Perception of noise from large wind turbines
(EFP-06 Project)

von Hünerbein, S; King, A; Hargreaves, JA; Moorhouse, AT; Plack, C

Abstract

Is noise from large wind turbines more annoying than noise from small wind turbines? This is a question that is discussed widely in the context of a new generation of large wind turbines replacing the traditional smaller ones. To date legislation takes noise levels and the tonality of noise sources into account. However, many more influencing factors are known from the psychoacoustic literature. Examples are the nature of the listening environment, spectral and temporal characteristics of the sound, and the influence of masking noise. An earlier part of the EFP-06 project established the measurable differences between large and small wind turbines. It concluded that spectral characteristics are generally very similar apart from a slight increase in the low frequency content of large turbines.

In this study on the perception of wind turbine noise, audibility thresholds and equal annoyance contours have been established for idealised wind turbine sounds containing low frequency tones. The listening test simulated an indoor scenario and an outdoor scenario with and without masking garden noise.

The focus has been on the question whether annoyance changes with the frequency of a tone. The test sounds consisted of a broadband spectrum with a specific tone at one of the frequencies 32, 44, 72, 115, 180 and 400 Hz. Idealised sounds with features broadly representative of wind turbine sounds were used. The participant were asked to imagine being in different scenarios. The outdoor scenario presented sounds broadly representative of a wind turbine at three A-weighted sound pressure levels, each with and without garden noise, whereas the indoor scenario omitted the garden noise since the facade attenuation rendered it inaudible. A comparative adaptive method was used to establish relative equal annoyance levels in the form of equal annoyance contours. The results enable comparisons between different scenarios, broadband levels, tone frequencies, masked and unmasked ‘wind turbine’ sound, and two different prominence levels for the reference tone. Temporal variation like “swishing” was avoided to keep the research questions well focused.

In a second part of the study wind turbine recordings from a large and a small wind turbine were compared in annoyance with steady traffic noise. The recordings were manipulated to include the effect of sound propagation and façade attenuation. They were also normalised to equal A-weighted levels.

The study concludes:

Tones in quiet were heard at levels that agree well with hearing thresholds published elsewhere. As the broadband noise level increases the tones were heard at levels that were determined by the masking level. Masking thresholds predicted by the ISO 1996-2 standard have been shown to agree well with the measured tonal audibility thresholds as long as the masking noise clearly exceeds the hearing threshold of the tones. As low levels can frequently occur indoors in the neighbourhood of wind turbines when the Danish noise regulations are observed it would be useful to extend the standard to include a method to evaluate the hearing threshold.

One possible method published by Pedersen (2008) to establish the audibility of broadband spectra has been successfully tested for two examples: a broadband spectrum of room background noise and the broadband spectra of wind turbine noise at levels close to the hearing threshold. The calculated critical band levels agree to within 2 dB with perceived audibility. Theoretical considerations support the conclusion that the method should be adequate for use in standard applications.

Low frequency tones had to be adjusted to higher tone levels above the masking threshold to be equally annoying as higher frequency tones. Garden noise was not shown to reduce annoyance because different scenarios could not be compared easily.

It was shown that increasing the tone level by 5 dB increases the equal annoyance level by a smaller value both for tone frequencies lower than 180 Hz and at 400 Hz. This casts doubt on the appropriateness of the adjustment used in the ISO 1996-2 standard which adds penalty adjustments which are increasing linearly with sound pressure level above masking.

Relative sensation levels were calculated from equal annoyance contours to determine whether low frequency tones are relatively more annoying than high frequency tones. The frequency dependence was not shown to be significant. The main influence on these levels is the tone level above masking level: Tones at higher levels are more annoying than tones at lower levels above masking. Both findings are common for the indoor and outdoor scenarios.

To compare real recordings of a large and a small wind turbine a test protocol was developed. This method was successfully trialled.

The comparison between normalised recordings showed the spectral characteristics of the small turbine to be more annoying outdoors than those of the large turbine recording. This has been attributed to the different spectral characteristics of the two turbines. These differences are effectively masked by garden noise and the equal annoyance ratings change accordingly. The indoor scenario does also not find the turbines to be differently annoying. If these results can be reproduced in other listening experiments then it follows that the specific differences in spectral content will determine the annoyance levels from a wind turbine more than whether it is a small or a large turbine. It would also mean that the differences in annoyance between wind turbines get smaller when sufficient masking noise is present. Presently, the finding that the small turbine is more annoying cannot be generalised to large and small wind turbines or to a wider range of wind and terrain conditions than were used in the test. The listener responses were however consistent and therefore demonstrate the potential of the comparison method.

Another significant achievement of this project was of technical nature: It was the design of an immersive sound reproduction system that is calibrated to high precision over the largest part of the audible frequency range including low frequencies down to 30 Hz. It has been shown that this design is possible and that the stimuli sound realistic. Future listening test with similar requirements will therefore be easy to design and fast to perform.

In answer to the initial question whether large turbines are more annoying than small wind turbines, the results of this study find no evidence for a significant difference in annoyance between small and large wind turbines as long as total noise levels and tonal characteristics are taken into account in the assessment. Temporal variations of wind turbine noise such as the level of swishing might also have to be evaluated in the future.

Citation

(EFP-06 Project)

• Report Type: Technical Report
• Deposit Date: Jan 13, 2011
• Publicly Available Date: Apr 5, 2016
• Publisher URL: http://www.madebydelta.com/delta/Business_units/TC/Services+by+technology/Acoustics/Low+frequency+noise/Low+frequency+noise+from+large+wind+turbines.page?