assessment of the acoustic impact of the proposed ... · assessment of the acoustic impact of the...

Assessment of the Acoustic Impact of the Proposed Fjällberg Wind Farm

Author: Andrew Birchby Date: 10 December 2012 Ref: 02372-000952

This document (“Report”) has been prepared by Renewable Energy Systems Ltd (“RES”). RES shall not be deemed to make any representation regarding the accuracy, completeness, methodology, reliability or current status of any material contained in this (“Report”), nor does RES assume any liability with respect to any matter or information referred to or contained in the Report. Any person relying on the Report (“Recipient”) does so at their own risk, and neither the Recipient nor any party to whom the Recipient provides the Report or any matter or information derived from it shall have any right or claim against RES or any of its affiliated companies in respect thereof. Recipient shall treat all information in the Report as confidential.

Prepared: Andrew Birchby Signed Electronically: 10-Dec-2012

Checked: Matthew Cassidy Signed Electronically: 13-Dec-2012

Approved: Jeremy Bass Signed Electronically: 14-Dec-2012

Document Reference: 02372-000952 Issue: 02 - Approved

Revision History

Issue Date Author Nature And Location Of Change01 08 Oct 2012 Andrew Birchby First Created02 10 Dec 2012 Andrew Birchby Alternative Layout Scenario

CONTENTS

INTRODUCTION 1

1.0 GENERAL OVERVIEW OF WIND TURBINE NOISE 1

2.0 METHODOLOGY 1

2.1 Noise Emission Characteristics for the Siemens SWT-3.0-113 Wind Turbine 22.2 Locations of Wind Turbines 32.3 Locations of Nearest Neighbours 32.4 Estimation of Noise Levels at Receivers 62.4.1 Noise Propagation Model 62.4.2 Correction for Surface Roughness 62.4.3 Conservatism in Propagation Modelling 72.4.4 Predictions 72.5 Acoustic Acceptance Criteria 92.5.1 SEPA Assessment Procedure 92.6 Acoustic Assessment 102.7 Turbine Management 102.7.1 Noise Reduced Modes for the Proposed Wind Turbines 102.7.2 Proposed Turbine Management 10

3.0 OTHER ASPECTS OF ACOUSTIC NOISE 11

3.1 Low Frequency Noise 113.2 Infrasound 123.3 Vibration 133.4 Aerodynamic Modulation 143.5 Wind Turbine Syndrome 143.6 Wind Shadow Impact 153.7 Conditions Favourable to Propagation 163.7.1 Fog 16

4.0 CONCLUSIONS 16

5.0 REFERENCES 17

APPENDIX A: FIGURES 19

GLOSSARY 21


1

Assessment of the Acoustic Impact of theProposed Fjällberg Wind Farm

INTRODUCTION

This report contains an assessment of the acoustic impact of the proposed Fjällberg wind farm according to guidelines issued by the Swedish Environmental Protection Agency (SEPA, 1978).

1.0 GENERAL OVERVIEW OF WIND TURBINE NOISE

Noise can have an effect on the environment and on the quality of life enjoyed by individuals and communities. The effect of noise can therefore be a material consideration in the determination of planning applications.

There are two quite distinct types of noise source within a wind turbine. The mechanical noise produced by the gearbox, generator and other parts of the drive train; and the aerodynamic noise produced by the passage of the blades through the air.

Since the early 1990’s there has been a significant reduction in the mechanical noise generated by wind turbines, and it is now usually less than, or of a similar level to, the aerodynamic noise.

Aerodynamic noise from wind turbines is generally unobtrusive; it is broad band in nature and, in this respect, is similar to, for example, the noise of wind in trees.

Aerodynamic noise is usually only perceived when the wind speeds are fairly low. In higher winds, it is generally masked by the normal sound of wind blowing through trees and around buildings. There has been rapid progress in reducing both aerodynamic and mechanical noise from wind turbines, and new designs are much quieter than those of a few years ago.

Over recent years, many wind farms have been constructed within Sweden and worldwide, and a better understanding has been gained into what constitutes an acceptable level of noise from wind farms. In Sweden, the relevant methodology for assessing the impact of noise from wind farms is described in Swedish Environmental Protection Agency report ‘78:5’ (SEPA, 1978). This provides a robust basis for assessing the noise impact of a wind farm and has therefore been used in this assessment.

Generally any noise restrictions placed on a wind farm must balance the environmental impact of the wind farm against the national and global benefits that would arise through the development of renewable energy sources, and not be so severe that wind farm development is unduly stifled.

2.0 METHODOLOGY

This acoustic impact assessment has been prepared by RES for a proposed layout consisting of 60 turbines.

In accordance with the guidelines issued by the Swedish Environmental Protection Agency (SEPA, 1978) the acceptance of the proposed wind farm is established by comparing the noise levels produced by the operation of the proposed wind turbines with an appropriate noise limit at nearby residential properties.

To make this assessment, the following steps have been taken, as detailed in this report:

determination of the noise emission characteristics of the wind turbines;

determination of the locations of the wind turbines;

determination of the locations of the nearest, or most noise sensitive, neighbours;


2

estimation of noise levels at the nearest neighbours due to the operation of the wind farm, using a sound propagation model;

determination of the acoustic assessment criteria in light of relevant guidance or regulations; and

evaluation of the acoustic assessment by comparing the estimated noise levels with the noise assessment criteria

Technical terms are explained in the glossary and all figures are included in Appendix A.

2.1 Noise Emission Characteristics for the Siemens SWT-3.0-113 Wind Turbine

Although not finalised, the turbine type for the proposed Fjällberg wind farm is likely to be acoustically similar to the Siemens SWT-3.0-113 machine. In this assessment all turbines are considered to be operating in their standard noise setting (the noisiest).

The acoustic data used in this analysis, detailed in Table 1, is as provided by the turbine manufacturer and represents the sound power level that would be warranted (Siemens, 2012). It has been assumed that tonal emission characteristics are such that no clearly audible tones are present at any wind speed1.

Table 1 Sound Power Levels for the Siemens SWT-3.0-113 Wind Turbine

Standardised 10m Height Wind Speed, v10 (ms-1)

A-Weighted Sound Power Level / dB(A) re 1 pW

4 97.0

5 103.2

6 106.3

7 107.0

8 107.0

9 107.0

10 107.0

11 107.0

12 107.0

Octave band data at the maximum sound power level that this turbine generates when operating in its standard noise setting is presented in Table 2 below. See section 2.4.2 for further information on the turbine values adopted.

1 RES standard practice is that no clearly audible turbine tonal component is acceptable. Before this turbine type could be employed at Fjällberg, RES would seek to obtain a warranty from the manufacturer.


3

Table 2 Octave Band Data for the Siemens SWT-3.0-113 Wind Turbine

Frequency (Hz)

A-Weighted Sound Power Level / dB(A) re 1 pW at 8m/s

63 86.9

125 94.8

250 100.2

500 101.1

1000 100.7

2000 99.7

4000 92.6

Total 107.0

2.2 Locations of Wind Turbines

The proposed Fjällberg wind farm is located near to Långbäcken, in the Lycksele and Åsele Kommunes, Västerbotten Län, Sweden at Swedish grid reference (1601000, 7149000)2. The turbine layout considered in this assessment consists of 60 turbines as shown in Figure 1 (Appendix A).

2.3 Locations of Nearest Neighbours

The locations of the nearest neighbours to the turbines have been determined by a house survey. In total, 127 buildings have been identified and the distances from each property to the nearest turbine are given in Table 3. Of these, 16 buildings have been identified as unoccupied and are highlighted in grey in Table 3. The remaining 111 properties are shown in Figure 1 (Appendix A). The minimum house–to–turbine separation for the properties considered in this assessment is 871m to house K11.

Table 3 Buildings near Wind Farm and Distance to Nearest Turbine

Buildings identified as, or pending agreement to become, unoccupied are highlighted in grey

House ID

Grid Co-ordinates Distance / m

Nearest TurbineX Y

K01 1610505 7148876 2856 E1K02 1610330 7148930 2811 E1K03 1610213 7149046 2858 E1K04 1610213 7149093 2900 E1K05 1609878 7149140 2812 E1K06 1608114 7147962 1688 E1K07 1605173 7150911 1047 A30K08 1604791 7150608 1214 A27K09 1604735 7150680 1124 A27K10 1602689 7150088 1642 A25K11 1600819 7152830 871 A24K12 1600503 7152931 1197 A24K13 1600494 7152913 1203 A24K14 1600275 7152774 1360 A22K15 1592846 7149519 1806 C13K16 1592671 7149962 1332 C13K17 1592481 7150118 1111 C13K18 1592430 7149766 1430 C13K19 1594201 7150628 1749 C9K20 1594263 7150533 1861 C9K21 1594268 7150570 1832 C9

2 Swedish National Grid co-ordinates, RT90 2.5 gon V datum


4

House ID


Nearest TurbineX Y

K22 1594355 7150740 1744 C9K23 1594332 7150720 1746 C9K24 1597051 7143426 3303 B20K25 1590111 7147598 3801 C12H1 1602050 7153590 1048 A24H2 1602920 7155520 3143 A24H3 1602990 7155520 3172 A24H4 1603010 7155470 3136 A24H5 1603010 7155490 3154 A24H6 1603070 7155540 3221 A29H7 1603050 7155520 3198 A24H8 1603060 7155440 3125 A29H9 1603210 7155360 3018 A29H10 1603490 7155230 2852 A29H11 1603500 7155180 2802 A29H12 1603560 7155240 2857 A29H13 1603570 7155230 2846 A29H14 1603650 7155240 2816 A32H15 1605800 7150710 1358 A30H16 1605770 7150720 1336 A30H17 1605850 7150670 1415 A30H18 1605950 7150620 1504 A30H19 1606040 7150490 1660 A30H20 1606190 7150410 1804 A30H21 1606180 7150360 1842 A30H22 1608390 7149060 2626 E1H23 1608320 7148960 2545 E1H24 1608510 7148810 2358 E1H25 1608640 7148680 2210 E1H26 1608500 7148530 2086 E1H27 1608510 7148520 2074 E1H28 1608570 7148480 2024 E1H29 1608580 7148480 2022 E1H30 1608620 7148410 1947 E1H31 1608770 7148560 2077 E1H32 1609740 7149230 2856 E1H33 1608820 7148630 2143 E1H34 1605030 7145740 1579 B33H35 1605010 7145740 1559 B33H36 1605050 7145740 1599 B33H37 1605130 7145630 1675 B33H38 1605190 7145460 1744 B33H39 1605470 7145340 2035 B33H40 1605580 7145250 2122 B35H41 1605590 7145260 2135 B35H42 1605780 7145080 2096 E8H43 1589200 7156110 3368 C1H44 1589190 7156140 3395 C1H45 1589240 7156130 3350 C1H46 1589260 7156130 3334 C1H47 1589050 7156060 3456 C1H48 1589040 7156070 3470 C1H49 1594290 7151440 1213 C9H50 1594320 7151420 1249 C9H51 1594410 7151400 1335 C9


5

House ID


Nearest TurbineX Y

H52 1594430 7151400 1352 C9H53 1594470 7151050 1595 C9H54 1594380 7150990 1572 C9H55 1594350 7150990 1552 C9H56 1594230 7150760 1656 C9H57 1594040 7150760 1557 C9H58 1594040 7150820 1505 C9H59 1593970 7150900 1401 C9H60 1593990 7150910 1402 C9H61 1592950 7149660 1731 C13H62 1592860 7149720 1635 C13H63 1593040 7149630 1803 C13H64 1593030 7149630 1798 C13H65 1597410 7143600 2910 B20H66 1597380 7143570 2952 B20H67 1597270 7143560 3047 B20H68 1597050 7143420 3307 B20H69 1599620 7142550 2367 B27H70 1599740 7142750 2135 B27H71 1599760 7142760 2118 B27H72 1603110 7139360 4872 B35H73 1602950 7139450 4808 B35H74 1602750 7139650 4651 B35H75 1602680 7139650 4666 B35H76 1602690 7139640 4674 B35H77 1602650 7139690 4634 B35H78 1602190 7140410 4093 B35H79 1602210 7140370 4123 B35H80 1602030 7140510 4067 B35H81 1601850 7140670 4005 B35H82 1601470 7140630 4185 B27H83 1601740 7140660 4067 B35H84 1601810 7140810 3902 B35H85 1601780 7140840 3891 B35H86 1601680 7140880 3896 B34H87 1601690 7141000 3780 B34H88 1601360 7141190 3615 B27H89 1601430 7141110 3708 B27H90 1612000 7140970 3910 E7H91 1611880 7141100 3747 E7H92 1611880 7141450 3417 E7H93 1611810 7141540 3308 E7H94 1611550 7141770 3005 E7H95 1611350 7141910 2816 E7H96 1611190 7142030 2664 E7H97 1610990 7142590 2081 E7H98 1610990 7142560 2110 E7H99 1611010 7142410 2260 E7H100 1611020 7141580 3083 E7H101 1611140 7141410 3268 E7H102 1611660 7141320 3470 E7


6

2.4 Estimation of Noise Levels at Receivers

2.4.1 Noise Propagation Model

In Sweden it is mandatory that noise immission levels are predicted using one of the noise propagation models described in ‘Ljud från Vindkraftverk’ (Naturvådsverket, 2010). Two propagation models are described for onshore use, the choice of which one to use depending on whether the separation distance between source and receiver exceeds 1 km or not. For the assessment presented here, a combination of both models has been used, as appropriate for the separation distance between each turbine and building.

The models described are simple, robust models that take account solely of geometric divergence and the atmospheric absorption of sound. As a result, they are likely to produce ‘worst case’ predictions and the implication is that the assessment presented here will be conservative.

To make these predictions, it is assumed that:

the turbines are identical

the turbines radiate noise at the power specified in Section 2.1

each turbine can be modelled as a point source at hub-height

2.4.2 Correction for Surface Roughness

The ‘Ljud från Vindkraftverk’ document (Naturvådsverket, 2010) also specifies that the sound power level data used for propagation modelling should be corrected where the site under consideration has different surface roughness characteristics to those implicitly assumed in the specification of the turbine’s sound power level, i.e. a roughness length, z0, of 0.05 m(Siemens, 2012).

Following the procedure specified in the Swedish document, a corrected sound power level, LWA,corr, is calculated from the measured value, LWA,meas, as follows:

hmeasWAcorrWA vkLL ,, [2.01]

where:

105.0ln05.0lnlnln 00 HhzhzHvv hh [2.02]

and where:

k is the rate of change of sound power level to 10 m height wind speed, in dB per ms-1;

H is the hub height of the turbine, here 115 m;

z0 is the actual roughness length of the site, in meters;

h is 10 m.

The value of LWA,corr determined using this procedure should be used in place of LWA,meas in propagation modelling.

However, for the turbine considered for Fjällberg, it is not considered that the correction methodology, as specified here, is strictly appropriate. The reason for this is the relationship between the sound power level of the machine and the hub height wind speed is complex and actually may not consistently increase with increasing wind speed. Equation 2.01 assumes a


7

generally increasing relationship between the two and this is clearly not appropriate for this turbine (see Table 1).

As an alternative, the maximum sound power level that this turbine generates has been adopted for LWA,corr.

2.4.3 Conservatism in Propagation Modelling

As mentioned in Section 2.4.1, the noise immission level predictions presented here are likely to be conservative, and the noise immission levels that would be experienced in practice would likely be significantly lower. The main reason for this is that the noise propagation model which has been used, the use of which is mandatory, is a simple, robust model that takes account solely of geometric divergence and atmospheric sound absorption. It does not consider other sound attenuating effects, e.g. barrier effects, ground effects andmeteorological effects and, as a result, is likely to produce sound level predictions which considerably over-estimate the ‘true’ values for the majority of the time.

It is interesting to note that if the commonly used ISO 9613 Part 2 environmental noise propagation model had been used in place of the mandatory model, noise level predictions at the nearest properties would be on average 2.8dB(A) lower than those predicted by the SEPA defined model.

2.4.4 Predictions

Table 4 shows the predicted noise immission levels at the assessed buildings considering the 60 turbine layout. The maximum predicted noise level is 41.0dB(A) at house K11.

Figure 1 (Appendix A) shows the isobel (i.e. noise contour) plot for this proposed layout at the maximum noise output according to the ‘Ljud från Vindkraftverk’ propagation model. Such plots are useful for evaluating the noise ‘footprint’ of a given development and are for informational purposes only.

Table 4 Predicted Maximum Noise Levels At Nearby Buildings

Shading indicates greater than 40dB(A)

House ID

Sound PressureLevel / dB(A) re. 20 μPa

K01 31.6K03 31.5K04 31.4K05 31.7K06 36.1K08 38.4K09 38.9K10 37.8K11 41.0K12 39.4K13 39.5K14 39.5K15 35.5K17 38.4K19 36.7K20 36.6K21 36.6K22 36.7K24 30.6H1 38.7H2 31.7


8

House ID


H3 31.7H4 31.8H5 31.8H6 31.7H7 31.7H8 31.9H9 32.1H10 32.5H11 32.6H12 32.5H13 32.5H14 32.5H15 36.6H18 36.0H19 35.4H20 34.8H21 34.8H22 32.8H23 33.0H24 33.4H25 33.7H26 34.2H27 34.2H29 34.4H30 34.6H31 34.1H32 31.6H33 33.8H36 37.1H37 37.0H38 36.9H39 36.3H40 36.2H41 36.2H42 36.0H43 29.2H44 29.1H45 29.3H46 29.3H47 29.0H48 29.0H51 37.5H53 36.9H55 37.0H56 36.9H57 37.1H58 37.2H59 37.6H60 37.5H61 35.9H62 36.1H63 35.7H64 35.7H65 31.5H66 31.4


9

House ID


H67 31.2H68 30.6H69 32.5H70 33.3H72 26.7H73 26.9H74 27.3H75 27.2H76 27.2H77 27.3H78 28.7H79 28.6H80 28.9H81 29.2H82 29.0H83 29.1H84 29.5H85 29.5H86 29.6H87 29.9H88 30.2H89 30.0H90 27.4H91 27.8H92 28.6H93 28.9H94 29.9H95 30.5H96 31.1H97 33.3H98 33.1H99 32.6H100 30.0H101 29.4H102 28.6

2.5 Acoustic Acceptance Criteria

The framework most commonly used to assess the acceptability of wind farm noise immission levels in Sweden is based on guidelines issued by the Swedish Environmental Protection Agency (SEPA, 1978). This has become essentially the de facto standard for such developments within Sweden and it is suggested that this should also be adopted for the Fjällberg site.

The impact of the proposed wind farm has been established, in accordance with the recommendations of SEPA, by comparing the calculated noise levels produced by the proposed wind turbines with the noise limits specified, at nearby properties.

2.5.1 SEPA Assessment Procedure

It is understood that the SEPA guidelines, originally issued in 1978, have now been superseded: first in 1983 and again more recently. SEPA’s web site states that 40dB(A) outdoors in residential areas should not be exceeded. In areas where background noise is low


10

and the sound environment is particularly important, SEPA recommends that 35dB(A) is not exceeded (SEPA, 2012).

2.6 Acoustic Assessment

For the 60 turbine layout scenario the 40dB(A) level is exceeded at one of the nearest neighbours.

2.7 Turbine Management

Turbine management is proposed to ensure that the predicted noise levels at all occupied houses meet the criteria level of 40dB(A). The turbine management strategy involves reduction of noise levels at each house to below the criteria level through management of selected turbines. This is a conservative approach as predictions are made under the assumption that all properties are always downwind of the turbines, therefore the proposed turbine management would not be required at all times.

It should be acknowledged that there will be many different combinations of turbine management that result in predicted noise levels lower than the specified criteria. The suggestion that follows represents a potential turbine management scheme which may feasibly not be the most efficient from an energy capture perspective but simply demonstrates the principle of the use of turbine management to mitigate noise levels at all properties to acceptable levels.

2.7.1 Noise Reduced Modes for the Proposed Wind Turbines

The noise emission characteristics of the proposed wind turbines operating normally are as detailed in Section 2.1. However, the operation of this type of turbine may be altered by changing the pitch of the wind turbine blades, resulting in a trade-off between power production and noise reduction. There are several alternative modes of operation available. This investigation aims to identify the turbines that would be required to operate in reduced noise mode in order to meet the criteria.

Octave band data for some of the reduced noise modes available for the proposed Fjällberg wind turbines are shown in Table 5. These are presented at the maximum sound power level for each of these operational modes and were extracted from the available acoustic data on these machines (Siemens, 2012).

Table 5 Octave Band Data for Siemens SWT-3.0-113 Noise Reduced Modes

Frequency (Hz)

A-Weighted Sound Power Level / dB(A) re 1 pW at 8m/s

Normal Operation

“-1dB” Mode

“-2dB” Mode

“-3dB” Mode

“-4dB” Mode

“-5dB” Mode

63 86.9 86.7 86.5 86.3 86.1 85.9

125 94.8 94.5 94.1 93.7 93.4 93.0

250 100.2 99.0 98.0 96.8 95.7 94.4

500 101.1 99.6 98.7 97.2 95.9 94.5

1000 100.7 99.8 98.7 97.8 96.8 95.7

2000 99.7 98.9 97.7 96.9 96.0 94.9

4000 92.6 92.6 91.0 91.0 90.5 89.8

Total 107.0 106.0 105.0 104.0 103.0 102.0

2.7.2 Proposed Turbine Management

The following turbine management strategy is proposed to reduce the predicted noise levels to below the criteria level of 40dB(A) at all occupied properties:


11

Operate turbine A23 in the “-1dB” reduced mode

Operate turbine A24 in the “-4dB” reduced mode

As previously mentioned, the exact noise management strategy employed may differ from that presented here. The above strategy is provided to illustrate that the 40dB(A) criteria level can be met.

Figure 2 (Appendix A) shows the isobel (i.e. noise contour) plot for the proposed layout with the identified noise management applied according to the ‘Ljud från Vindkraftverk’ propagation model. Such plots are useful for evaluating the noise ‘footprint’ of a given development and are for informational purposes only.

3.0 OTHER ASPECTS OF ACOUSTIC NOISE

The main focus of the acoustic impact assessment of operational noise from the wind farm presented here is based on the two most relevant types of noise emission for modern wind turbines: broadband and tonal noise, both of which are types of ‘audible noise’. Implicitly incorporated within this assessment is the normal character of the noise associated with wind turbines (commonly referred to as “swish”) and consideration of a range of noise frequencies, including low frequencies.

3.1 Low Frequency Noise

The frequency range of ‘audible noise’ is generally taken to be 20 Hz to 20,000 Hz, with the greatest sensitivity to sound typically in the central 500 Hz to 4,000 Hz region. The range from 10 Hz to 200 Hz is generally used to describe ‘low frequency noise’, and noise with frequencies below 20 Hz used to describe ‘infrasound’ (Leventhall, 2003), although there is sometimes a lack of consistency regarding the definition of these terms in both common usage and the literature.

Low frequency noise is always present, even in an ambient ‘quiet’ background (Leventhall, 2003). It is generated by natural sources, including the sea, earthquakes, the rumble of thunder and wind. It is additionally an emission from many artificial sources found in modern life, such as household appliances (e.g. washing machines, dishwashers) and all forms of transport.

Noise emitted from wind turbines covers a broad spectrum from low to high frequencies. In relation to human perception of the broadband noise produced by wind turbines, the dominant frequency range is not the low frequency or infrasonic ranges (Ontario Ministry of the Environment, 2010). The reason for this is that the perception threshold for hearing in these ranges is much higher than for speech frequencies of between 250 Hz and 4000 Hz. As a result of this decreased sensitivity, wind turbine noise at the lowest frequencies of the range described as ‘low frequency noise’ would be below the average hearing threshold.

A comprehensive literature review of ‘Low Frequency Noise and Infrasound Associated with Wind Turbine Generator Systems’, undertaken for the Ontario Ministry for the Environment in 2010, indicates that low frequency noise from wind turbines crosses the threshold boundary, and thus would be considered to become audible, above frequencies of around 40-50 Hz (Ontario Ministry of the Environment, 2010). The degree of audibility depends upon the wind conditions, the degree of masking from background noise sources and the distance from the wind turbines (Ontario Ministry of the Environment, 2010).

Although audible under some conditions, a paper; ‘Infrasound and low frequency noise from wind turbines: exposure and health effects’ (Bolin et al., 2011), published by the authors of a literature review on the subject prepared for the Swedish Environmental Protection Agency in 2011 (SEPA, 2011), concludes that the level of low frequency noise produced by wind turbines does not exceed levels from other common sources, such as road traffic noise (Bolin et al., 2011).


12

In response to an article published in the UK national press in 2004, alleging that low frequency noise from wind turbines may give rise to adverse health effects, the UK Department of Trade and Industry (DTI) commissioned the Hayes McKenzie Partnership to perform an independent study to investigate these claims (Hayes, 2006). The UK Government released the following advice based on the report’s findings:

“The report concluded that there is no evidence of health effects arising from infrasound or low frequency noise generated by wind turbines.” (DTI, 2006)

This is re-iterated in the review undertaken for the Ontario Ministry for the Environment (Ontario Ministry of the Environment, 2010), which concludes that publications by medical professionals indicate that; at typical setback distances, the noise levels produced by wind turbines, including noise at low and infrasound frequencies, do not represent a direct health risk (Ontario Ministry of the Environment, 2010).

Whilst the low frequency content of the noise from wind farms has been considered through the use of octave band specific noise emission and propagation modelling within the assessment presented here, it is considered that a specific and targeted assessment on the low frequency content of noise emissions from the proposed wind farm is unjustified.

3.2 Infrasound

In relation to infrasound in general; frequencies below 20 Hz may be audible, although tonality is lost below 16 - 18 Hz, thus losing a key element of perception (Leventhall, 2003). In relation to modern, upwind turbines; there is strong evidence that the levels of infrasound produced will be well below the average threshold of human hearing (Ontario Ministry of the Environment, 2010). The aforementioned DTI report (Hayes, 2006) extended this conclusion to more sensitive members of the population:

“Even assuming the most sensitive members of the population have a hearing threshold which is 12 dB lower than the median hearing threshold, measured infrasound levels are well below this criterion” (Hayes, 2006).

As such:

“infrasound from wind turbines is not audible at close range and even less so at distances where residents are living” (Bolin et al., 2011).

In February 2005, the BWEA3 published background information on low frequency noise from wind farms (BWEA, 2005). The conclusion states that:

"It has been repeatedly shown, by measurements of wind turbine noise undertaken in the UK,Denmark, Germany and the USA over the past decade, and accepted by experienced noise professionals, that the levels of infrasonic noise and vibration radiated from modern upwind configuration wind turbines are at a very low level; so low that they lie below the threshold of perception, even for those people who are particularly sensitive to such noise, and even on an actual wind turbine site". (BWEA, 2005)

The BWEA report goes on to quote Dr Geoff Leventhall, author of the DEFRA report on “Low Frequency Noise and its Effects” (BWEA, 2005), as saying:

"I can state, quite categorically, that there is no significant infrasound from current designs of wind turbines". (BWEA, 2005)

With regard to health effects, the DTI report quotes the document ‘Community Noise’, prepared for the World Health Organisation (WHO), which states that:

3 BWEA is now known as RenewableUK, a group representing the concerns of companies in the Renewable Energy Industry in the UK


13

“there is no reliable evidence that infrasound below the hearing threshold produce physiological or psychological effects” (Hayes, 2006)

The DTI report goes on to conclude that:

“infrasound associated with modern wind turbines is not a source which will result in noise levels which may be injurious to the health of a wind farm neighbour” (Hayes, 2006)

Furthermore, researchers at Keele University in the UK explain that:

“The infrasound generated by wind turbines can only be detected by the most sensitive equipment, and again this is at levels far below that at which humans will detect the low frequency sound. There is no scientific evidence to suggest that infrasound has an impact on human health.” (Styles and Toon, 2005)

Therefore, in accordance with the literature, it is not considered appropriate or relevant to undertake specific assessment in relation to infrasound for the proposed wind farm.

3.3 Vibration

Structure borne noise, originating in vibration, is also low frequency, as is neighbour noise heard through a wall, since walls generally block higher frequencies more than lower frequencies.

A report by Snow gives details of low frequency noise and vibration measurements made ata wind farm (Snow, 1997). Measurements were made both on the wind farm site, and at distances of up to 1 km. It was found that the vibration levels at 100 m from the nearest turbine itself were a factor of 10 lower than those recommended for human exposure in the most critical buildings (i.e. laboratories for precision measurements), and lower again than the limits specified for residential premises (BSI, 1992). Noise and vibration levels were found to comply with recommended residential criteria, even on the wind turbine site itself, and the acoustic signal was below the generally assumed frequency range of audible noise i.e. below 20 Hz. In addition, it was found that there was no clear relationship between vibration levels and wind speed, and that some vibrations appeared to come from other sources, as they were found even when the turbines were switched off.

More recently, in 2004/2005, researchers at Keele University (in the UK) investigated the effects of the extremely low levels of vibration resulting from wind farms on the operation of the seismic array at Eskdalemuir (UK) - one of the most sensitive installations in the world. The results of this study have frequently been misinterpreted and, to clarify the position, the authors have explained that:

"The levels of vibration from wind turbines are so small that only the most sophisticated instrumentation and data processing can reveal their presence, and they are almost impossible to detect" (Styles and Toon, 2005).

They go on to say:

"Vibrations at this level and in this frequency range will be available from all kinds of sources such as traffic and background noise - they are not confined to wind turbines. To put the level of vibration into context, they are ground vibrations with amplitudes of about one millionth of a millimeter. There is no possibility of humans sensing the vibration and absolutely no risk to human health” (Styles and Toon, 2005)

Therefore, in accordance with literature, it is not considered appropriate or relevant to undertake specific assessment in relation to vibration caused by the operation of the proposed wind farm.


14

3.4 Aerodynamic Modulation

The noise normally associated with wind turbines and commonly referred to as “Swish” is the modulation of aerodynamic noise produced at blade passing frequency (the frequency at which a blade passes a fixed point). This noise character is acknowledged by, and accounted for, in the recommendations of ETSU-R-974, the guidance used to assess and rate noise from wind energy developments in the UK (ETSU, 1996). However the aforementioned DTI report (Hayes, 2006) researching low frequency noise and/or infrasound emitted by wind turbines noted that a related phenomenon known as ‘Aerodynamic Modulation’ (AM) - alternatively referred to as ‘Amplitude Modulation’, was, in some isolated circumstances, occurring in ways not anticipated by ETSU-R-97. Such AM above and beyond that considered by ETSU-R-97 is often referred to as Excess, or Other, AM.

To investigate whether or not Other AM was an issue which might require attention in the context of the rating advice in ETSU-R-97, the UK Government subsequently commissioned the University of Salford (UK) to undertake further research in the area (DTI, 2006).

On 1 August 2007, the UK Government issued a statement (BERR, 2007) regarding the findings of the University of Salford report into (Other) AM of wind turbine noise (University of Salford, 2007) published earlier in 2007 which found that, of 133 operational wind farms in the UK at the time of the report, there were only 4 cases where AM may have been a factor. It is known that complaints have now subsided for 3 of these cases (one due to introduced mitigation by a wind farm control system) and in the remaining case a settlement has been reached. The statement says that:

“…the Government does not consider there to be a compelling case for further work into AM and will not carry out any further research at this time.”

Several potential causes for these occurrences of this Other AM have been suggested including: high wind shear; stall; yaw error; blade-tower interaction; inflow turbulence; and wake interference between closely located turbines. There is, however, currently no clear evidence to support any of the proposed causative mechanisms of Other AM. This is partly due to the difficulty in obtaining sufficiently detailed measurements of Other AM and the conditions under which it occurs, this being as a direct consequence of the infrequency of occurrence and the small number of sites at which high levels of Other AM have been reported. Consequently, the cause of Other AM is still a subject of ongoing research.

Therefore, in accordance with literature and advice, it is not considered appropriate or relevant to undertake specific assessment in relation to Other AM that may be potentially produced by the operation of the proposed wind farm.

3.5 Wind Turbine Syndrome

The condition proposed by paediatrician Dr Nina Pierpont in her report ‘Wind Turbine Syndrome: A Report on a Natural Experiment’ (Pierpont, 2009) cites a range of physical sensations and effects as being caused by living near a wind farm. This study is based on a series of interviews comprising a study group of 10 families. It is a self published report with none of the research being published in any peer reviewed medical journal.

In a UK National Health Service (NHS) response to the Pierpont report, a report titled ‘Are wind farms a health risk?’ (NHS, 2009) states that there is no conclusive evidence that wind turbines have an effect on health or are causing the set of symptoms described as ‘wind turbine syndrome’. It was noted that the group study by Pierpont was not sufficient to grant the claims stated.

4 ETSU is the Energy Technology Support Unit, UK


15

A scientific advisory panel conducted a review of current literature available on the issue of perceived health effects of wind turbines ‘Wind Turbine Sound and Health Effects - An Expert Panel Review’ (Colby et al., 2009). This was carried out by the American and Canadian Wind Energy Associations and the conclusion on Wind Turbine Syndrome was that it is

“not a recognized medical diagnosis, is essentially reflective of symptoms associated with noise annoyance and is an unnecessary and confusing addition to the vocabulary on noise.”

The report went on to say:

“There are no unique symptoms or combinations of symptoms that would lead to a specific pattern of this hypothesized disorder.”

An independent review of the state of knowledge about the alleged health condition was carried out (RenewableUK, 2010). This report includes three expert opinions provided by: Richard J.Q. McNally - Reader in Epidemiology at the Institute of Health and Society Newcastle University; Geoff Leventhall – an independent consultant specialising in low frequency noise, infrasound and vibration; and Mark E. Lutman - Professor of Audiology at the University of Southampton. Their critique of Pierpont’s study concludes that the reported symptoms are the effects mediated by stress and anxiety when exposed to an adverse element in their environment. There is no evidence that they are patho-physiological effects of wind turbine noise.

The authors of ‘A literature review of infra and low frequency noise from wind turbines: exposure and health effects’, prepared for the Swedish Environmental Protection Agency in 2011 (SEPA, 2011), identify several limitations with the work undertaken by Pierpont which make the conclusions of the study unjustified (Bolin et al., 2011). No consistent associations between wind turbine noise exposure and symptom reporting were found (Bolin et al., 2011).

A paper by Pedersen explores data from three cross-sectional studies comprising A-weighted sound pressure levels of wind turbine noise, and subjectively measured responses from 1,755 people, to find the relationships between sound levels and aspects of health and well-being. It was concluded that there is no consistent association between wind turbine noise exposure and the symptoms associated with Wind Turbine Syndrome (Pedersen, 2011).

Therefore, in accordance with literature, it is not considered appropriate or relevant to undertake specific assessment in relation to Wind Turbine Syndrome potentially caused by the operation of the proposed wind farm.

3.6 Wind Shadow Impact

Wind shadow zones are defined by the Environmental Protection Agency as areas where wind speed is of the order of 50% less than on the wind farm site. Any properties located in such sheltered areas may therefore experience lower background noise levels and the impact of a given noise level may be increased as a result.

As stated in Section 2.4.3, the propagation model used in this assessment does not consider sound attenuating effects such as barrier effects, ground effects and meteorological effects. As a result, the sound level predictions are likely to considerably over-estimate the ‘true’ values for the majority of the time. Barrier effects and meteorological effects are particularly relevant to the discussion of wind shadow impact. Areas which are sheltered from the wind may also provide some shelter from the noise, however, any resulting attenuation has not been considered here. Furthermore, it has been assumed that a given property would be downwind of all turbines simultaneously. In reality, the noise level at a given property will be less when the property is crosswind or upwind of the


16

proposed wind turbines although, again, this has not been taken into account in this assessment.

In summary, the impact of wind shadow would be reduced should barrier effects and meteorological effects be included. As such, it is considered that the conservative nature of the propagation model in not including such effects makes an allowance for wind shadow impact.

3.7 Conditions Favourable to Propagation

Meteorological conditions favourable to the propagation of sound include situations when the receiver is downwind of the source and also temperature inversions, where the air at ground level is cooler than the air above it, such as commonly occurs on clear, calm nights. The ‘Ljud från Vindkraftverk’ propagation model used in this assessment has been demonstrated (see Section 2.4.3) to be more conservative than the commonly used ISO 9613-2 propagation model which is itself conservative in that it assumes meteorological conditions favourable to propagation are present. As such, the conservatism of the simple propagation model used in this assessment implies that it is likely to considerably over-estimate the noise levels under most meteorological conditions.

3.7.1 Fog

The increased humidity associated with fog will have only a second order, minimal impact on sound propagation. However, fog is a characteristic of highly stable atmospheric conditions which may influence sound propagation in extreme cases. Such highly stable conditions, however, are also associated with low wind speeds where the noise levels produced by the wind farm will be lower than those detailed in this assessment where the maximum sound power level of the proposed turbines has been used. This, along with the conservatism of the propagation model detailed in Section 2.4.3, is likely to considerably over-estimate the noise levels under most meteorological conditions.

4.0 CONCLUSIONS

The acoustic impact for the proposed Fjällberg wind farm on nearby neighbours has been assessed in accordance with the guidance on wind farm noise issued by the Swedish Environmental Protection Agency (SEPA, 1978).

This assessment has presented a layout comprising 60 Siemens SWT-3.0-113 wind turbines.

Noise levels have been predicted at 111 identified buildings. Predicted noise levels do not exceed the 40dB(A) assessment criteria level at any of the nearest properties with the application of a noise management strategy as detailed in Section 2.7.

It is concluded that noise, in the low frequency (10 to 200 Hz) range, and vibration from the proposed installation is unlikely to be a problem. The proposed wind farm is not therefore predicted to have an amenity effect on local properties due to vibration, infrasound or low frequency noise.


17

5.0 REFERENCES

BERR, 2007. “Government statement regarding the findings of the Salford University report into Amplitude Modulation of Wind Turbine Noise”, URN 07/1276, dated July 2007 (http://www.berr.gov.uk/files/file40571.pdf)

Bolin, K., Bluhm, G., Eriksson, G. and Nilsson, M.E., 2011. “Infrasound and low frequency noise from wind turbines: exposure and health effects”, Environmental Research Letters 6, September 2011

BSI, 1992. British Standards Institution, “Guide to Evaluation of human exposure to vibration in buildings (1 Hz to 80 Hz)”, BS 6472, 1992

BWEA, 2005. “Low Frequency Noise and Wind Turbines”, January 2005, http://www.bwea.com/ref/lowfrequencynoise.html & Technical Annex: http://www.bwea.com/pdf/lfn-annex.pdf, dated February 2005

Colby, W.D., Dobie, R., Leventhall, G., Lipscomb, D.M., McCunney, R.J., Seilo, M.T. and Søndergaard, B., 2009. “Wind Turbine Sound and Health Effects - An Expert Panel Review 2009”, prepared for American Wind Energy Association and Canadian Wind Energy Association.

DTI, 2006. “Advice on findings of the Hayes McKenzie report on noise arising from Wind Farms”, URN 06/2162, dated November 2006 (http://www.berr.gov.uk/files/file35592.pdf)

ETSU, 1996. “The Assessment and Rating of Noise from Wind Farms”, The Working Group on Noise from Wind Turbines, ETSU Report for the DTI, ETSU-R-97, September 1996

Hayes, 2006. “The Measurement of Low Frequency Noise at Three UK Wind Farms”, Contract Number W/45/00656/00/00, URN 06/1412 http://www.berr.gov.uk/files/file31270.pdf

Leventhall, G., 2003. Dr Geoff Leventhall, “A Review of Published Research on Low Frequency Noise and Its Effects”, Report for DEFRA, May 2003

Naturvådsverket, 2010. “Ljud från vinkraftverk”, Report 5933, April 2010

NHS, 2009. “Are wind farms a health risk?”, www.nhs.uk/news/2009/08August/Pages/Arewindfarmsahealthrisk.aspx

Ontario Ministry of the Environment, 2010. “Low Frequency Noise and Infrasound Associated with Wind Turbine Generator Systems, a Literature Review”, OSS078696, December 2010

Pedersen, 2011. “Health aspects associated with wind turbine noise—results from three field studies” Noise Control Engineering Journal, Volume 59, Issue 1

Pierpont, N., 2009. “Wind Turbine Syndrome - A Report on a Natural Experiment”, K-Selected Books

RenewableUK, 2010. “Wind Turbine Syndrome (WTS) - An independent review of the state of knowledge about the alleged health condition”, www.bwea.com/pdf/publications/HS_WTS_review.pdf

SEPA, 1978. Swedish Environmental Protection Agency, Report 78:5, dated 1978

SEPA, 2011. “A literature review of infra and low frequency noise from wind turbines: exposure and health effects”, prepared for Swedish Environmental Protection Agency, November 2011

SEPA, 2012. Riktvärden för ljud från vindkraft, http://www.naturvardsverket.se/en/Start/Verksamheter-med-miljopaverkan/Buller/Vindkraft/Riktvarden-for-ljud-fran-vindkraft/

Siemens, 2012. “Acoustic Emission, SWT-3.0-113”

Snow, D.J., 1997. “Low Frequency Noise & Vibration Measurements at a Modern Windfarm”, ETSU W/13/00392/REP, 1997.

Styles, P. and Toon, S., 2005. "Wind farm noise" a letter by (Prof) Peter Styles, President, Geological Society of London and Sam Toon, Keele University, Staffordshire, printed in the


http://www.naturvardsverket.se/en/Start/Verksamheter-med-miljopaverkan/Buller/Vindkraft/Riktvarden-for-ljud-fran-vindkraft/

http://www.naturvardsverket.se/en/Start/Verksamheter-med-miljopaverkan/Buller/Vindkraft/Riktvarden-for-ljud-fran-vindkraft/

http://www.bwea.com/pdf/publications/HS_WTS_review.pdf

http://www.nhs.uk/news/2009/08August/Pages/Arewindfarmsahealthrisk.aspx

http://www.berr.gov.uk/files/file31270.pdf


http://www.bwea.com/pdf/lfn-annex.pdf

http://www.bwea.com/ref/lowfrequencynoise.html


18

Scotsman newspaper as a rebuttal of claims made by the Renewable Energy Foundation, August 2005

University of Salford, 2007. “Research into Aerodynamic Modulation of Wind Turbine Noise: Final Report”, URN 07/1235, dated July 2007 (http://www.berr.gov.uk/files/file40570.pdf)



19

APPENDIX A: FIGURES

Figure 1 Predicted Noise Footprint for the Proposed Fjällberg Wind Farm

Grid Intervals at 1 kmThe LAeq descriptor has been used

The noise footprint has been calculated when turbines are operating at their maximum noise emissionusing the ‘Ljud från Vindkraftverk’ propagation model and is for indication only


20

Figure 2 Predicted Noise Footprint with Noise Management

Grid Intervals at 1 kmThe LAeq descriptor has been used

The noise footprint has been calculated when turbines are operating at their maximum noise emissionusing the ‘Ljud från Vindkraftverk’ propagation model and is for indication only


21

GLOSSARY

Broadband NoiseNoise which covers a wide range of frequencies (e.g. from 10 Hz to 5 kHz).

dB(A)The decibel (dB) is a logarithmic unit used in acoustics to quantify sound levels relative to a 0 dB reference (a sound pressure level of 2*10-5 Pa). The ‘A’ signifies A-weighting which is a frequency-response function that applies an international weighted scale of sound levels in each frequency band (octave band or third octave band) providing a good correlation with the sensitivity of the human ear which is less sensitive to very high and very low frequencies.

FrequencyThe pitch of a sound in Hz or kHz. See Hz.

HzSound frequency refers to how quickly the air vibrates, or how close the sound waves are to each other (in cycles per second, or Hertz (Hz)).

Leq

The equivalent continuous noise level is a notional steady noise level, which over a given time, would provide the same energy as the intermittent noise. Noise standards often specify the length of time over which noise should be measured.

L90

Sound pressure level exceeded for 90% of the time for any given time interval. For example, L(A)90,10min

means the A-weighted level that is exceeded for 90% of a ten minute interval. This indicates the noise levels during quieter periods, or the background noise level. It represents the lower estimate of the prevailing noise level, and is useful for excluding the effects of, for example, aircraft or dogs barking on background noise levels.

LW

Sound power level is the acoustic power (W) radiated from a sound source. This power is essentially independent of the surroundings, while the sound pressure depends on the surroundings (reflecting surfaces) and distance to the receiver.

Noise EmissionThe noise energy emitted by a source (e.g. a wind turbine).

Noise Immission The sound pressure level detected at a given location (e.g. nearest dwelling).

Octave BandRange of frequencies between one frequency (f0*2

-1/2) and a second frequency (f0*2+1/2). The quoted centre

frequency of the octave band is f0.

Sound FrequencyRefers to how quickly the air vibrates, or how close the sound waves are to each other (in Hertz).Frequency is subjectively felt as the pitch of the sound. The lowest frequency audible to humans is 20 Hz and the highest is 20,000 Hz. The human ear is most sensitive to the 1 kHz, 2 kHz and 4 kHz octaves and much less sensitive at the lower audible frequencies.

SpectrumDescription of the sound pressure level of a source as a function of frequency.

Third Octave BandThe range of frequencies between one frequency (f0*2

-1/6) and a second frequency equal to (f0*2+1/6). The

quoted centre frequency of the third octave band is f0.

Tonal NoiseNoise which covers a very restricted range of frequencies (e.g. a range of <=20 Hz). This noise is more annoying than broadband noise.


assessment of the acoustic impact of the proposed ... · assessment of the acoustic impact of the...

Documents