Presentation Palmer ASA Seattle May_2011

8/6/2019 Presentation Palmer ASA Seattle May_2011

1/26

Collecting Data on Wind Turbine

Sound to Identify Causes ofIdentified Concernsor

Wind Turbine Noise - Clues to the Mystery of

Why People are Hurting

William K.G. (Bill) Palmer P. Eng.

[email protected], RR 5, Paisley, ON N0G 2N0

Canada

Presented to the Acoustical Society of America

- 161st Meeting


2/26

Outline of Presentation People are hurting, but why?

This paper describes a method for

collection of reliable data at wind turbine

approved locations where sound level is

predicted to not exceed 40 dBA for wind

speeds up to 6m/s and at control locations

2 to 10 times the Ontario approved distance

of 550 metres (1800 ft or 0.34 of a mile)

Analysis of the data suggests causes for the

hurt, and ongoing investigation shows

evidence continues to illuminate the truth


3/26

Data Collection Readings taken at a series of locations (10 km

radius from centre) within 2 hour window Same terrain, same general proximity toforested areas, same weather, fundamentallysame wind speed. Difference is proximity ofturbines.

Recorded time, wind speed, turbine output(IESO), weather conditions (wind direction,pressure, temperature), dBA and dBC soundlevels.

Performed a minimum 30 second digitalrecording using M-Audio output to Audacityprogram on iBook G4 at 44,100 Hz sample rate

Over 250 samples taken in all seasons, from

March 2010 to March 2011 (and continuing)


4/26

Typical Monitoring Setup


5/26

Data Processing Listen to ensure no extraneous sources of

noise (road traffic, birds, dogs, wind)

Plot Spectrum (Hanning Window) using FFTAnalysis of 30 second window usingAudacity, into 16,384 bins from 2.6 to 22 k Hz


6/26

Plot FFTAnalysis & Export


7/26

Background - No Turbines Operating


8/26

No Turbines - A&C Weighting Only


9/26

Turbines Operating - Low Power


10/26

Octave Analysis - Turbines Low Power


11/26

Comparing No Turbine Case to

Case with Turbines at 0%

No turbine case shows all monitoring sitesclosely match each other except forexplainable differences (flies, birds, roadtraffic)

Case with turbines just connected to grid butat very low power shows sound levelincreases at approved locations by 15 to 20dB at all octaves up to 1000 Hz compared to

mean base case. Overall dBA at controllocation unchanged (or decreased) while atapproved location increases over 15 dB.

Only at octaves of 4000 and 8000 Hz is

situation relatively unchanged


12/26

Turbines at 25% Output


13/26

Octave Analysis - Turbines 25%


14/26

Turbines at 88% Output


15/26

Octave Analysis - Turbines 88%


16/26

Observations from Comparisons

With wind speed at 8m/s (Beaufort Force 5,

fresh breeze) dBA sound level at controllocation rises to be roughly the same at atapproved locations when wind speed thereis 1m/s, and turbines still at 0% output.

Meantime, at approved locations sound levelunder these conditions has increased by 25 to30 dB at all octaves up to 1000 Hz.

Even at zero power, C weighted sound at

approved locations is 60 dBC, while controllocation is about 44 dBC and 32 dBA ( 28dB)

At high power, C weighted sound atapproved locations is 75 dBC, while control

location is about 55 dBC and 41 dBA (


17/26

Shifting Frequency in Modulation At the 4th International Wind Turbine Noise

meeting in Rome, in April, several presenters(Carlo di Napoli & Sidney Xue) identified thatdominant frequency as well as sound level canchange from low to high (or vice versa) as

angle of attack changes from wind impingingon the blade. They suggested this can occurfrom microclimate wind shifts, or from changein speed across the rotor due to wind shear

Review of the data shows this evidence canbe detected, which increases the audiblepattern of the turbine - just like variablefrequency police sirens are used to increaserecognition and response instead of a steady

horn


18/26

Effect of Modulation - Frequency Shifts


19/26

Review of Proponent Monitoring

Resident had complained of noise at night,unable to sleep, other adverse impacts

Acoustical consultant gathered over 150 days of

data. Report looked at high noise conditions

identified road noise, birds, wind sound, andlawnmower, but could not detect wind turbines.

Lawnmowers at night? Lets look again. Chose

midnight readings, sorted by Leq.

Eliminated ~ 40 days with Leq > L90 by more

than 6 dB, indicative of traffic, or other transient


20/26

Proponent Monitoring


21/26

SMI Nov 2010

0450

Calculator 53.4 dBA,

66.9 dBC

Meter says 43 dBA

3 m/s @2.5m


22/26

CSK - May 2110 0411Note Cyclic Pattern - Dishwasher?

Calculator42 dBA,

54.1 dBC

Meter says

37.5 dBA,52.1 dBC

1.5 m/s

@2.5 m


23/26

SCH Feb 1511 2351Closest Turbine Shutdown

Calculator

45.1 dBA,

58.7 dBC

Meter says

41.5 to 45

dBA

0.5 m/s

@2.5 m


24/26

Conclusions projecting simple dBA sound levels from

wind turbines and performing dBA auditing ofsound levels does not protect citizens

a simple, repeatable method of monitoring

sound and performing octave band analysishas been demonstrated, which identifies

special characteristics of wind turbine sound

approved locations are 15 to 20 dB above

control locations for all octaves up to 2000 Hzunder all conditions - calm or stormy, when

wind turbines operating

cyclical sound pattern and frequency shift

makes turbines articularl noticeable


25/26

Acknowledgements

John Coulter for loan of Knowles microphone

and M-Audio Fast Track USB interface

Werner Richarz for Excel octave band calculator

Carlo di Napoli and Sidney Xue for raising

frequency shift as an issue George Kamperman, Rick James, and Harvey

Wrightman for review and comments on draft

The people suffering effects of wind turbines who

permitted recordings taken at their homes (CS,VS, NS, JH, HF, JT, GB, SJ, KA, TW and others)

The encouragement of countless friends and

those suffering from the effects of wind turbines


26/26

Presentation Palmer ASA Seattle May_2011

Documents