Amplitude Modulation In Wind Farm Noise

Acoustic ‘character’, such as tonal or impulsive content, is anything that makes noise more noticeable than if it were a bland steady-state noise. Regular or irregular variation in noise is called amplitude modulation (AM) and, for wind turbine noise, this can be a regular variation caused by the rotation of the blades. Not all wind turbines exhibit AM at levels that cause annoyance, but where AM levels are sufficiently high it can cause disturbance to residents around a wind farm site. At Hayes McKenzie, we’ve had extensive experience in measuring AM from wind farms and single turbines, and have carried out a significant number of measurements prior to, and since, the most recent UK publications in the area of wind farm AM from the Institute of Acoustics (IOA) and the Department for Energy and Climate Change (DECC), which are detailed further below. This article looks at the recent developments in the UK and our experience of applying the latest guidance in practice.

What is Amplitude Modulation of Wind Farm (or Wind Turbine) Noise

Amplitude modulation of wind turbine noise relates to the change in amplitude (loudness) occurring at the blade passing frequency (three times the rotational speed for three-bladed machines). This is commonly described as swish, whoomp, thump, or whoosh. The level of AM can be quantified by the modulation depth (MD) or measured ‘peak’ to ‘trough’ level, with higher MD equating to greater levels of AM. In some cases, the modulation depth can be quantified by subtracting the L95 (noise level exceeded for 95% of the time) from the L5 (noise level exceeded for 5% of the time) over a specified a time period as shown in the figure below for two 10 second periods.

In practice, however, this procedure can be significantly affected by other noise not related to wind turbine AM referred to as ‘false positives’. A more robust method for quantifying wind turbine AM is by using the metric described in the Institute of Acoustic’s (IOA) Amplitude Modulation Working Group (AMWG) report to derive AM ratings for successive 10-minute periods. Wind farm noise data is usually referenced to 10-minute overall averaging periods as wind farm and associated meteorological data is normally referenced to 10-minute intervals.

All conventional horizontal axis wind turbines exhibit blade swish to a certain extent, but this tends to become less noticeable with increasing distance such that it is often not normally a significant feature at residential properties. There are times when the swish becomes more of a thump or whoomp and this is likely to be due to the blades entering stall conditions at the highest point of their sweep, usually under higher wind shear conditions when the wind speed varies significantly between the highest and lowest points of the blade swept area.

Recent UK Publications and Meetings on Wind Turbine Amplitude Modulation

The UK has been progressive in its investigation of AM and a significant amount of research has been undertaken in this area. Renewable UK published a study of Wind Turbine AM in 2013. The IOA AMWG then published its report, A Method for Rating Amplitude Modulation in Wind Turbine Noise, on 9th August 2016. The IOA AMWG analysis method provides a robust way of identifying periods of AM during any 10-minute period to enable an AM rating level to be calculated specific to that period. The method allows for the identification of wind turbine AM whilst minimising the likelihood of non-wind turbine related sources affecting the results by restricting the assessed modulation frequency to that corresponding to the range of blade passing frequencies for the wind turbines under investigation and in frequency bands where such modulation occurs.

Subsequently the Wind Turbine AM Review carried out by WSP Parsons Brinkerhoff for the (then) Department of Energy and Climate Change (DECC) was published on 25th October 2016 (although the report itself is dated August 2016). Phase 1 of the report sets out its approach and methodology, and the Phase 2 report includes a literature review, its research into human response to AM, and recommends how excessive AM might be controlled through the use of a planning condition. The report includes recommendations of how AM should be addressed based on assessment according to the IOA AMWG method. It should be noted that the Wind Turbine AM Review report states that ‘the condition has been designed only for new planning applications, and applicability for use in Statutory Nuisance investigations on existing wind turbine sites has not been considered as part of this review’.

The report recommends a two tier approach whereby the first tier would be to seek a reduction in the depth and/or occurrence of AM with a rating level (according to the IOA AMWG method) ≥3 dB. Whether remedial action is required depends on the prevalence of any complaints, and how often AM rating levels ≥3 dB occur. The second tier is that if AM is deemed to be a significant issue, and if nothing can be done to reduce the level of AM, then a penalty scheme has been proposed whereby a penalty ranging from 3 dB (for a rating level of 3 dB) up to a maximum of 5 dB (for a rating level of 10 dB and above) should be added to the measured level before measured levels are compared with the relevant noise limits.

The IoA then held a one-day Wind Turbine Noise meeting in Birmingham on 7th December 2016 which included a lively debate on the implications of these two documents following field measurements and analysis having been undertaken by a number of acoustic consultancies involved in wind turbine noise.

How to Measure and Rate Amplitude Modulation using the IOA Metric

Noise measurements are carried out at the same location as required for normal wind farm noise measurements, but it is necessary to measure continuous audio or continuous one-third octave band LAeq,100ms­ data. Measuring continuous audio data is memory intensive and may have privacy implications. Measuring continuous one-third octave band LAeq,100ms­ data is, therefore, usually preferable, although it is useful to include such audio recordings so that the data can be auditioned if there are periods of interest. At HMPL, we usually supplement the 100 ms data with 2 minutes of audio recording (subject to the resident’s consent) at the beginning of each 10-minute period as a reasonable compromise between storage space and having the ability to listen to and review some periods of interest. Having 2-minute audio recordings also allows for the data to be analysed for tonal content which usually requires analysis of 2-minute audio samples. AM would usually have a much greater variation over a 10-minute period than tonal noise which is why continuous data is required.

The IOA AMWG method splits the modulation depth analysis into 3 bands; 50 to 200 Hz, 100 to 400 Hz, and 200 to 800 Hz. The AM rating levels in the 50 to 200 Hz and 200 to 800 Hz bands are compared against the ‘reference’ 100 to 400 Hz modulation depths. The frequency band giving the highest levels is used for further analysis. The summary of the IOA AMWG method performed on each band is outlined below:

• The input signal (a time series of band-limited, A-weighted, 1/3-octave Leq­ data in 100 millisecond samples) is split into blocks of 10 seconds;
• It is transformed to the frequency domain using a Fourier analysis to obtain a modulation spectrum;
• If a clear (prominent) peak is present at a rate expected from the turbines, a window around that frequency (and the next two harmonics) is selected (subject to some tests);
• An inversion Fourier transform is applied to the filtered spectrum to reconstruct a filtered time-series;
• The modulation depth in the filtered time-series is then determined
• A value for each 10-minute period is calculated from a combination of the 10-second modulation depths within that period.

How the resultant data is then used in determining the severity of any AM is not prescribed in the IOA AMNWG report or the DECC report. The DECC report is relatively clear that AM rating levels ≥3 dB can be considered to be excessive, but whether or not any remedial action needs to be taken is dependent on whether there are any complaints, and the amount of time that AM above the threshold occurs. Careful analysis of the results is, therefore, required to correlate the AM rating levels with wind speed and direction conditions on site, time of day, and periods of complaint so that the significance of any AM can be determined. It is envisaged by the authors of the DECC report that as more experience is acquired using the method, a significance rating matrix could be developed by Local Planning Authorities to determine what amount of AM may be acceptable.

Further AM Developments

There is currently a proposal for an international standard relating to AM to be included within the International Electrotechnical Commission (IEC) 61400 series of standards that deal with wind turbines. There is some multi-national interest for the addition of a second standard to accompany IEC 61400-11 Acoustic noise measurement techniques (see our summary article on the most recent version of this standard here) to deal with quantifying AM at residential distances from wind turbines or wind farms. This proposal is in its early stages and offers an interesting platform whereby AM could be discussed internationally, and should enable the views of the rest of the world to be taken into account. A significant problem in quantifying an unacceptable amount of AM is that, when listening tests have been carried out, the metric used to determine the levels of AM are often different, which means that results are not always comparable on a like-with-like basis. An international standard for the analysis of AM would never prescribe limits or penalties for AM as these are always country specific, but if an AM rating level can be determined using a common method it would enable more focused further human perception research.

Our Expertise and Input into Wind Farm and AM Research

Malcolm Hayes (Director) was on the Institute of Acoustics (IOA) noise working group that produced, A Good Practice Guide to the Application of ETUS-R-97 for the Assessment and Rating of Wind Turbine Noise (GPG) which he now chairs. He was also on the working group which produced the original ETSU-R-97 document.

Andy McKenzie (Director) was on the peer review panel for the IOA GPG, and drafted Supplementary Guidance Note 5, Post Completion Measurements, to the IOA GPG.

Rob Shepherd (Principal Consultant) is a member of the BSI committee EH/1/3 on residential and industrial noise as the Association of Noise Consultant’s representative on wind turbine noise.

Tom Levet (Senior Consultant) is a member of the IOA amplitude modulation working group which produced the Method for Rating Amplitude Modulation in Wind Turbine Noise document.

Robin Woodward (Senior Consultant) is a member of the British Standards IEC 61400-11 Shadow Committee.

By Rob Shepherd

If you'd like to know more about AM, it's analysis or our services, don't hesitate to get in contact with us using our 'Get in Touch' form or by giving us a call using our Contact page.