Download Noise Metrics and Regulations - Noise Control - Handout and more Exercises Noise Control in PDF only on Docsity!
4. NOISE METRICS AND
REGULATIONS
4.1 INTRODUCTION
Noise metrics are an attempt to emulate the manner in which humans respond to sound. They enable us to repeatably predict the impact of a given noise on the average person. They are extensively used to predict loudness, annoyance and potential for hear loss. Each government agency seems to have its own motivation and method for quantifying noise. The Federal Highway Administration (FHWA), Federal Aviation Administration (FAA), OSHA and EPA all have different methods for assessing noise. A variety of descriptors and calculation procedures are used. These methods attempt to quantify the complex characteristics of human hearing and human psychology. While they are all based on the decibel scale (dB), there is no agreement on a single best measure. Different procedures have been developed for different applications, such as aircraft, traffic, factory and community noise.
Objectives of this Section:
This section describes the various methods that are commonly employed and will help you to navigate through the alphabet soup of noise metrics. The reader will understand basic noise metrics, their application, calculation and limitations.
References:
Books:
- Acoustic Noise Measurements , Bruel and Kjaer, 1979.
- Handbook of Noise Measurements , A. P. Peterson, 9 th^ Edition, General Radio Inc., 1980.
- Engineering Noise Control , D.A. Bies and C. H. Hansen, Unwin Hyman Ltd, 1988.
- Industrial Noise Control , L. H. Bell and D. H. Bell, 2nd^ Edition, Marcel Dekker, 1994.
- Noise Control for Engineers , H. Lord, W. Gatley and H. Evensen, Robert Krieger Publishing Co., 1987.
- Effects of Noise on Man, K. Kryter, 2 nd^ Ed., Academic Press, 1985.
- Handbook of Acoustical Measurements and Noise Control, C.M. Harris, 3 rd^ Edition, McGraw-Hill, 1991.
Papers:
- H. Fletcher and W. Munson, “Loudness, its definition, measurement and calculation”, JASA, 5(2), October 1933, pp82-108.
- D. W. Robinson and R. Dadson, “A re-determination of the equal loudness relations for pure tones”, Bri J of Appl Phys, Vol 7, May 1956, pp 166-181.
- S.S. Stevens, “Perceived level of noise by Mark VII and Decibels(E)”, JASA, 51(2) part 2, Feb. 1972, pp 575-601.
- S.S. Stevens, “Procedure for calculating loudness, Mark VI”, JASA, 33(11), Nov 1961, pp 1577-1585.
- E. Zwicker, H. Fastl and C. Dallmayr, “Basic program for calculating the loudness of sounds from their 1/3 octave band spectra according to ISO532B”, Acustica vol 55, 1984, pp 63-67.
- R. Hellman and E. Zwicker, “Why can a decrease in dB(A) produce an increase in loudness?”, JASA, 82(5), November 1987, pp 1700-1705.
- T. J. Schultz, “Synthesis of social surveys on noise annoyance”, JASA 64(2), August 1978, pp377-405.
- Protective Noise Levels: Condensed Version of EPA Levels Document, Rep. EPA- 550/9-79-100, Nov. 1978. (Available from NTIS as PB82 138 827).
4.2 WEIGHTING NETWORKS
Weighting networks (implemented with electronic filters) are built into sound level meters to provide a meter response that tries to approximate the way the ear responds to the loudness of pure tones. These weighting curves are directly derived from the Fletcher/Munson equal loudness contours.
Figure 1 (ref. fig 1.12 lord, gatley and evenson)
Table 1. A, C, and D weighting correction values
Center Frequency Hz
A- Weighting Correction dB
C- Weighting Correction - dB
D- Weighting Correction- dB 10 -70.4 -14. 12.5 -63.4 -11. 16 -56.7 -8. 20 -50.5 -6. 25 -44.7 -4. 31.5 -39.4 -3. 40 -34.6 -2. 50 -30.2 -1.3 -12. 63 -26.2 -0.8 -10. 80 -22.5 -0.5 -9. 100 -19.1 -0.3 -7. 125 -16.1 -0.2 -5. 160 -13.4 -0.1 -4. 200 -10.9 0 -2. 250 -8.6 0 -1. 315 -6.6 0 -0. 400 -4.8 0 -0. 500 -3.2 0 -0. 630 -1.9 0 -0. 800 -0.8 0 -0. 1000 0 0 0 1250 0.6 0 2. 1600 1.0 -0.1 4. 2000 1.2 -0.2 7. 2500 1.3 -0.3 10. 3150 1.2 -0.5 11. 4000 1.0 -0.8 11. 5000 0.5 -1.3 9. 6300 -0.1 -2.0 7. 8000 -1.1 -3.0 5. 10000 -2.5 -4.4 3. 12500 -4.3 -6.2 -1. 16000 -6.6 -8. 20000 -9.3 -11.
Quantities which you can read directly from a typical, basic sound level meter include: LP = Overall unweighted sound pressure level [designated as dB(lin) or just dB] LA = Overall A-weighted SPL (dBA) LC = Overall C-weighted SPL (dBC)
4.3. LOUDNESS AND ANNOYANCE RATINGS FOR STEADY
NOISES:
Loudness or annoyance measures are not generally available on basic sound level meters, since they require some additional calculations or time averaging. They provide much more information than the overall sound pressure level (with or without frequency weighting).
Loudness level - (Stevens - Mark VI) This measure provides a quantitative measure of the overall loudness, as well as the relative contribution of each octave band to the overall loudness. It is useful for comparison purposes and gives important information for the cost effective application of noise control treatments. It was derived from empirical data with relatively flat spectra (no pure tones) and diffuse sound fields.
Loudness levels in each octave band are determined from Table 2.1. The composite loudness level L for all the octave bands is then:
Loudnessindexofthe thoctave band
Loudnessindexofloudestoctaveband
CompositeLoudnessLevel(sones) =. 7. 3
max
max
Si i
S
L S Si
Example Calculation:
Octave band center frequency - Hz 31 63 125 250 500 1000 2000 4000 8000 Octave band level – dB lin 76 72 70 75 80 74 65 65 66 Band loudness index Si 3.2 3.7 5.0 8.3 13.5 11.1 7.8 9.3 11. Ranking 9 8 7 5 1 3 6 4 2
Using the table that follows:
LoudnessLevel 89.8phons (usingcolumns 10 and 11)
. 7 13. 5. 3 73. 7 31. 56 sones
max
= × + × =
= å =
L
S
S (^) total Si
Stevens - Mark VII (ref 10) This is an improvement to Mark VI which uses 1/3 octave data and includes some effects of masking. (see ref. 2, General Radio Handbook for calculation details)
ISO532B – Zwicker Method This method is similar to the MarkVII method but also accounts for the upward spread of masking and can handle complex sounds with broadband and/or pure tone components. It uses 1/3 octave data and can account for frontal or diffuse sound fields. This seems to be the best method for quantifying annoyance of sound and is now an international standard. (see ISO532B standard for details)
PNL – Perceived Noise Level This is a similar procedure as Mark VI loudness, but uses equal noisiness contours. It is commonly applied to aircraft noise
EPNL – Effective Perceived Noise Level This is a refinement of PNL to include a correction for the noise duration and the presence of clearly audible discrete tones. It is used for aircraft noise (FAA) and involves relatively complicated calculations. (See ANSI S6.4-1973 for details)
NC curves – Noise Criterion Noise levels below 80 dBA are considered safe from a hearing loss perspective. However, they can still be highly annoying and interfere with the effective performance of occupational tasks or other activities. The Noise Criterion method, developed in 1957, rates the background levels in buildings and rooms. It is used to judge the appropriateness of the acoustic environment for various activities. The actual spectrum (octave band levels) is compared to standard NC curves (shown in Figure below). The highest NC level penetrated is the NC rating. This will be discussed further in the section on room acoustics.
RC, NCB :
A potential deficiency of the NC method is that it does not adequately rate the quality of the spectrum. An HVAC system may sound rumbly (low frequency sound) or hissy (high frequency sound), or both if the spectrum matches a particular NC contour. To improve on NC, a number of more recent, and more conservative room measures have been proposed. These include: RC (ROOM CRITERION), (Blazier, 1981) takes into account lower frequency (down to 16 Hz) and attempts to achieve better balance between low frequency (rumble) and high frequency (hiss) components. It is the preferred method of ASHRAE. NCB (NOISE CRITERION BALANCED), (Beranek, 1989). NCB also covers the octave bands from 16 through 8000 Hz. It allows significantly higher levels in the 16 and 31.5 Hz bands that does the RC method.
4.4 SPEECH INTERFERENCE
Interference with speech is one of the more negative consequences of excessive noise. Speech intereference causes frustatration, annoyance and irritation. When oral communication is disrupted, worker efficiency can suffer, and the potential for error due to miscommunication is increased. Several methods have been proposed to predict and quantify speech intelligibility including: A-weighted sound level, Speech Interference Level (SIL), Articulation Index (AI) and Speech Transmission Index (STI).
A-weighted sound level is the simplest method for predicting speech intelligibility. The background sound level is measured and a chart such as Figure 16.8 is used. This technique works best if the noise spectrum is flat, the noise levels are steady, and the acoustic environment is non-reverberant (reverberation time less than ~2 seconds).
Figure 3. (ref. Harris, Handbook of acoustical measurements and noise control)
Speech interference level (SIL, PSIL) The Speech Interference Level is intended to quantify the effectiveness of speech in the presence of noise. Numerically it is the numerical average of four octave band levels - 500, 1000, 2000 and 4000 Hz. A commonly used, related measure is the Preferred Speech Interference Level (PSIL) which uses three bands (500, 1000, and 2000 Hz).
PSIL = L^500 + L^1000 + L^2000 Equation #
Equivalent sound level - Leq
In normal occurrence, sound levels vary during the course of the day. Levels temporarily increase such as when a truck passes, when a dog barks, when an airplane flies over, when an air condition compressor turns on, or when a factory begins its production shift.. Equivalent Sound Level (LEQ ) is analogous to an average level and is defined as the hypothetical constant sound level over a period of time which results in the same overall sound energy as the actual time varying sound. Since sound energy is proportional to intensity, which is in turn proportional to the square of sound pressure:
P Pa
L
T
L dt T
L
P
T
dt P
P t T
L
REF
EQ A A
A REF
A EQ
where referencepressure 20 μ
A-weightedlevel(dBA) 0
10 log^11010
rmsA-weightedpressure 0 2
10 log
10
10
ú
ú û
ù
ê
ê ë
é = ò
ú
ú û
ù
ê
ê ë
é = ò
Equation 2
for discrete data samples:
å
å = ≅
≅
∆
å ∆
≅
n
i
L n
L P
n P t ni
L
t
ti
n
i P
P t T
L
EQ A REF
A EQ
REF
i EQ
1
(^11010) 2 10 log^10
(^2) ()
1
10 log^1
forconstant timeintervals (onemeasurementeveryhouristypical):
1 2
1 2 () 10 log
10
10
Equation 3
Day-Night Level - LDN
LDN is similar to LEQ but adds a 10 db penalty at night from 10 pm to 7 am. It is widely used in US to compensate for the increased undesirability of noise during sleep periods. EPA recommends a maximum residential level of 55 Ldn. For hourly measurements:
(7am- 10 pm) (10pm- 7 am)
10 log 10 ú
ú û
ù ê
ê ë
é å =
å
i
L
i
L LA A
DN (^) Equation 4
A steady noise of 48.6 dBA equates to 55 L (^) DN
Exceedance Level - LN
Defined as: the noise level which is exceeded N% of time during a day.
A value of 60 dB L 10 means that the sound level exceeds 60 dB for 10% of the day. This measure is commonly used for traffic noise measurement. It is also useful for separating fluctuating noise from steady noise. L 90 is a good measure of background noise L 50 is the median noise, which is not necessarily the same thing as LEQ (the mean) L 10 is a good measure of intermittent or intrusive noises, such as traffic, aircraft flyovers, barking dogs, etc.
4.6 HEARING DAMAGE RISK
In order to adequately prevent permanent hearing loss, we need a way to measure the severity of noise and correlate the noise level with risk of hearing damage. There is considerable disagreement over which criteria to use. It is agreed that in general, hearing damage is a function of noise level and exposure time. Figure 5 shows the percentage risk of developing hearing loss from sustained occupational noise.
Figure 5 Percentage risk of hearing loss from sustained occupational noise (Fig 2.6 LG&E)
Figure 6a below shows a compilation of published data (Beranek 1971, Burns and Robinson 1970) showing median hearing loss as a function of the percentage risk of incurring that loss for a specified exposure time and level. It is assumed that a person would be exposed to the stated level for about 1,900 hours during each year.
This principle says that hearing loss is proportional to the product of sound energy and exposure time:
exposure Hearing Loss∝ p^2 × T Equation 7
This implies that a 3 db increase in SPL is equivalent to (causes same hearing loss risk as) doubling the exposure time. European and Australian noise exposure standards are based on the equal energy principle (3 dB rule).
There is some justification for either 3 dB or 6 dB exchange rule. However, U.S. standards are based on a 5 dB rule, where a 5 db increase in SPL is assumed equivalent to (causes same hearing loss risk as) doubling the exposure time.
CRITERIA FOR CONTINUOUS NOISE EXPOSURE
Internationally, it has been agreed that 90 dBA is the maximum acceptable level for an 8 hour work day. This number represents a compromise between health concerns and economic constraints. However, according to Fig 5, this will cause hearing damage in approximately 25% of the population. To minimize the risk of hearing damage, 80 dBA or less is required. Higher levels are compromises between the cost of noise control and the risk of hearing damage and resulting compensation claims. 70 dBA exposure over 24 hours will protect 97% of people at all frequencies.
OSHA Noise Standard –Permissible Occupational Noise Levels (1978)
OSHA (Occupational Safety and Health Administration) Act of 1970 and standards developed in response to the Act in 1978 set maximum permissible levels and specify employer remedial action if levels are exceeded. Table 3 OSHA permissible noise exposure Level, dBA (slow)
Permissible Exposure (hours) 90 8 92 6 95 4 97 3 100 2 102 1. 105 1
115 .25 or less
If these levels are exceeded:
- When employees are subjected to sounds exceeding permissible exposure levels (noise dose >1.0), feasible administrative or engineering control shall be utilized.
- If such controls fail to reduce sound levels to within the permissible limits, personal protective equipment shall be provided and the proper use enforced.
- In all cases where the sound levels exceed values specified as permissible limits (> dBA or noise dose of 0.5), an effective hearing conservation program shall be administered for as long as the noise levels exceed those permitted by law. Hearing conservation programs include the following components: a) Exposure monitoring b) Employee notification c) Audiometric testing (baseline and annual) d) Employee training e) Hearing protection f) Record keeping
Additional action is required if an employee exhibits a “standard threshold shift” i.e. average of 10 dB or more in the 2K, 3K and 4K bands in either ear.
When noise levels vary with time, the total equivalent noise “dose” is either measured with a dosimeter, or calculated. Noise dose D for two or more periods at different levels (should never exceed 1) is calculated by:
2 (^90 )^5
totalpermittedexposuretimetoL
timeofexposureatsoundpressurelevelL
D= ... 1
N
N
2
2 1
1
N EQ
N
N
N
N
L
T
T
C
T
C
T
C
T
C
Equation 8
Only levels above 80dBA are considered in the calculation of noise dose. In addition, impulsive noises of greater than 140dB peak sound pressure level are not permitted.
4.7 COMMUNITY NOISE
EPA Guidelines (1974)
Public awareness of noise as a national problem in the United States was dramatically increased in 1970 with the establishment of the Occupational Safety and Health Administration and the extension of noise standards to virtually all American industry in 1971. OSHA’s chief concern was to protect against hearing loss due to excessive noise levels in the workplace. Soon thereafter, EPA published its “Report to the
Note: Explanation of identified level for hearing loss: the exposure period which results in hearing loss at the identified level is a period of 40 years.
- (^) Refers to energy rather than arithmetic averages
Reference: Information on Levels of Environmental Noise Requisite to Protect Public Health and Welfare with an Adequate Margin of Safety , U.S. Environmental Protection Agency, 550/9-74-004, March 1974. World Health Organization (1993)
In 1993, the World Health Organization (WHO) published recommended guidelines for protection against noise. These values are primarily oriented towards criteria such as sleep disturbance, annoyance, and speech interference. They are based on results from numerous laboratory and field studies and are very similar to the EPA guidelines. The WHO target values are:
- To protect the majority from being moderately annoyed, the noise level (L (^) EQ ) should not exceed 50 dB.
- To protect the majority of people from being seriously annoyed during daytime, the level (L (^) EQ ) from steady, continuous noise in outdoor living areas should not exceed 55 dB.
- At night, outdoor levels (L (^) EQ ) should not exceed 45 dB, so that the recommended level of 30 dB inside bedrooms for steady state continuous noise can be met with the windows open.
People react differently to the same noise source. What is annoying to one person, may not be noticeable to another. No matter how low the sound level, as long as it is audible, someone will object to it for one reason or another. Numerous studies have been done of the effect of noise from traffic and aircraft sources. A compilation of these studies shown in Figure 7, shows a surprising correlation between the measured noise level (measured in LDN ) and the % of people who are highly annoyed by that level. A curve fit of this data results in an equation:
Equation 9
Using this equation, a LDN value of 55 would result in an estimated 4% of the population being highly annoyed.
Figure 7. Summary of annoyance data from eleven surveys that show close agreement (ref. T. J. Schultz, Synthesis of social surveys on noise annoyance, Journal of the Acoustical Society of America, Vol. 64(2), August 1978)
%Highly Annoyed = 0.8553 L DN − 0 0401. L (^) DN^2 +0 00047. LDN^3
Day-nite level (L (^) DN ) % Highly Annoyed (by equation 9) 50 1. 55 3. 60 8.
Various State and Local Regulations State regulations for noise are unusual but one state that has taken a leadership role in this area is Connecticut. Permissible levels are specified for community and environmental noise. The most stringent level is for residential areas at night (45 dBA). Connecticut regulations also define limits for impulse noise, prominent pure tones , ultrasonic and infrasonic noise and the presence of high ambient background levels. The complete text of the Connecticut regulation is included in the Bell and Bell text. Where no state rules are applicable, all municipal governments will generally have at least some sort of “nuisance law” which can be applied to noise which disturbs the peace (i.e. generates a complaint from a citizen). However, these laws are highly subjective in application, and are difficult to enforce and adjudicate. To avoid these problems, many municipalities have well defined, quantified noise ordinances. A representative exmple is the noise standard for New York City. Noise limits are in terms of Leq, measured over a 1 hour period. Consideration is given to the land use zoning, with higher levels permitted in industrial and commercial areas, and during daylight hours (7am-10pm). A relatively simple sound level meter, with A-weighting, and equipped to perform a one hour average, is needed to acquire the data.
The noise regulations for the city of Boston are also shown in Bell and Bell. In this case, maximum levels are specified in octave bands from 31.5 to 8000 Hz. This requires more data to be recorded and a more elaborate (and expensive) sound level meter. The overall limits for Boston (if the individual octave band levels are mathematically combined)of 60 dBA for daytime and 50 dBA for nightime in residential areas, are equivalent to the New York values. Both cities exceed the EPA guidelines by 5 dB, probably due to the traffic and economic realities of big city life. Smaller, more rural communities generally place a higher value on quality of life and may be less tolerant of noise. This attitude is epitomized by the Canadian noise standard which specifies that any noise source which can be heard over the background level of traffic is too loud. One last example is cited, that of Ferguson Township, PA, a mixed rural, light industrial, and residential area which includes part of State College. No daytime noise limits are specified. Nightime limits (from 7 pm - 7 am) are 55 dBA at the boundaries of residential zones and 62 dBA at the boundaries of commercial zones.
