Download presentation
Presentation is loading. Please wait.
Published bySarah Eaton Modified over 9 years ago
1
Research Study on Wind Turbine Acoustics DRAFT March 7, 2014 Interim Results II for WNTAG
2
2 03.07.2014 RSG Interim Report II Interim Report II for WNTAG focuses on a comparison of sound metrics with sound modeling to help inform and synchronize pre-construction estimates with post-construction monitoring. CONTENTS New terms Review of data collection Sound monitoring metrics Pre-construction sound predictions Attended sound monitoring Statistical confidence
3
3 03.07.2014 RSG Review of data collection Four sites to date – -all in Massachusetts -all 1.5 MW or greater Five sound monitoring locations at each site -1/3 octave bands + other metrics at 100 ms to 1 s intervals -Type I sound monitors Infrasound monitoring at one location (inside and outside) at one site One LIDAR location at each site One 10-meter met tower at each site Turbine operating conditions collected by operator Over 120,000,000 data records logged Over 150 sound level, meteorological, operational, and observational variables
4
Sound monitoring metrics
5
5 03.07.2014 RSG New terms Site Location Background sound level vs ambient L90 LAf max (1-sec) L90 of the L90
6
6 03.07.2014 RSG Consideration of new sound monitoring metric for Turbine sound – L90 of Laf max(1-sec)
7
7 03.07.2014 RSG Consideration of new sound monitoring metric for Turbine sound – L90 of Laf max(1-sec)
8
8 03.07.2014 RSG Consideration of new sound monitoring metric for Turbine sound – L90 of Laf max(1-sec)
9
9 03.07.2014 RSG Sound monitoring metrics – Background sound
10
10 03.07.2014 RSG Sound monitoring metrics – Background sound
11
11 03.07.2014 RSG Sound monitoring metrics – Background sound
12
12 03.07.2014 RSG Background L90 - Variability
13
13 03.07.2014 RSG Effect of wind speed on L90 – wind shear Wind Shear -0.20.3 0.6 Wind shear exponents
14
14 03.07.2014 RSG Background L90 and Wind Speed are significantly correlated Slopes of 80-meter wind speed vs sound level for various methodologies
15
15 03.07.2014 RSG Wind speeds vary during any measurement period Example of a 10-minute period at one site, showing the frequency of occurrence of 0.5 m/s bins for 9 m/s average wind speed ws 90 1.3 standard deviations
16
Pre-construction sound predictions
17
17 03.07.2014 RSG Example of pre-construction modeling methodology for one site 370 meters downwind
18
18 03.07.2014 RSG Example of pre-construction modeling methodology for one site 370 meters downwind Slope of brown line in db/meter per second
19
19 03.07.2014 RSG Example of pre-construction modeling methodology for one site 370 meters downwind
20
20 03.07.2014 RSG Example of pre-construction modeling methodology for one site 370 meters downwind
21
21 03.07.2014 RSG Example of pre-construction modeling methodology for one site 370 meters downwind
22
22 03.07.2014 RSG Measured L90s of turbine sound levels
23
23 03.07.2014 RSG Perfect modeling of wind turbine sound
24
Attended sound monitoring
25
25 03.07.2014 RSG Filtering background sound
26
26 03.07.2014 RSG Filtering background sound
27
27 03.07.2014 RSG Filtering background sound
28
Statistical Confidence in Measurements
29
29 03.07.2014 RSG New terms Statistical Bias Accuracy Precision Confidence Interval Standard Deviation Standard Error
30
30 03.07.2014 RSG Comparing background to turbine-on measurements
31
31 03.07.2014 RSG Estimate means and confidence intervals
32
32 03.07.2014 RSG Estimate means and confidence intervals
33
33 03.07.2014 RSG Suggested strategy for using different metrics for background and turbine-on measurements
34
Conclusions
35
35 03.07.2014 RSG Some specific conclusions from the report Background sound levels vary by time of year, time of day, and day of week. Natural short-term variation is partly a function wind speed and wind shear Sound levels measured on the ground increase when 80 meter wind speed increases Wind shear variation is highest at night and at low wind speeds Background sound will contaminate measurements of wind turbine sound -Wind alone can have a significant effect -By definition, 90% of the turbine-on measurements have background levels that are higher than the L90 When measuring over five or 10 minutes, the wind speed exceeded 90 percent of the time is likely to be at a lower integer wind speed than the mean wind speed
36
36 03.07.2014 RSG More specific conclusions from the report Since L90 and wind speed are correlated, this means that the L90 is also likely to occur at a lower wind speed relative to the mean. Adjustments can be made to account for this. The 10 th percentile wind speed is a function of the mean and standard deviation of the measured wind speed over a period Considerations of sound metrics -Using L90 of Laf max (1-sec) for both background and turbine-on measurements -Improving predictability by establishing a turbine-only sound limit based on background measurements during pre- construction -Incorporating some type of statistical analysis to improve confidence in compliance measurement -Adjust turbine-on sound metric (if different from background metric) to account for higher background sound.
37
37 03.07.2014 RSG General conclusions Overall, real-world systems are dynamic. Methods developed should take into account likelihood that Conditions change during the measurement Conditions change over time Measurements including everything that produces sound in the environment Methods to measure and model sound will have biases Methods to measure and model sound will have variability
38
www.rsginc.com Contacts www.rsginc.com Contact Kenneth Kaliski, P.E., INCE Bd. Cert. Senior Director Ken.kaliski@rsginc.com 802-295-4999
Similar presentations
© 2024 SlidePlayer.com. Inc.
All rights reserved.