Confirming Tonality at Residences Influenced by Wind Turbines

For 5 years, since the start-up of an array of 140 wind turbines, residents have filed complaints with the Ontario Ministry of the Environment (the regulator), and K2 Wind (the operator). Residents complained that the turbines produce a tonal sound, and that the irritation this produced impacted their sleep, their health, and the enjoyment of their property. To confirm tonality from the wind turbines, this research examined over 200 data examples from two families. These families collected data by two independent methods, a continuously recording system, and by making selected audio recordings. The recorded data was correlated with the wind turbine operational performance, and local weather conditions. The correlated data was analyzed for tonality using international standard evaluation methods. The analysis confirmed over 84% correlation between complaints of irritating conditions, and tonality from 5 dB to over 20 dB. The research also identified deviation between the recommended method for assessing wind turbine tonality of an expert group panel for the industry and the method for compliance monitoring now prescribed by regulations. The deviation can incorrectly reduce tonality calculated to significantly below the actual tonality. Finally, the results showed that the assumption of the regulator to only require assessment of compliance when the resident was downwind of the nearest wind turbine was incorrect. Most complaints arose from other wind directions. Neither was the regulator’s assumption correct that curtailing the wind turbine operation to continue operating at only partially reduced outputs would give remediation. The research concludes that tonality arises consistent with the wind turbine operation, identifying a critical need to revise the practices to prevent chronic irritation. DOI: 10.14302/issn.2642-3146.jec-20-3359 Corresponding author: William K.G. Palmer, Independent Researcher, TRI-LEA-EM, 76 Sideroad 33-34, RR 5, Paisley, ON N0G 2N0, Canada. Email: trileaem@bmts.com Running title: Confirming Tonality at Residences Influenced by Wind Turbines


Introduction
Coincident with the 2015 start-up of a wind power development 1  In 2018-2019 an acoustic audit was performed for the regulator at each residence. Residents identified specific times that the conditions were most irritating.
The intent of this paper is to provide an evaluation of the tonality conditions occurring at times identified by the residents during the acoustic audit. Rigorous compliance with international standards and reference to recent peer reviewed literature on tonality, is included to assist regulators in assessing the conditions that impact the citizens, and to provide a basis for revision of compliance protocols. The hypothesis tested was that tonality caused by the wind turbines as identified in international standards exists as a chronic condition, causing irritation impacting health, and reducing enjoyment of normal use of property.

Experimental Procedure
1) Assessment was made of the data collected by the regulator during initial acoustical assessments to determine indications of tonality they showed.
2) Tests were conducted to determine the optimum locations of microphones at the first residence to permit monitoring data by the two channel continuously recording system to ensure the data collected would represent free field conditions. The objective was to select monitoring sites that would permit recording data during all environmental conditions. Free field locations suffer from the problem that rainfall or snow pellets striking the windscreen can influence the recordings. It was necessary to show that sheltered sites would provide unimpeded acoustic access to the source of the sound, without introducing artifacts.

Results
A summary of the main results for each of the 5 sub-sections outlined in the Experimental Procedure is given in Table 1.  Figure 1 occurring at times the wind turbine output was "curtailed" (operating at less than capable output due to electrical system lack of need.) Figure 1 shows the uncalibrated output of an Audacity spectrum analysis for two ministry  1) Analysis of the initial data collection from the SAM Scribe identified a concern in obtaining data during all weather conditions. The resident was required to cover, or bring the microphones indoors prior to rain to protect the microphones, and as a result data was not available for some time periods. A test was conducted at a third home in Niagara Region, using  Both microphones are visible in Figure 6. Figure 7 shows that while the sound pressure levels    1) The Simplified Tone-to-Noise Ratio Method a) Using the example of one case from the first residence, the simplified Tone-to-Noise Ratio method used to assess tonality for the potential cases identified as of tonal concern followed these steps:

4)
i) The sound recording from microphone  c) The FFT display was then used to "zoom-in" on the peak at 461 Hz as shown in Figure 17. with no tonality present.
v) The 8 remaining times noted by the residents as "not tonal" but by descriptive words such as thunderous, whooing, crashing, like an airport, or like a train were also analyzed. None of the samples that were identified as being "not tonal" were determined to be tonal.
c) The strong correlation (84%) between identification of tonality by the residents and that determined to be over 5 dB by the Tone-to-Noise Ratio (TNR) method as well as the fact that none of the samples identified as  2) The report discusses the determination of tone   The IEC 61400-11 Ed 2.1 specifies in Sub-Section 8.5.1 "General Methodology" of Section 8.5 "Tonality" that the narrowband analysis Frequency  Table 4. The assessment showed that as a typical tonal calculation for a 12 -10 second set of samples transitions from an FFT window width of 0.5 Hz to 5 Hz, it can render a case meeting all the criteria of being tonal to one being not tonal.

Discussion
Early work by Kryter, 24   • The evidence is sufficient to establish a causal relationship between exposure to wind turbine noise and annoyance.
• There is limited evidence to establish causal relationship between wind turbine noise and sleep disturbance.
• The evidence suggests a lack of causality between wind turbine noise and hearing loss.

Principal Findings
• This paper demonstrated acceptable readings from a microphone located on the gable end of a home that did not need to be taken out of service during rain or snow to allow continuous recording of conditions, as an alternative to a free-field location which is not suitable for continuous recordings in adverse weather which can render data unusable.
• This paper showed that analysis of the sound files collected at the homes of two resident families confirmed high correlation between times the residents described as tonal and the presence of tonality by a graphical method of comparing the tonal peak to the magnitude of the sound outside the critical bandwidth centered on the tonal peak.
• The paper showed that regulations which require analysis only of situations when a resident is downwind of the nearest wind turbines, or when wind turbines are at high power will miss the majority of irritating cases.
• The paper showed that the majority of irritating tonal cases did not occur when the resident was downwind of the nearest wind turbines.
• The paper shows that "curtailing" the wind turbine operation (partially reducing the output) may result in enhanced tonality.
•   The difference in definitions means that as the sample width gets smaller, reducing from 5Hz to 2 Hz or smaller as recommended by current research, an error is introduced into the masking noise bandwidth, and thus into the tonality • Some regulatory documentation based on the earlier IEC-61400-11 Ed. 2.1 version (as in Ontario) requires that the narrow band analysis use a sample width of 2 or 5 Hz for tonal cases below 2000 Hz.
Current research recommends a sample width of no larger than 0.25% of the tonal frequency. For the tonality present in this study in the order of 450 Hz, that would require a sample width in the order of 1 Hz or smaller. The paper demonstrates the effect of the change is that compliance with the regulations based on IEC 61400-11 Ed 2.1 (as in Ontario) can reduce cases that were clearly tonal by a listening test when using smaller sample windows to "not-tonal" when using the window width required by the regulation.

Conclusion
This research has demonstrated that tonality as defined by international documentation is present at the site under investigation 1 (K2 Wind in Ontario) as a chronic condition, and has been for about 5 years without abatement. Tonality is also present at a different location 34 (Armow Wind) employing the same variety of wind turbines. Tonality was present at some times on each day the wind turbines were operating when the capability was higher than about 50% even when the turbines were curtailed (operating at less than forecast output), and data showed tonality was continually present on some days for periods of up to 8 hours. This

Funding
This research project was funded by the investigator and received no external funding. It is noted that the instrumentation purchased by the residents to collect the data was at their expense, and is their personal property.