have been able to quantify the actual level of data contamination. This is becoming more of an issue for wind farms as they are often located near motorways and trunk roads for ease of access for the heavy loads. Road traffic leads to highly variable levels of background noise depending upon traffic flows and wind direction. This was even the case for the flat topography site where all monitoring locations were exposed to what should have been identical conditions.

This finding has implications with regard to background noise measurements undertaken to assess the acceptable levels of specific noise radiation from wind farms. Users of any technique that sets wind farm noise levels relative to measured background noise levels should be aware of the differences between measurements.

Further work is suggested to establish whether the large measured differences are real changes in background noise level or whether they are due to the differing susceptibility of individual items of measuring equipment to wind induced noise. This slide shows a typical case where motorway noise has a major effect on background noise as any local resident knows depending upon wind direction and traffic flow:.

Without directional screening the derived background noise level will be much too high for the worst case situation of northerly to easterly winds. For obvious reasons, wind farm developers have generally resisted directional screening of background noise data whereas the evidence shows wind direction is a very important factor in background noise levels and subsequent noise nuisance.

A robust directional screening would result in two or more sets of background noise data and noise limits that would be dependant upon wind direction.

The new draft guidance is unclear regarding the requirement and lacks a process to ensure robust screening for directional effects. There are scientific problems in how the background data are processed to arrive at summary measures that can be compared with the turbine noise prediction.

There are also problems in how the background data are processed to arrive at summary measures that can be compared with the turbine noise prediction.

Note that whatever model is used to describe the measured data, this should not be extrapolated outside the range of the measured wind speed data. No physical reasoning is put forward to guide the choice of curve to be fitted but the examples given are all polynomials of order up to order 4 quartic.

In practice cubic or quartic polynomials are usually used. The method evolved when, rather than being a very large data file downloaded from an automatic recording device, each and every data point was likely to be hard won by careful hand measurement.

Autocorrelation can be understood by a simple thought experiment. In other words successive data are correlated with themselves. Yet statistical inference assumes that each case is independent or uncorrelated with the others. Any of these curves could have been used in a noise assessment and all fit the data reasonably well. These are real data accepted and used to set noise conditions at a recent wind farm inquiry.

They are typical and the summary they suggest was used in the assessment process by the planning inspectorate, an action justified by the notion that the analysis was conducted properly since it followed the ETSU guidance.

ETSU actually suggests that this approach might be used but for some reason the industry of acoustics consultants who appear for developers seem to have quietly ignored this recommendation. It has two advantages. First, no curve is fitted — we rely solely on a summary of the measured data plus a bit on the bins chosen. Second, it makes explicit the fact that at each wind speed we have a range of background that might, or might not, mask any turbine generated noise.

The curve on the right hand side has been fitted to the same data but with a constraint that as it intersects the Y axis it has zero gradient — this is a simple logical constraint that must be true.

In a real example where we have done this we find that the fit is only marginally reduced and the predictions of the noise climate are more consistent at all the receptors than you get from the standard unconstrained approach. See ref 11, for more detail. This paper has been submitted to the Institute of Acoustics and will be published in the Acoustics Bulletin May edition. We have thus far assumed that the predicted turbine noise levels at each and every receptor is reasonable.

The next series of slides show that this is simply not the case. There are three critical issues:. ETSU makes only minimal reference to wind shear.

The loA Bulletin, March [ref 5] was an attempt to clarify the treatment of wind shear, which we believe:. This slide taken from WEIS [ref 3] shows a plot of wind shear for the Winwick site in Northamptonshire over a typical 7-day period showing how highly variable wind shear can be in the short term:. The horizontal axis is time based, the numbers relate to the 10min data sets over the 7-day period.

The vertical axis is the shear exponent. The shear exponent is a simple plot of log ten minute mean wind speed vs. In effect the magnitude of the exponent provides an index of how rapidly wind speed changes with height AGL, which is the major part of the wind shear it neglects directional change that more properly should be included. Note that shear goes negative on occasions, indicating a reduction in wind speed as one ascends.

This is found to occur at very low wind speeds in stable atmospheres and may well indicate the presence of low level jet flow. Despite this apparent random behaviour, over the long term the shear profile is regarded as being very reliable and predictable year after year. This is why wind farm developers record wind data for at least a year in order to determine the wind resource for the site. The average wind shear throughout the 12 month period for this site Winwick calculated from 51, valid 10 minute data sets was found to be 0.

However, plotting the average hourly wind shear for the 12 month period shows typical diurnal variation, with a minimum value of 0. There is thus a very large daily variation by a factor of around 3.

For reference in the literature, shear exponents above 0. If you refer to Figure 6 at WEIS reproduced here you will see a breakdown of hourly shear into vertical bands based on percentage of time. Note this relates to 12 months worth of data for the Winwick wind farm site:.

Analysis of the data for this site to determine the durations for shear levels during each hourly period shows:.

We show on this slide some other sites where data has been analysed by the authors and others. Some sites show even higher levels of wind shear than is shown for Winwick. They show the percentages of time during the day when shear is within the bands shown.

Note that 0. What these numbers do not show but the slide 2 previous does is that the higher levels of shear occur mostly during the evening and night time hours so contribute to noise nuisance during these hours.

We can conclude from this data that high levels of shear where the exponent exceeds 0. The view put forward by the industry that high shear is confined to low lying flat land is simply incorrect. These concerns relate to both the NWG draft document [ref 13] and the IoA Bulletin article [ref 5] dated from which it was developed.

Unfortunately, the NWG proposed shear methodology is quite complex making it extremely difficult to understand. It is doubtful if all members of the IoA NWG fully understand it and our experience to date is that only those who have conducted detailed analysis of raw mast data are able to appreciate fully the true implications of this methodology.

It will most likely be incomprehensible to planning decision makers. In theory, correcting for wind shear either by adjusting the measured background noise or the predicted turbine noise should produce the same assessment result.

However, it is how the shear correction is implemented that generates problems, with the devil being very much in the detail. In fact we believe that only one method of adjusting the turbine noise level for wind shear can be made to work as ETSU originally intended. The 1st bullet point at annex E para 7. This is also the most logical method from the point of view of a person on the ground since wind shear affects the turbine noise not the background noise.

The noise limits produced by this method are based on the actual background noise levels at a particular receptor residence location with the actual wind speed at 10m AGL at the wind farm site as the reference point. The actual 10m wind speed at the receptor is likely to be more closely correlated with the site 10m height wind speed than to hub height wind speed.

If the limits are subsequently breached it is most likely because wind shear has caused the turbine noise to be louder that was predicted. This is the method described in WEIS and is demonstrated at the next slide.

In this case the noise limits are based on using hub height wind speed as the reference level. An average line is then taken through these data points to determine the reference background noise. As we have seen wind shear is highly variable in the short term days and weeks but over a longer period such as a year the overall shear profile is considered to be highly consistent and repeatable year after year.

If we base the noise limits on the short term shear measurements obtained during a typically 2- or 3-week background noise survey then there will be a high risk of sampling error. Additionally, the noise limits will not relate to the actual site 10m AGL wind speeds.

At the planning stage it therefore becomes very difficult to assess whether a noise breach is likely to occur. In the event of any later noise complaints, it will be very difficult to determine whether a breach of the limits has occurred and will require close cooperation by the turbine operator since hub height wind speed data are required to determine if a breach has occurred.

The correction is applied by offsetting the predicted turbine noise curve to the left for shear in excess of the standard 0. This is described in WEIS. This method can provide a limit based on the high level of wind shear that may occur at a particular site for what is considered a significant percentage of the time. The problems associated with this NWG methodology have been demonstrated by Moroney [ref 6]. This slide taken from WEIS Table 2 shows the levels of shear offset that are required for a range of wind speeds and shear exponents.

The next few slides demonstrate the problems associated with the second method being proposed by the IoA NWG. REF Moroney [ref 6] demonstrates the effect of the this second Bulletin method on the background noise data plots and it is a simplified version that is shown here. The slides shows a starting point with a cleaned up background noise curve dB A against measured 10m AGL wind speed. The blue line is a typical curve for averaged background noise as measured at rural sites obtained during a 2 to 3 week survey period and for clarity the blue data points have all been removed.

No shear correction has been applied to this line. We find this method introduces a high degree of scatter reducing the accuracy on the derived noise limit. The substantial increase in scatter is a direct result of this methodology as each is corrected for instantaneous wind shear.

The data points are effectively offset to the right. These data cover a two week period that includes some periods of relatively high wind shear, so there are many points significantly lower than the simple ETSU line of background noise versus actual metre wind speed. It may be noted that at higher wind speeds the shift of the red points from the blue points is much smaller; this is because wind shear tends to decrease at higher wind speeds.

The next step in deriving a background noise condition according to the IoA NWG methodology requires fitting a smooth polynomial to the transformed data, effectively averaging out the data.

Furthermore, at the lowest and highest wind speeds, the polynomial fits the data particularly poorly. This demonstrates that shear was very high for a significant percentage of the time and higher than apparent due to the averaging effect of the polynomial curve. Crucially, it follows that any noise condition based on this fit will be less reliable than one based on the original ETSU method.

The key problem with the NWG methodology is that the averaged derived background noise curve from the corrected data points will be too high. A secondary problem is that short survey durations will result in a high risk of collecting unrepresentative data with regard to the long term shear profile. The real issues here is not whether correction should be to the background or turbine noise data, or whether wind speeds are referenced to 10m height or hub height, what needs to be considered is how to provide reasonable noise protection for receptors in a transparent way such that planning decision makers can easily understand it and to give noise limits that in practice can be enforced by Environmental Health Officers acting for the local planning authority.

We conclude that the methodology being promoted by the IoA NWG fails to provide the promised levels of protection, opening the way to a prospect of enormously complex future legal difficulties.

The next two slides provide a summary of his argument. Wind shear is the condition where wind speed varies with height above ground level. The normal atmospheric condition is one of positive wind shear, where wind speed increases with height. Discussion of wind shear is usually confined to the problem of causing differences between wind speed at different heights and the implications for sound power generated at the turbine and background noise generated at ground level.

A completely different and if not more important aspect of wind shear is its role in determining, via refraction, the propagation path and intensity of outdoor noise. The role of wind shear in outdoor sound propagation has been well understood since the founding work of Stokes, Reynolds and Rayleigh in the 19th century, and yet full discussion of the effect of wind shear appears to be totally neglected in wind farm noise assessments. Wind shear shapes the propagation paths of outdoor noise in all directions, strongly affecting the intensity of sound at receiving locations.

Wind shear is a principal cause of noise levels being often unexpectedly enhanced at locations a long way downwind of a noise source. Motorway noise provides a useful everyday example of noise propagation. If you walk away from the motorway on the upwind side, i. This is not true of walking away on the downwind side, noise will persist at significant levels for many hundreds of meters. Were there to be a 25 mph gale blowing across the motorway, with the same wind speed at all heights above the ground, the motorway would be equally noisy at long distances upwind as downwind; this is an unexpected conclusion and not our normal experience at all.

However it would occur in such a situation of zero wind shear. It would happen because the speed of sound in air is around mph, so movement of all the air at 25mph one way or the other will have very little effect. In practice the wind is generally stronger at greater heights, which progressively changes the speed of sound with height; this in turn changes the curvature of the wave fronts and hence the direction of their propagation. Sound waves consequently bend back down to earth on the downwind side at longer distances from the source; the opposite happens on the upwind side.

This gives us our everyday experience of motorway noise being enhanced at long distances downwind. It was not validated for use with high-level noise sources under high wind shear, turbulence and high wind speeds as apply for wind turbines. The current guidance, ETSU and the NWG draft guidelines are totally deficient from a scientific perspective with respect to noise prediction. Amplitude modulation is the most important noise characteristic of wind turbines so excluding it from the IoA review of the guidelines must be of concern.

Additionally, no credibility can be given to any future report on this subject commissioned by the industry trade association RenewableUK. This will have the same credibility as studies into health effect of smoking conducted by the tobacco industry.

Turbine manufacturers will understand the problem very well but are not going to share such information. This diagram is taken from page 39 of the MAS Environmental report ref 10 dated 4 Jan that analysed the noise assessment for the New Albion wind farm. This shows the short-term variability of turbine noise at Warboys when the dB LAeq is measured at ms intervals. In this case short term background noise must be at least 3dB lower than the lowest LAeq readings that are inclusive of background noise at a level of around 33dB.

As a result we have turbine noise peaks of up to 10dB above an actual background level of 30 dB and 5. Quoting from the MAS report at para A1. This was not the worst case as it was not directly downwind but at 60 degrees angle to the nearest turbine. The modulation was of 3—5dB. It was clear and audible both outside and inside my car. There were only rarely any other sounds within the noise environment. In conclusion we see that the public have very little statutory protection from any harmful effects of EAM.

We recognised that the existence or otherwise of health effects attributable to it is controversial. The IoA NWG does not have the inputs to enable health impacts to be determined, yet their conclusion have implications for such possible effects.

Where present at the assessment location, such features are taken into account by adding 5dB to the specific noise level to obtain the rating level. Sect 8. This certainly describes wind turbine noise even without what is described as excess amplitude modulation. With turbine noise levels just below the 43 dB LA90 limit, complaints would be virtually guaranteed in quiet rural locations.

There is evidence from studies elsewhere in the world that LFN is a problem. For the first time we have agreement by consultants acting for both sides in this dispute that:.

It should be addressed beyond the present practice of showing that wind turbine levels are magnitudes below the threshold of hearing at low frequencies.

The highly significant fact here is that the noise while either not audible or barely audible is causing adverse health effects. Measuring equipment suitable for detecting LFN was used during this survey revealing high levels of LFN inside the homes. Yet again more research is needed. Sweeping possible issues under the carpet in order to allow the onshore wind industry access to virtually every site it wants will not help.

The turbines used most frequently at UK wind farms now have rotor diameters of 80 or 90m so will operate at close to the critical blade passing frequency as shown here:. The octave band data we have seen presented by developers, obtained by turbine manufacturers only goes down to 63Hz whereas the Wisconsin report is discussing frequencies down to 10Hz.

This is the first time we have clear evidence of high levels of LFN and that A weighting measurement is inappropriate. We should also add here that IEC [ref 19] at sections A2 and A3 recognises the potential for infrasound below 20Hz and LFN between 20 Hz and Hz causing annoyance even though barely audible and that noise may be underestimated if assessed using only an L AEQ value.

Noise nuisance and ill heath effects are occurring due to turbine LFN even when the noise is inaudible or barely audible. We have also seen that turbine noise prediction is not an exact science and that there are problems in the sampling and measurement of the background.

There are also uncertainties in how the background data are processed to arrive at summary measures that can be compared with the turbine noise prediction. How do these all combine to affect the security of the overall noise assessment? The responsible scientist should recognise these uncertainties, and draw attention of any planning decision makes to them.

The implications for wind farm evaluation and the planning system are not of direct scientific concern and anyhow will vary from site to site. These may well be conservative estimates. This represents a halving or doubling of loudness and should be considered against the claimed headroom for the selected receptors.

In many cases this is less than 3dB, sometimes even less than 1. It is virtually certain that a proportion of wind farms consented under ETSU will generate justified noise complaints. ETSU does not afford the protection to citizens that its originators thought it would.

You are commenting using your WordPress. You are commenting using your Google account. You are commenting using your Twitter account. You are commenting using your Facebook account. Notify me of new comments via email. Notify me of new posts via email. You would need 8oo turbines to produce the output of a coal-fired power station. Email Address:. Sign me up! Calling for a moratorium on wind farms in the Highlands and everywhere!

Skip to content. Educating our Children or Frightening them to Death! Huhne and Hendry-The Comedy Act? Share this: Share Facebook Email Print. Like this: Like Loading About Dougal Quixote Slightly mad. Always believes a cup is half full so continues to tilt at Wind Turbines and the politicians that seem to believe it is their god given right to ruin Scotland for a pot of fool's gold. This entry was posted in Wind farms. Bookmark the permalink.

Leave a Reply Cancel reply Enter your comment here Fill in your details below or click an icon to log in:. Email required Address never made public. Name required. Search for:. Click for full details. As Einstein once said,"It does not take a hundred scientists to prove me right consensus? It only takes one to prove me wrong". Click for Details. Photos The Future for the residents of Kiltarlity?

One wind turbine is a kinetic sculpture. Thousands of them are a visual pollutant. Energy is not a devolved power!