noise control handbook
TRANSCRIPT
Environmental Noise Control
ii
TABLE OF CONTENTS
Page
INTRODUCTION………………………………………………………………………. 1
DEFINITIONS………………………………………………………………………….. 1
Sound and Noise ……………………………………………………………………… 1
Sound Waves…………………………………………………………………………... 1
Speed of Sound………………………………………………………………………… 2
Wavelength and Frequency…………………………………………………………... 3
Noise Spectrum………………………………………………………………………… 5
Octave Bands…………………………………………………………………………... 8
Decibel and A-Weighted Decibel (dBA) Scale……………………………………… 10
Loudness………………………………………………………………………………... 12
Sound Pressure Level (SPL) and Sound Power Level (PWL)……………………. 14
BASIC CALCULATIONS……………………………………………………………... 17
Calculating Sound Power from Sound Pressure…………………………………… 17
Calculating the Total PWL for a Single Noise Source……………………………... 19
A-Weighting the PWL of a Single Noise Source……………………………………. 19
Calculating the Total PWL of Numerous Noise Sources……………………….…. 20
SOURCE-PATH-RECEIVER…………………………………………………………. 23
Source Specifics……………………………………………………………………….. 23
Path Specifics………………………………………………………………………….. 25
Receiver Specifics……………………………………………………………………... 34
ACOUSTIC MATERIALS……………………………………………………………... 38
Sound Absorbing Materials…………………………………………………………… 38
Transmission Loss or Barrier Materials……………………………………………… 39
Resonator-Type Materials…………………………………………………………….. 40
Damping Materials…………………………………………………………………….. 41
Vibration Isolators……………………………………………………………………… 41
iii
TABLE OF CONTENTS – CONT’DPage
ATTENUATION………………………………………………………………………… 42
Buffers…………………………………………………………………………………… 42
Natural Barriers………………………………………………………………………… 42
Barriers………………………………………………………………………………….. 42
Acoustical Enclosures…………………………………………………………………. 43
Acoustical Buildings……………………………………………………………………. 44
Silencers………………………………………………………………………………… 46
Acoustic Plenums……………………………………………………………………… 49
Acoustic Louvers……………………………………………………………………….. 50
Acoustic Lagging……………………………………………………………………….. 51
NOISE CONTROL APPLICATIONS………………………………………………… 51
ATCO’s Acoustic Assemblies………………………………………………………… 51
ATCO’s Balanced Approach………………………………………………………….. 53
Testing and Guarantees………………………………………………………………. 58
USEFUL SOURCES………………………………………………………………… 61
iv
FIGURES
PageFigure 1: Behavior of Sound Waves…………………………………………………. 2
Figure 2: Wavelength………………………………………………………………….. 4
Figure 3: Wavelength and Frequency……………………………………………….. 5
Figure 4: Example of a Noise Level Spectrum……………………………………… 5
Figure 5: Discrete Frequency (Tonal) Noise………………………………………... 6
Figure 6: Broad Band Noise………………………………………………………….. 7
Figure 7: Impulsive Noise……………………………………………………………... 8
Figure 8: Narrow Band, One-Third Octave Band and Octave Band……………... 9
Figure 9: Comparison Between the Pascal and Decibel Scales………………….. 10
Figure 10: A, B, C and D Weighting Networks..……………………………………… 12
Figure 11: Doubling Sound Pressure Adds 3 dB…………………………………….. 13
Figure 12: Equal Loudness Contours…………………………………………………. 14
Figure 13: Comparison of Sound Power (PWL or Lw) and Sound Pressure (SPLor Lp)………………………………………………………………………….
18
Figure 14: Structure Borne Noise……………………………………………………… 23
Figure 15: Near Field and Far Field…………………………………………………… 26
Figure 16: Sound Intensity…………………………………….……………………….. 28
Figure 17: Sound Pressure Decreases 6 dB for Each Doubling of Distance……... 29
Figure 18: Sound Propagation from a Line Source………………………………….. 30
Figure 19: 3 dB Near Field and 6 dB Far Field Guideline for a Point Source…….. 31
Figure 20: What Happens When Sound Waves Encounter an Obstacle…………. 32
Figure 21: Refraction of Sound………………………………………………………… 33
Figure 22: Equivalent Continuous Sound Pressure Level (Leq)…………………….. 35
Figure 23: Common Noise Level Criteria Used by Regulators…………………….. 36
Figure 24: Transmission Loss (TL) for Two Walls…………………………………… 39
Figure 25: Example of Parallel Baffles………………………………………………... 47
Figure 26: Example of an Absorptive-Reactive Silencer……………………………. 49
Figure 27: Example of an Acoustic Plenum………………………………………….. 50
v
FIGURES – CON’T
Page
Figure 28: Example of an Acoustic Louver…………………………………………… 51
Figure 29: Example of a Noise Management Assembly………………………….. 52
Figure 30: Noise Contour Levels at a Power Plant Before Acoustic Treatment….. 54
Figure 31 Noise Contour Levels at a Power Plant After Acoustic Treatment……. 55
Figure 32 Example of ATCO’s Balanced Approach………………………………… 57
Figure 33 Sample Acoustical Test……………………………………………………. 59
vi
TABLESPage
Table 1: Relationship Between Sound Power (PWL or Lw) and SoundPressure (SPL or L p)………………………………………………... 16
Table 2: Examples of Sound Power Levels for Select Equipment byOctave Band Frequency…………………………………………….. 19
Table 3: Sampling of Noise from Sources at a Peaking Power Plant……. 21
Table 4: Table Method for Adding or Subtracting Decibels……………….. 22
Table 5: Correction for Background Noise…………………………………... 25
Table 6: Examples of Community Noise Guidelines……………………….. 36
Table 7: STC Ratings and Their Relationship to Sound ProofingProperties….…………………………………………………………. 45
1
ENVIRONMENTAL NOISE CONTROL
IINNTTRROODDUUCCTTIIOONNThe objective of environmental noise control is to improve the acoustic environment in a
community by reducing noise levels. Noise from industrial operations can affect
neighboring residential areas, ranging from intolerable noise levels to structural
vibrations. Well-planned noise control can eliminate a major component of an industrial
site’s impact on its surrounding environment. Noise control is based on what we know
about how sound behaves. For this reason, our look at some of the fundamentals of
environmental noise control begins with basic descriptions of sound and noise, acoustic
materials, and attenuation.
DDEEFFIINNIITTIIOONNSS
SOUND AND NOISE
Noise is usually defined as annoying or unwanted sound. Sound may be defined as
any pressure variation (in air, water or other medium) that the human ear can detect.
A barometer measures pressure changes in air. However, the arrival of a warm or cold
front is too slow and the changes too gradual to be heard and, hence, called sound.
The human ear hears the rapid changes in air pressure that barometers can’t
measure—changes that are at least 20 times per second. Pressure changes are
caused by the action of a vibrating object—such as a turbine casing—on the
surrounding air.
SOUND WAVES
Pressure variations (sound energy) travel through air or other elastic media (such as
water) in the form of sound waves from the sound source to the receptor (microphone,
listener’s ears). When a solid object hits the air and does so repeatedly—as a vibrating
2
object does—the air alternately compresses and expands around it and waves of lower
and higher pressure are sent out in all directions from the object. What we sometimes
feel in our ears, and express as sound, is the change from the lower to higher pressure.
Figure 1: Behavior of Sound Waves
Sound vibrations alternately compress and expand air in front of the loudspeaker cone,moving away in the form of a sound wave.
SPEED OF SOUND
The speed at which sound travels varies with the medium. In air, a familiar rule applies.
Do you recall counting three (3) seconds per kilometer (five (5) seconds per mile) every
time you saw lightning to the time you heard thunder? The time lapse corresponds to
the speed of sound in air of 1,238 kilometers (770 miles) per hour. For purposes of
sound measurement, the speed of sound is expressed as 340 meters (372 yards) per
second (at sea level and 15° Celsius).
3
WAVELENGTH AND FREQUENCY
The number of pressure changes per second is called the frequency of the sound.
Units of frequency used to be given in cycles per second, but now they are called Hertz
(Hz), to honor H.R. Hertz, the physicist who discovered electromagnetic waves. One
cycle of pressure change is called the period. The period is also called the reciprocal
of the frequency and is given as follows:
Period (T) = 1
Frequency
Knowing the speed and frequency of a sound allows the calculation of its wavelength.
A wavelength is the distance a sound wave travels in the time it takes to complete one
cycle or period.
Wavelength (λ) = Speed of Sound ( c )
Frequency (Hz)
4
Figure 2: Wavelength
When designing an acoustical solution to industrial noise, it is important to know the
wavelength of the different frequencies. In general, the object in the sound path
must be larger than one wavelength to significantly disturb the sound. At 20 Hz, a
wavelength is about 17 meters (56 feet), so an object must be larger than 17 meters
wide and high to block the sound waves. At 20,000 Hz, the wavelength shortens to 1.7
centimeters (.7 inches). Low frequency noises have long wavelengths and high
frequency noises have short ones. The longer wavelength of a low frequency sound
allows it to slip easily around or over barriers.
5
Figure 3: Wavelength and Frequency
NOISE SPECTRUM
Most sound is made up of a number of frequencies just as light is made up of different
colors. A color spectrum results when light passes through a prism. A sound or noise
spectrum is produced when sound is passed through a spectrum analyzer.
Figure 4: Example of a Noise Level Spectrum
6
Two types of noise exist: steady noise and non-steady noise. Steady noise with audible
discrete tones is called discrete frequency noise and is the most common noise found
in industry. This type of noise has the characteristic of pure tones over a number of
frequencies. Discrete frequency noise is caused by rotating parts of machines such as
fans, internal combustion engines, transformers and pumps.
Figure 5: Discrete Frequency (Tonal) Noise
The second most common form of industrial noise is called broad band noise. Broad
band noise is steady noise without discrete frequency tones. Sounds are of longer
duration and vary little over time. However, acoustical energy may be heavily
concentrated in one or more areas of the spectrum. Large gas turbines emit peak noise
levels in the lower frequencies. This is called pink noise and is analogous to the pink
and red light at the lower frequencies of the color spectrum. If the noise has
frequencies evenly distributed throughout the audible range, white noise results.
7
Figure 6: Broad Band Noise
The noise levels shown in Fig. 6 were emitted by the engine exhaust of a Solar Mars Centaur
40S.
Other industrial noises are non-steady and consist of fluctuating noise (noise that
doesn’t remain at any constant level over a given period of time), intermittent noise
(noise that returns to the ambient or background level), and, more commonly,
impulsive noise (sounds of short duration with high peak pressures). Peak pressures
rise at least 40 dB in 0.5 seconds.1
1 Henning E. Von Gierke and Charles W. Nixon, “Damage Risk Criteria for Hearing and Human Body Vibration,”in Noise and Vibration Control Engineering: Principles and Applications. Leo L. Beranke and Istaván L. Vér, eds.New York.: John Wiley & Sons, Inc., p. 595.
8
Figure 7: Impulsive Noise
OCTAVE BANDS
Frequencies are divided into octaves, just like octaves on a piano. An octave band is
defined as a range of frequencies extending from one frequency to exactly double that
frequency. For example, the 1000 Hz octave band is centered at 1000 Hz and extends
from 707 Hz to 1414 Hz. Nine octave bands are most often used when measuring
sound.
Most Commonly Used Octave Bands in Industrial Noise Studies
31.5 Hz 63 Hz 125 Hz 250 Hz 500 Hz 1000 Hz 2000 Hz 4000 Hz 8000 Hz
When analyzing noise at an industrial site, a noise spectrum is studied. However, it is
not practical to examine the acoustic energy generated at every frequency at the same
time – this would create an enormous amount of data. Instead, the frequency range is
apportioned into a set of broader ranges, each containing lesser amounts of detail.
Examples of the three most common types of frequency analyses are narrow band,
one-third octave band and the octave band.
9
Figure 8: Narrow Band, One-Third Octave Band and Octave Band
For most industrial noise analysis, the octave band provides a sufficient level of detail.
Occasionally, a finer breakdown than an octave band is required, particularly when the
noise emitted is tonal. Tonal or discrete frequency sounds are characterized by spikes
of high energy at specific frequencies in an otherwise continuous noise spectrum. To
pinpoint these energy spikes, a more detailed noise analysis (using one-third octave
band) is undertaken. For even greater accuracy, a narrow band analysis over specified
narrow frequency ranges can be performed.
The frequency of a sound produces its distinctive tone. The rumble of the lowest notes
of the largest pipe organ has a low frequency, while a flute produces a high frequency
tone. Machines like gas turbines generate both low and high frequency sounds. Some
sources don’t cause various frequencies of sound. Instead, they generate a single
frequency or pure tone.
10
DECIBEL (DB) AND A-WEIGHTED DECIBEL (DBA) SCALE
The size or amplitude of pressure changes is measured in decibels or dB. The
weakest sound the human ear can hear has an amplitude of around 20 millionths of a
Pascal (20µPa) – the scale used to measure barometric pressure. A pressure change
of 20µPa is equivalent to 5 billion times less than normal atmospheric pressure.
Because the range of sound pressures in a typical room is so huge, using the Pascal
scale to measure noise would be close to impossible. The decibel scale was devised to
make calculations of noise levels manageable.
The decibel (dB) is a unit of logarithmic measure, which uses 2 x 10 –5 Pa as the
starting point of zero (0) dB. Zero dB or 2 x 10 –5 Pa is the lowest pressure a young
adult can detect of a pure tone at 1000 Hz. Most continuous noise sources emit sound
pressure levels between 0 to 150 dB. A level of 150 dB is equivalent to a jet aircraft at
take off. Noise levels over 150 dB can occur. For example, a blowdown vent opening
can produce sounds of 170 dB, while the space shuttle is recorded at 180 dB.
Figure 9: Comparison Between the Pascal and Decibel Scales
Sound Pressure (Pascals) Sound Pressure Level (Decibels)
Jet Engine (25 m)
Rock Concert
Heavy Truck
ConversationalSpeech
Unsilenced Turbine Inlet (3 m)Unsilenced Turbine Exhaust (3 m)
Inside Turbine EnclosureCooling Tower (3 m)Transformers (3 m)
HRSGInside Powerhouse Building
Lube Oil Cooler (3 m)
Inside Control Room
Equipment ExamplesExamples
11
The decibel scale is a closer approximation to the sounds heard by the human ear than
the Pascal scale, because the human ear is able to react to exponential changes in
sound pressure. However, the decibel scale still doesn’t replicate what the human ear
actually hears. This is because the human ear is more sensitive to sound at
frequencies between 1000 and 5000 Hz and less sensitive to higher and lower
frequency sounds. To quantify the sensitivity of humans to sound the A-weighted
decibel or dBA scale (also written dB(A)) was created. A correction factor was devised
to change unweighted decibels (dB), also known as the linear scale, to A-weighted
decibels (dBA).
For purposes of noise control, both the dB and dBA scale can be used interchangeably.
Sometimes it is necessary to convert from the dB to dBA scale and vice versa. For
example, a manufacturer might provide the noise level of a machine in dB, whereas the
community noise requirement is stated for dBA. In this case, initial calculations of the
noise level might be made in dB, then converted to dBA.
There exist three additional weighting networks — B, C, and D — which are either
used in special circumstances or are obsolete. When low frequency noise is of
concern, C weightings are used because they attenuate low frequencies much less than
the other weightings. D weightings are used when very high frequencies, like those
emitted from jet engines, need to be attenuated. The B weightings, emphasizing middle
frequencies, are no longer in use.
Example: A 100 dB sound in the 31.5 Hz band has a correction factor of –39.4. Subtract 39.4 from 100 dB(i.e., 100 dB – 39.4 = 60.6 dBA). The answer—60.6 dBA—is how “loud” the 100 dB sound isperceived by the human ear in the 31.5 Hz band. By contrast, the same 100 dB sound is perceived bythe human ear exactly as 100 dBA when frequencies are in the 1000 Hz band (i.e., 100 dB – 0 = 100dBA).
12
Figure 10: A, B, C and D Weighting Networks
Frequency Curve A dB Curve B dB Curve C dB Curve D dB
10 -70.4 -38.2 -14.3 -26.512.5 -63.4 -33.2 -11.2 -24.516 -56.7 -28.5 -8.5 -22.520 -50.5 -24.2 -6.2 -20.525 -44.7 -20.4 -4.4 -18.5
31.5 -39.4 -17.1 -3 -16.540 -34.6 -14.2 -2 -14.550 -30.2 -11.6 -1.3 -12.563 -26.2 -9.3 -0.8 -1180 -22.5 -7.4 -0.5 -9
100 -19.1 -5.6 -0.3 -7.5125 -16.1 -4.2 -0.2 -6160 -13.4 -3 -0.1 -4.5200 -10.9 -2 0 -3250 -8.6 -1.3 0 -2315 -6.6 -0.8 0 -1400 -4.8 -0.5 0 -0.5500 -3.2 -0.3 0 0630 -1.9 -0.1 0 0800 -0.8 0 0 01000 0 0 0 01250 0.6 0 0 21600 1 0 -0.1 5.52000 1.2 -0.1 -0.2 82500 1.3 -0.2 -0.3 103150 1.2 -0.4 -0.5 114000 1 -0.7 -0.8 115000 0.5 -1.2 -1.3 116300 -0.1 -1.9 -2 108000 -1.1 -2.9 -3 8.5
10000 -2.5 -4.3 -4.4 612500 -4.3 -6.1 -6.2 316000 -6.6 -8.4 -8.5 -420000 -9.3 -11.1 -11.2 -7.5
LOUDNESS
Sound is defined as any pressure variation heard by the human ear. This translates
into a range of frequencies from 20 Hz to 20,000 Hz for a healthy human ear. In terms
of sound pressure, the human ear’s range starts at the threshold of hearing (0 dB) and
ends at the threshold of pain (around 140 dB).
13
The human ear is less sensitive to sound pressure variations in the low frequencies
compared to the higher frequencies. A 50 Hz tone must be 15 dB higher than a 1000
Hz tone at a level of 70 dB to be perceived as the same loudness by the listener. As a
rule of thumb, a doubling in the loudness of the sound occurs with every increase of
10 dB in sound pressure. Similarly, for each 10 dB decrease in sound pressure, the
loudness is cut in half.
The 10 dB loudness rule is not the same as a common guideline used when decibels
are added (or subtracted) together. In the latter guideline, a doubling in sound pressure
results in a 3 dB increase in the noise level (not a 10 dB increase as with loudness).
The 3dB rule applies only when identical noise sources are added (or subtracted). For
example, adding together two identical noise sources of 85 dB results in a total noise
level of 88 dB (85 dB + 85 dB = 88 dB).
Figure 11: Doubling Sound Pressure Adds 3dB
14
The human ear’s ability to hear logarithmic changes in sound pressure explains why
loudness increases 10 dB but the noise level from identical sources increases by only
3dB. In practice, loudness plays a small role in noise control because it is subjective
and varies from person to person. What is interpreted as loud noise by one individual
may not be loud or noise to another. Of note is that human beings do not hear sounds
in the very low frequencies. However, you may recall “feeling” rather than “hearing”
sound. Vibrations from very low frequency sounds can rattle dishes and shake home
foundations even though they can’t be heard.
Figure 12: Equal Loudness Contours
Equal loudness curves show the relative lack of sensitivity of the human ear to lowfrequencies.
SOUND PRESSURE LEVEL (SPL) AND SOUND POWER LEVEL (PWL)
Sound pressure is the change in pressure of the air above and below the average
atmospheric pressure. When dealing with sound, the changes an acoustical engineer
records can be huge—from as small as a millionth of a Pascal (also recorded in
15
pounds per square inch, abbreviated as psi) to larger pressure changes like
explosions inside reciprocating engines and gas turbines.
To measure such wide pressure changes (or amplitude), sound pressure is converted
into decibels, and referred to as the Sound Pressure Level (SPL or Lp ). The scale
starts at zero decibels and the international standard of pressure change of 2 x 10 –5 Pa.
The equation used to calculate the Sound Pressure Level is:
SPL or Lp = 10 log10 (p2 / p20) [dB]
Or, in a simpler form as:SPL or Lp = 20 log10 p + 94 [dB]
Where:SPL or Lp = Sound Pressure Levelp = root-mean-square (rms) sound pressure (Pascals or Pa)p0 = international reference pressure of 2.0 x 10 –5 Pa
Most manufacturers will make available the Sound Pressure Levels of their machines.
These machines, such as gas turbines, emit energy in the form of power, heat and
sound. The power is expressed in horsepower, the unit used to describe your car’s
performance. The acoustic energy radiating from a machine is termed sound power.
Sound power is defined as the average rate at which sound energy is radiated from a
sound source. It is measured in watts (W). The Sound Power Level, abbreviated as
PWL or Lw, is sound energy after it is converted into decibels. Like sound pressure, a
reference sound power has been established. This reference is 10 –12 x watt (W).
The equation used to calculate the Sound Power Level is:
PWL or Lw = 10 log10 (W / W0) [dB]
Or, in a simpler form as:
PWL or Lw = 10 log10 (W) + 120 [dB]Where:
PWL or Lw = Sound Power LevelW = acoustic energy of the source given in watts (W)W0 = international reference sound power of 10 –12 Watt (W)
16
The PWL or Lw is constant for a given source and is independent of the acoustic
environment. It cannot be measured directly, but must be calculated from the Sound
Pressure Level. This is because PWL can be thought of as similar to the watt rating of
a light bulb. SPL, on the other hand, is like the amount of light produced at a given
distance from the bulb in a specific environment. Sound pressure is relatively easy to
measure—the pressure variations felt by the human eardrum are the same pressure
variations detected by a microphone used to record the sound.
Table 1: Relationship between Sound Power (PWL or L w) and SoundPressure (SPL or Lp)
Pressure and Pressure Level:
Source Pascal (Pa) Decibels (dB)
Average hearing threshold 2 x 10 –5 0
Whisper 2 x 10 –3 40
Conversation 4 x 10 -2 65
Train Station 2 x 10 0 100
Jet aircraft at takeoff 6 x 10 1 130
Power and Power Level:
Source Watts (W) Decibels (dB)
Conversational voice 10 –5 70
Piano 10 –2 100
Orchestra 10 0 120
Jet aircraft at takeoff 10 2 140
Space shuttle 10 6 180
Example:1.0 watt of acoustic energy is the equivalent of 120 dB:
PWL or Lw = 10 log (1 watt / 10 –12 watts)= 10 log (1012 )= 10 (12)= 120 dB
17
Note: Unless otherwiseindicated, all acousticcalculations involvingdistance use metric units.
BBAASSIICC CCAALLCCUULLAATTIIOONNSS
CALCULATING SOUND POWER FROM SOUND PRESSURE
The Sound Power Level (PWL or Lw) of noisy
equipment is what we use to determine the amount of
attenuation needed to meet the noise level
requirement. As mentioned, the PWL cannot be
measured; it must be calculated. To calculate the PWL, we first measure the Sound
Pressure Level—usually at one meter from the machine. Also needed to calculate the
PWL is the size (or dimension) of the noise source. Manufacturers will often make
available the SPL and equipment dimensions upon request.
An equation that gives an approximate calculation of the PWL from the SPL of a noise
source is:2
PWL or Lw ≅ SPL + 10 log (A ) [dB]
Where:
SPL = Sound Pressure Level of the sound source at a specified distance
Area = height x width x length in square meters (m2)
As mentioned, the Sound Pressure Level is relatively easy to measure; a microphone
picks up the same pressure changes as the human ear. However, the sound pressure
2 The precise equation is:
PWL = SPL + 10 log [P02 * A/W0 ρ C]
Where:SPL = Sound Pressure Level of the sound source at a specified distanceP0
2 = reference pressure of 20 x 10 –5
A = area of sound source in square meters (m2 )C = speed of sound which is 340.3 meters per secondρ = density of medium; 1.225 kilograms per cubic meter in air
Since Po2 = (20 x 10 –5 )2 Pa 2
W0ρC = 1 x 10 –12 x 1.225 kg/m3 x 340.3 m2
And Po2 ÷ W0 ρC= 0.96 and 10 log (0.96) = -0.18;
Hence the formula, PWL or Lw ≅ SPL + 10 log (A ) represents an approximation of the Sound Power Level.
18
measurement doesn’t represent the acoustical energy (sound power) of a machine. To
use an analogy from another kind of energy — electrical energy — heating the head of
a pin and a stovetop element to exactly the same temperature takes different levels of
energy. The amount of electricity used to heat the pin is much less than the energy
emitted by the element. This same analogy can be applied to sound. A radio and
orchestra might produce the same Sound Pressure Level (e.g., 85 dB) at a certain
distance, but the orchestra emits substantially higher amounts of acoustical energy with
a correspondingly greater impact on the environment.
Figure 13: Comparison of Sound Power (PWL or Lw) and Sound Pressure (SPLor Lp)
The PWL also needs to be calculated in each octave band. Recall the noise peaks that
occur at discrete frequencies for most industrial equipment. The peak noise level is
often the level that is attenuated, particularly when it is causing discomfort to residents
in the neighborhood.
19
Table 2: Examples of Sound Power Levels for Select Equipment by OctaveBand Frequency *
Sound Power Level (PWL or Lw) in dB (relative to 10 –12 Watts)Octave Band Frequency (Hz)
Equipment Item 31.5 63 125 250 500 1000 2000 4000 8000LM6000 Enclosure 124.5 120.5 117.5 113.5 106.5 100.5 84.5 87.5 77.5HRSG Body 122.0 114.0 106.0 103.0 99.0 97.0 98.0 96.0 89.0Inlet Filter 116.0 120.0 112.0 108.0 107.0 113.0 107.0 102.0 92.0* PWLs for select equipment at 110 MW power station in Iroquois Falls, Ontario.
CALCULATING THE TOTAL PWL FOR A SINGLE NOISE SOURCE
After a machine’s PWL is calculated for each octave band frequency, the next step is to
enter the calculated PWLs into the following formula to obtain the Total PWL:
n
Total Sound Power Level (PWL) = L w, Total = 10 * log10 [ Σ 10 Lw, i /10 ]
i = 1
Where:Lw, I = Sound Power Level or PWL for each octave band frequency∑ = sum of number of PWLs
The total PWL should always be higher than the highest PWL recorded by octave
band—a quick way to check whether your calculation is on track.
A-WEIGHTING THE PWL OF A SINGLE NOISE SOURCE
Sometimes it is necessary to A-weight the Sound Power Level if a community’s noise
by-law is stated in dBA. To obtain the total A-weighted PWL for single noise source, a
Example:Calculating the total PWL for a LM6000 enclosure at Iroquois Falls, Ontario:
PWL or LwTotal = 10 * log10 (10 124.5/10 + 10120.5/10 + 10117.5/10 + 10113.5/10 + 10106.5/10 + 10100.5/10 + 1084.5/10 + 1087.5/10 + 1077.5/1)
PWL or Lw Total = 10 * log 10 (4.783 x 1012)PWL or Lw Total = 126.8 dB
20
correction factor, given in Figure 10, is added to the unweighted PWL (known as the
linear PWL) at each octave band frequency. Then, the A-weighted PWLs for each
octave band are inserted into the formula for calculating the Total Sound Power Level to
obtain the PWL expressed in dBA.
CALCULATING THE TOTAL PWL OF NUMEROUS NOISE SOURCES
In most industrial facilities, sound is emitted from many sources. Table 3 gives a
sampling of some of the major noise sources associated with a single gas turbine at a
peaking power plant, which are often driven by two or more gas turbines.
Example: Calculating A-weighted PWL’s using the table method. Taking the linear PWL at each frequency for acombustion exhaust, apply the correction factor from Table 3 to obtain the A-weighted result.
31.5Hz 63Hz 125Hz 250Hz 500Hz 1000Hz 2000Hz 4000Hz 8000HzTake Unweighted PWLs LM 6000
Enclosure124.5 120.5 117.5 113.5 106.5 100.5 84.5 87.5 77.5
Add A-Weighted Correction Factor -39.4 -26.2 -16.1 -8.6 -3.2 0 1.2 1.0 -1.1
Obtain A-Weighted PWL Result 85.1 94.3 101.4 104.9 103.3 100.5 85.7 88.5 76.4
21
Table 3: Sampling of Noise from Sources at a Peaking Power PlantSound Power Levels at Center Octave Bands – dB (relative to 10 –12 Watts)
Source Description 31.5Hz 63Hz 125Hz 250Hz 500Hz 1000Hz 2000Hz 4000Hz 8000HzTotal
dBInlet Gas Turbine 100.8 99.9 93.0 95.2 93.5 87.9 86.7 87.0 90.0 105.0
Turbine Vent Fan 108.2 108.2 110.1 104.1 101.0 92.1 96.8 95.3 91.1 114.5
Load Compartment Vent Fan 103.1 103.1 100.3 96.5 90.2 85.6 85.10 79.5 78.0 107.6
Load Compartment Discharge 89.0 102.0 93.0 96.0 95.0 102.0 103.0 110.0 98.0 112.3
Lube Oil Demister Vent 92.0 96.0 96.0 98.0 99.0 91.0 83.3 72.0 87.0 104.1
Accessory Module 103.0 106.0 99.3 97.1 95.9 95.4 97.7 91.7 88.3 109.5
Inlet Plenum 86.4 89.0 86.1 88.0 86.9 87.7 96.9 87.4 76.8 99.5
Turbine Compartment 108.1 109.9 104.5 102.8 100.4 98.4 103.5 98.7 93.8 114.1
Exhaust Diffuser 114.5 112.0 110.0 103.3 102.4 99.8 98.1 96.9 93.8 117.8
Load Compartment 103.1 104.9 104.8 100.3 94.9 92.7 96.6 92.7 85.8 110.2
Generator 101.9 101.8 101.4 98.0 100.3 98.8 98.0 93.0 84.0 108.9
Expansion Joint 100.8 108.8 105.8 98.8 87.8 84.8 77.8 60.8 43.8 111.3
Transition Duct 101.4 109.4 108.4 103.4 91.4 93.4 81.4 51.4 36.4 112.9
Exhaust Stack Casing 92.3 85.3 63.3 53.3 38.3 46.3 45.3 41.3 30.3 93.1
Exhaust Stack Opening 131.0 142.0 146.0 145.0 137.0 139.0 132.0 115.0 98.0 150.1
Fin Fan Cooler 57.0 96.0 88.0 93.0 92.0 90.0 89.0 88.0 69.0 100.3
Total dB 131.2 142.0 146.0 145.0 137.0 139.0 132.0 116.4 103.4 150.1
The same formula for adding (or subtracting) PWLs for a single noise source is used for
adding (or subtracting) multiple-source PWLs. The difference is that all source PWLs
are typically added (subtracted) up over a single octave band (down a column), then a
grand total is calculated for all noise sources over the nine octave bands. However,
you can add (subtract) over the individual noise sources first (across a row) and arrive
at the same grand total.
Example:Calculating the total PWL for all the noise sources in Table 3 at the 31.5 Hz octave band is:
PWL or Lw, Total = 10 * log10 (10 100.8/10 + 10 108.2/10 + 10 103.1/10 + 1089.0/10 + 1092.0/10
+ 1086.4/10 + 10108.1/10 + 10114.5/10 + 10103.1/10 + 10101.9/10 +10100.8/10 + 10101.4/10 + 1092.3/10 + 10131.0/10 + 1057.0/10)
PWL or Lw, Total = 10 * log 10 (1.28 x 1013 )PWL or Lw, Total = 131.2 dB
22
A popular method for adding (or subtracting) PWLs is the table method. For example,
first find the difference between the two loudest sources in PWLs. Next, go to Table 5
and add the specified number of dB that correspond to the difference. The sum should
then be combined with the highest remaining level and so on, until all levels are
combined.
Table 4: Table Method for Adding or Subtracting Decibels
Difference between levels – dBNumber of dB to be added to the
higher level0 3.01 2.52 2.13 1.84 1.55 1.26 1.07 0.88 0.69 0.5
10 0.412 0.314 0.216 0.1
Example:Using the table method to determine the PWL of three of the power plant noise sources in the 31.5octave band in the example in Table 4: turbine vent noise level of 108.2 dB, a generator noise level of101.9 dB and lube oil demister vent noise level of 92.0 dB. Start by subtracting the noise level of theturbine vent noise level from the generator (108.2 dB – 101.9 dB = 6.3 dB). Looking at Table 5, a 6.3dB difference means 1.0 should be added to the highest noise level.
108.2 dB – 103.4 dB = 6.1 dB; 6.1 dB converts to 1.0 dB108. 2 dB + 1.0 dB = 109.2 dB for turbine vent and generator noise
Add the lube oil demister vent noise to the subtotal. The difference between 109.2 dB and 92.0 dB is17.2 dB. Looking at 17.2 dB in Table 5, 0.1 is added to the subtotal.
109.2 dB + 0.1dB = 109.3 dB for total noise.
23
SSOOUURRCCEE--PPAATTHH--RREECCEEIIVVEERR
All noise propagation can be broken into three parts:
♦ The source
♦ The path
♦ The receiver
The source radiates sound based on its sound power (PWL). The path is how the
sound travels through the air. The receiver is what the sound impinges upon (person,
microphone, etc.).
SOURCE SPECIFICS
In industry, the most common noise sources are described as a point source, like a
gas turbine, or a line source, like a pipeline. In the free field, sound propagates
outward from point sources in uniform, concentric circles and from line sources as a
cylindrical wave, much like a weather front. Free field conditions exist when no
obstacles block the sound path. Noise from a source can either be air borne or
structure borne. Noise that travels through the air and through building walls and
openings is called air borne noise. Structure borne noise is a term used to describe
mechanical vibrations carried from machinery through to a building’s structure.
Figure 14: Structure Borne Noise
24
Whether a point or line source, occupational health standards in most countries limit
employees’ exposure to the noise. For example, the Occupational Safety and Health
Administration (OSHA) sets 85 dBA over an eight hour period as the maximum
admissible noise exposure limit in the workplace. The OSHA standard is representative
of a source noise limit. With this standard in mind, plant equipment is typically
ordered to emit sounds of no more than 85 dBA at one meter (3 feet).
Normally 10 to 12 measurements of the sound pressure around the periphery of a
machine at one meter (3 feet) are taken to obtain the source noise level. However, the
number of measurements vary by machine shape and size. National and international
standard institutes, such as ASTM (American Society for Testing and Materials), ANSI
(American National Standards Institute), CSA (Canadian Standards Association) and
ISO (International Standards Organization) publish guidelines on how to construct a grid
over equipment and gather point measurements at different frequencies.
Microphones are located at the points and, a sound level meter set to A-weighting,
measures sound levels at mid-band frequencies of 63, 125, 250, 500, 1000, 4000, and
8000 Hz. The measurements are averaged for each frequency and corrected for the
machine’s measuring surface area to find the Sound Power Level. The floor is
assumed to reflect the sound energy and so it is not included in the measuring surface.
Often in industry, background or ambient noise exists along with the source noise.
Industrial parks, for example, can emit high ambient noise levels from the many
industries on site. To get an accurate reading of noise from a specific source, the noise
level of the source must be at least 10 dB higher than the ambient noise level.
The following steps are recommended to obtain measurements of noise for a source
under conditions of background noise:
1. Measure the total noise level with all equipment running.
2. Shut down all equipment and measure the background noise level alone.
3. Determine the difference between the two measurements.
25
If the total noise level is 10 dB greater than the ambient noise level, then
background noise won’t interfere with a true measurement of the total noise level. If the
background noise level is 3 dB or less, then an accurate measure is not possible. If the
background noise is between 3 dB and 10 dB, a correction is necessary. To make
corrections the following table method can be used.
Table 5: Correction for Background Noise
dB difference between sound pressurelevel and background sound pressure level
alonedB to subtract from sound pressure level
Less than < 6 1.06 1.07 1.08 0.59 0.5
10 0.5Greater than > 10 0.0
Source: ANSI, S12.34 - 1988
PATH SPECIFICS
Under free field conditions, point sources produce noise that spreads uniformly as a
sphere, much like water ripples on a pond. By contrast, sound flows from line sources
as a cylindrical wave. The sound field within close proximity to a noise source is called
the near field. A person is considered to be standing in the near field if he or she is
within one size of the noisy object in distance away. Size is measured according to
the largest dimension of the object. So, if the object is a building and the largest
dimension is the building’s height, then the near field would start at the point away from
the building that is equivalent to its height.
26
Figure 15: Near Field and Far Field
Standing 3 meters (10 feet) away from this 15 meter (50 feet) high power plant would put aperson in the near field. Standing at a distance more than 15 meters away places her in thefar field.
In the free field, the SPL increases the closer you move toward the noise source and
decreases the further you move away. More precisely, the SPL increases or
decreases as the inverse square of distance. The formula used to calculate the SPL
at a known distance away from a noise source in the free field is:
Lp(R2) = Lp(R1) – 20 log 10 ( R2 ) [dB] R1
Where:Lp (R1) = Sound Pressure Level at the initial locationLp (R2) = Sound Pressure Level at the new locationR1 = distance from the noise source to the initial locationR2 = distance from the noise source to the new location
27
A popular method is to decrease the SPL by 6 dB for every doubling of distance away
from the source. If you are located one meter away from a point source, then move one
meter further away, the SPL drops by 6 dB. If you move to 4 meters away, it drops by
12 dB, at 8 meters by 18 dB, and so on. This method is derived from the inverse
square law of sound intensity.
Sound intensity is defined as the sound power per unit area. To understand the
concept of sound intensity, think of sound radiating outward from a point source. Under
free field conditions, this sound is of uniform intensity (power per unit) in all directions.
The sound power passing through a small area (d) near the sound source is the same
sound power passing through areas further away (2d, 3d, and 4d), but each successive
area gets larger while the sound intensity decreases with distance.
Example:The sound level specification you are given is 75 dB for the compressor package at 50 meters away.You have a residence 800 meters away from the facility. The SPL at the residence would be 51 dB,calculated as follows:
SPL or Lp (R2) = Lp (R1) – 20 log 10 ( R2, decibels) R1
SPL or Lp (800 meters) = Lp (50 meters) – 20 log (800/50)SPL or Lp (800 meters) = 51 dB
Example:Using the 6 dB rule, you also get 51 dB at 800 meters, the equivalent of using the formula:
Distance (m) Sound Level (dB)50100200400800
1600
756963575145
28
Figure 16: Sound Intensity
The same sound energy is distributed over successively larger areas as distance from thesound source is increased.
The uniform, concentric circles are actually spheres. As the area of a sphere is 4πr2 ,
the area of a small segment on the surface of the sphere varies in relation to the square
of the radius. “Doubling the distance from d to 2d reduces the intensity to ¼, tripling the
distance reduces the intensity to 1/9, and quadrupling the distance reduces the intensity
to 1/16. Intensity of sound is inversely proportional to the square of the distance
in a free field.” 3
The inverse square law for intensity becomes the inverse distance law for sound
pressure. That is, sound pressure varies inversely as the first power of distance.
When sound pressure is plotted against distance units, this means that sound pressure
is reduced 6 dB for each doubling of the distance. This is called the 6 dB rule.
3 F. Alton Everest. The Master Handbook of Acoustics. 3rd Ed. New York.: Tab Books, 1994, page 68.
29
Figure 17: Sound Pressure Decreases 6 dB for Each Doubling of Distance
The inverse square law holds true only for discrete distance points and under free field
conditions. If sound values between distance points (e.g., 425.5 meters) are required,
the calculation rather than the table method is used.
For a line source, the sound spread equates to a 3 dB loss per doubling of distance.
The formula for calculating noise levels at different distances from a line source is:
Lp (R2) = L p(R1) – 10 log 10 ( R2 ) [dB] R1
Where:Lp (R1) = Sound Pressure Level at the initial locationLp (R2) = Sound Pressure Level at the new locationR1 = distance from the noise source to the initial locationR2 = distance from the noise source to the new location
30
Figure 18: Sound Propagation from a Line Source
In the near field, noise from a point source diverges from the –6 dB guideline. Because
point sources are typically housed in buildings, the building behaves as a plane source,
rather than a point source. Sound is radiating outward from a flat surface. With plane
sources like buildings, there is minimal noise reduction until the radial distance (r = b/π,
where b is the width of the building) is reached. The radial distance is roughly one-third
a building’s width. At this point and as far as the far field, the Sound Pressure Level
Example:The sound level specification you are given is 55 dB for a paper recycling bailer at 200 meters away. You havea residence 800 meters away from the facility. The SPL at the residence would be 51 dB, calculated as follows:
L (R2) = L(R1) – 10 log 10 ( R2, decibels) R1
L (800 meters) = L (200 meters) – 10 log10 (800/200)L (800 meters) = 55 dB – 10 log10 (800/200)L (800 meters) = 51 dB
31
diverges at the same rate as a line source (-3 dB per doubling of distance), then
changes to –6 dB in the far field.
Figure 19: 3dB Near Field and 6 dB Far Field Guideline for a Point Source
The near field-far field guideline applies only in the free field. In practice, sound waves
regularly collide with obstacles. Think of the static on your car radio as you drive into a
tunnel. When a sound wave encounters an obstacle, five phenomena can occur:
absorption, reflection, transmission, diffraction and refraction.
Some of these conditions can occur at the same time. Part of a sound wave’s energy is
absorbed and part is reflected when it strikes a surface. This fact is important when
considering how to attenuate noise. For example, the more porous a surface, the more
sound is absorbed rather than reflected.
32
When an object is a certain thicknesslike a wallpart of the sound wave’s energy is
transmitted through it. In general, more sound energy will pass through a thin wall
than a thick one. If sound-absorbing material is also added inside of the wall, then the
amount of noise that gets through to the other side will be less than if the wall were left
“untreated”. The amount of noise lost when sound waves pass through a wall or barrier
is called Transmission Loss (TL). This is the difference between the noise level
measured on the source side of a noise barrier, and the level measured on the receiver
side.
Figure 20: What Happens When Sound Waves Encounter an Obstacle
Diffraction is a change in the direction of travel of sound when the sound encounters
an obstacle. Objects capable of diffracting (bending) sound must be large compared
to the wavelength of the sound. For low frequency noise, with its long wavelength, a
barrier must be acoustically large (larger than the wavelength of the sound) to change
the sound path.
33
Refraction changes the direction of travel of the sound by differences in the speed of
propagation. Wind and temperature changes are most common causes of refraction.
Sound travels faster in warmer air than in cooler air causing the tops of the wavefronts
to go faster than the bottom parts. Under normal conditions, air temperatures decrease
as altitude increases. This causes sound waves to refract upwards which decreases
audibility along the ground. Sometimes, the temperature is higher above the ground
than near the grounda condition called a temperature inversioncausing sound
waves to bend back toward the ground and increase audibility. Temperature inversions
are especially common at dawn, dusk, and in cold winter conditions.
Also, because winds aloft are usually faster than at ground level, the upper part of a
sound wave travels faster than the lower part when travelling with the wind. The
sound wave travels slower when traveling against the wind. Refraction of the noise
toward the ground occurs in the first instance and refraction away from the ground in the
latter case.
Figure 21: Refraction of Sound
34
RECEIVER SPECIFICS
Most municipalities set a dB or, more frequently, a dBA limit at the nearest sensitive
receiver (NSR), usually defined as the property line of an industrial, commercial or
residential building or its outside wall. A property line noise limit is typically used to
control noise from stationary sources like power plants and compressor stations. A
time limit, during which noise is either prohibited or required to stay below a certain
dBA level, is frequently combined with the property line limit. When using time limits, an
allowable day-time noise level is specified which is higher than a night-time noise
level.
Some localities define permissible noise levels for areas. In the case of area limits,
noise is restricted to a dBA level at the boundary of the nearest sensitive area (NSA).
Industrial zones allow higher noise levels than residential areas that have higher noise
levels than noise sensitive ones like hospitals or nursing homes. What becomes
interesting from a noise control perspective is when industrial areas abut noise sensitive
zones.
The dBA limit in noise guidelines is sometimes qualified with the symbol Leq. Leq is
defined as the equivalent continuous sound pressure level, and represents an
average of the noise history at a given site or location. The Leq is used when it is
important to consider variations in Sound Pressure Levels over time. It is usually
appraised hourly and then averaged over 24 hours, using the following formula:
n
Leq = 10 log (1/T Σ ti 10 Li/10)
i = 1
Where:T = total time (usually 24 hours)ti = usually an hourly time interval (with Σ ti = T)Li = Sound Pressure Level at time ti, measured in dBA (and converted to dB, if
required)
35
People are more sensitive to noise at night than they are during the day. Background
levels drop during the night-time when people are at home asleep. The day-night level,
Ldn, is an energy average of the 24 hour Leq for a day, with a 10 dBA penalty added to
the sound level for the hours between 10 p.m. and 7 a.m. The CNEL (Community
Noise Exposure Level) is the same as the Ldn but with a 5 dBA penalty added to the 10
dBA penalty from 10 p.m. to 7 a.m.
Figure 22: Equivalent Continuous Sound Pressure Level (Leq)
Other communities base their noise requirements on the existing background sound
level, L90 or L95 (the noise level present 90% or 95% of the time) with noise levels
allowed to reach a certain level over the ambient level (e.g., 5 dBA). Other communities
specify that the sound level must not exceed a certain limit 75% of the time (L75), 50 %
of the time (L50), or 10% of the time (L10). Still other communities specify noise limits for
each octave band.
36
Figure 23: Common Noise Level Criteria Used by Regulators
Table 6: Examples of Community Noise Guidelines
Municipality Sound Level Location
Miami, Florida Ambient + 10 dBA or 75 dBA Industrial property line
Toronto, Ontario 83 dBA L90 15 meters from equipment
World Health Organization
(WHO)
55 dBA Leq
Daytime
At residence
Puerto Rico 75 dBA L10 Industrial property line
Denver, Colorado 80 dBA Industrial property line
Salinas, California 60 CNEL
80 Ldn
Industrial property line
New York City, New York 70 dBA 25 feet from equipment
37
To measure the effect of noise from an industrial site on the NSR, an ambient noise
survey is conducted. Of interest is the total Sound Pressure Level generated at the
NSR by the many sound sources on the industrial site. The level of ground absorption,
site topography, placement of buildings, and atmospheric conditions influence the
sound pressure levels at the NSR. Sound pressure measurements at the receiving
property are typically taken every hour over a 24 hour period under calm and dry
weather conditions. Microphones are placed at a height of 1.5 meters (5 feet) above
the ground or surface and away from any natural or artificial structure.
For most noise, an octave band analysis suffices. When audible discrete frequency
tones exist, a narrower band analysis is usually performed (either one-twelfth or one-
third octave band). If noise is fluctuating, the maximum and minimum values during the
time the noise is “on” are recorded. For intermittent noise, the average noise level is
recorded during the “on” time. The maximum or peak noise level in addition to the
average noise level is captured when impulsive noise is the problem.
Ambient measurements are especially important when siting a plant or station. How the
facility is situated has a strong bearing on how much noise it will contribute at the NSA
or NSR. By configuring the plant design so that noise is channeled away rather than
toward the NSR or NSA, significant cost savings for attenuation can be realized.
Measuring the ambient noise level at a fully operational plant is sometimes necessary.
The need arises when documentation is required to determine the source and level of
noise affecting an NSR. Taking noise measurements at built-up sites may be
complicated. Sound pressure patterns are often disturbed by buildings and other
structures as well as landscaping. Since it is important to take measurements under
free field conditions, sound pressure may have to be measured in locations away from
structures, then extrapolated out to the NSR or back to the noise source.
Directional noise from existing facilities is also common. Sound from building
openings, such as exhaust stacks and ventilation and combustion outlets, emit more
38
noise in the front of the openings than to the sides. Frequency and the area of the
opening influence the directivity effect. The higher the frequency and larger the
opening, the greater the sound’s impact. Sound pressure measurements at more than
20 locations may be needed to determine the directivity effect.
AAccoouussttiicc MMaatteerriiaallss
Acoustical materials are divided into the following basic types:
1. Sound absorbing materials
2. Transmission loss or barrier materials
3. Resonator-type materials
4. Damping materials
5. Vibration isolators
SOUND ABSORBING MATERIALS
Sound absorbing materials are porous materials such as rock wool, mineral wool,
glass fiber, and foam. The effectiveness of acoustical material to absorb sound depends
on its thickness, amount of airspace, and density. For every inch of thickness of a
porous material (e.g., rock wool) sound loss is about 1 dB at 100 Hz to 4 dB at 3000 Hz.
The amount of sound absorbed at the surface of a material is described by an
absorption coefficient ( α ). The absorption coefficient relates to sound reflection,
where a high α equals low reflected energy and a low α equals high reflected energy.
Marble slate has an absorption coefficient of 0.01 (almost no absorption and high
reflection). Some specially constructed sound rooms score as high as 1.0 (total
absorption and no reflected energy).
The absorption coefficient of a material typically increases with frequency. At low
frequencies, porous materials absorb less sound, so that materials must be thicker to be
effective. The overall performance of a sound-absorbing material is often described by
39
the Noise Reduction Coefficient (NRC). The NRC is the arithmetic average of the
absorption coefficient at 250, 500, 1000, and 2000 Hz.
Sound absorption differs from sound insulation. Sound absorption relates to sound
reflection, whereas sound insulation relates to the amount of acoustic energy able to
pass through material. The sound absorption provided by a 10 centimeter-thick (4-inch
thick) fiberglass acoustical blanket is high, but its insulation quality is low. Sound is able
to travel through the material to the other side. By contrast, a lead wall absorbs almost
no sound but it is a very good insulator.
TRANSMISSION LOSS OR BARRIER MATERIALS
Lead is an example of a transmission loss or barrier material. Barrier materials are
dense and rigid and are defined in terms of their Transmission Loss (TL).
Transmission Loss is defined as the logarithmic ratio of the sound power on one side of
a barrier (wall or partition) to the sound power transmitted to the other side. The higher
the TL, the better a barrier material is at limiting or attenuating the amount of sound
travelling through it. For example, a wall or barrier having a TL of 45 dB reduces a 120
dB interior noise level to 75 dB. A wall with a TL of 60 dB reduces the same amount of
noise to 60 dB.
Figure 24: Transmission Loss (TL) for Two Walls
40
TL is calculated using the following equation:
TL (dB) = 10 log 1/τ = 10 log Wi/Wt
Where:τ = sound transmission coefficient; ratio of the PWL incident on one side to PWL on
the other sideWi = incident sound power (PWL on source side)Wt = transmitted sound power (PWL on the receiver side)
As a general rule, the heavier and thicker the wall the greater the attenuation of the
sound or higher the TL. This is because it is difficult for sound waves in air to move
or excite a dense, heavy wall. Sound transmission through walls, floors or ceilings
varies with sound frequency, and the weight and stiffness of the construction. This
gives rise to the effect known as the mass law in acoustics which states that for each
doubling of the surface weight of the wall, there will be about 5 or 6 dB less transmitted
sound. The mass law also states that for each doubling of the frequency (Hz) there will
be about 5 or 6 dB less transmitted sound. Doubling of the frequency has about the
same effect as doubling the surface weight.
RESONATOR-TYPE MATERIALS
Perforated metal wall liners or tiles are examples of resonator materials. The holes in
the liner or tile act as resonate types of sound absorbers. A common resonator is the
opening of a pop bottle or jug; blowing across the opening produces a tone at its natural
frequency of resonance. When the diameter of the hole or length of cavity behind it is
changed—as when a larger pop bottle is used or you fill the bottle with water—the
frequency of resonance also changes.
When a metal perforated liner is applied, sound impinging on the holes is absorbed into
the cavities, but a portion is reradiated back toward the sound source in the form of a
hemisphere. Because the sound energy is bounced back toward the source in semi-
41
circular waves, sound is actually diffused and noise levels are reduced. The holes of
liners can be sized and aligned in such a way that sound is absorbed and diffused at
specific frequencies.
DAMPING MATERIALS
Damping materials are used to reduce structure borne noise. Structure-borne noise
is a term used to describe mechanical vibrations carried from machinery through to a
building’s structure. For example, an engine bolted onto a metal skid that’s bolted to the
floor transmits huge amounts of acoustical energy through to the structure. Vibrations
from rattling machinery travel easily through solid structures like wood, steel, concrete
or masonry. With wood, concrete and bricks, vibrations are attenuated 2 dB in 30
meters (100 feet), while steel requires 20 times the distance for the same attenuation.
Damping materials create mechanical resistance to the structure-borne sound by
converting sound energy into heat through friction. An example of a damping material is
the spray-on coating compound placed under automobiles.
VIBRATION ISOLATORS
Vibration isolation is also used to reduce the transmission of noise through a
structure. Vibration isolators lower the vibration at its source. They are elastic
elements, such as coiled springs, and rubber, felt, cork or glass fiber materials, which
are as different as possible from the structure or mechanism. Vibration isolators can be
made from elastomers (compressed or shear, ribbed Neoprene); other compressed
material (cork); fibrous mats (felt and glass fiber); and metal springs. Vibration isolators
are often used in conjunction with damping materials. For example, steel springs are
undamped and placing them on elastomer pads, improves their level of vibration
isolation.
42
AATTTTEENNUUAATTIIOONN
Once the noise sources are identified and measured, the next step is to attenuate the
noise. Attenuation is defined as the difference in dB or dBA between two points in and
along the path of sound propagation. The aim of attenuation is to reduce or divert the
amount of sound energy reaching the receiver. The key to attenuation is to apply noise
control materials and measures that are both effective and economical. Noise controls
range from the simple to complex.
BUFFERS
One of the simplest attenuation methods is to place enough distance between the noise
source and the NSR so that noise is not a concern. Establishing a buffer zone is
possible when land is readily available. However, it usually takes a large amount of
land to stop noise from affecting the surrounding environment. Recalling the 6 dB rule,
it could take as much as 1,800 meters (approximately 5,900 feet) to reach 75 dB at the
NSR when the source noise is a high as 140 dB.
NATURAL BARRIERS
Shrubs, trees and berms are often used as natural noise blockers. For trees to be
effective barriers, they must be in a continuous stand, 50 feet tall, 100 feet deep, have
dense foliage down to the ground, and be evergreen. When only a line of deciduous
trees is planted, noise easily travels through the stand, particularly during the winter
when trees lose their foliage. Berms are more effective in stopping high frequency
noise. Low frequency noise, with its long wavelength, can easily slip over berms.
BARRIERS
Barriers are free-standing walls or structures intended to block source noise. The
barrier functions by absorbing a large amount of the sound energy and/or deflecting it
away from the source. Barriers reduce sound levels, but work best at reducing high
frequency noise. Barriers are most effective when they are at least three times larger
43
than the wavelength of the major noise contributor.4 For best results, barriers should
have a high transmission loss and be highly absorptive. Barriers made from a
combination of sound-absorbing and transmission loss materials give highest acoustic
performance. Concrete walls are often used as barriers. As a dense material, concrete
is a better sound insulator than sound absorber, so barriers made from concrete reflect
sound rather than absorb it.
When a barrier is wrapped around a noise source, it acts as a partial enclosure.
Partial enclosures come in a variety of configurations: two-sided, three-sided with a
roof, four-sided without a roof, and so on. Barriers and partial enclosures can be
effective and economical noise reducers, lowering noise levels by up to 12 or 15 dB.
ACOUSTICAL ENCLOSURES
If more than 12 to 15 dB of noise reduction are required, a total enclosure may be
needed to contain the noise. Typically, acoustic enclosures are modular boxes with
relatively high transmission loss and absorptive internal surfaces placed over noise
sources. The Insertion Loss (IL) is a measurement of enclosure performance, defined
as the reduction of sound pressure level at some position that occurs after the
enclosure is installed.
Insertion loss of an acoustic enclosure can be estimated as:
IL = TL + 10 log α
Where:TL = Transmission Lossα = absorption coefficient
By virtue of their design, enclosures can create heat build-up. Heat build-up is handled
by adding a ventilation blower, with silencers for intake and exhaust air. Fans and
4 Paul Jensen, Charles R. Jokel, and Laymon N. Miller, Industrial Noise Control Manual. Reprint. Cambridge,Massachusetts: Bolt Beranek and Newman, 1984: p.56.
44
internal ducting also are needed to supply cool air and remove hot air. The minimum
flow rate of cooling air, Q (in cubic meters or feet per minute), depends on W, the watts
of heat generated, and ∆T, the temperature rise permitted. At sea level, Q = 1.76 W/
∆T.
Most enclosures need openings to provide gas, water and/or steam, electricity and
lighting. Access to the machine through doors or removable panels is also required for
maintenance and servicing. The enclosure must be air tight to reduce the amount of
interior noise radiating through ventilation openings, engine intake and exhaust ducts,
cracks under doors and at panel joints, pipe penetrations and other openings. Even a
slight opening (such as which occurs along an ill-fitting panel joint) can cause a huge
reduction in attenuation (as high as 30 dB).5
ACOUSTICAL BUILDINGS
Sometimes, acoustical equipment enclosures are not enough to reduce noise to
required levels. Standard enclosures provided by manufacturers are designed to meet
an 85 dBA limit (at one meter), but higher attenuation is sometimes needed.
Customized, highly acoustical enclosures or acoustical treatment of the building in
addition to the enclosure provide alternatives.
An acoustical building is similar to an enclosure, but on a larger scale. The building
walls and roof are termed the acoustic envelope. In the design of the envelope, mass
law applies so that thick, dense walls provide better attenuation. However, few walls or
barriers behave exactly according to the mass law; they have elasticity so that
vibrations can occur. Because of this, the envelope is usually comprised of all the
materials used to attenuate sound: acoustical materials, barrier materials, damping
materials, and vibration isolators.
5 Lewis H. Bell. Industrial Noise Control: Fundamentals and Applications. New York and Basel: Marcel Dekker,Inc., 1982.
45
The acoustic performance of a wall structure of a building is often described by an STC
(Standard Transmission Class) rating. The American Society for Testing Materials
(ASTM) has introduced the Standard Transmission Classification (STC) to allow for the
comparison of various types of acoustical walls and roofs according to their
Transmission Loss properties. The STC rating is derived from the TL value of a wall
measured at different octave band frequencies. The TL values are plotted on semi-log
paper against a reference contour produced by the ASTM, producing the STC value.
The higher the STC rating, the better a wall or roof insulates against noise. For
example, a wall of STC 50 dB has greater attenuation capability than a wall of STC 40
dB. Without the STC, comparisons are difficult because actual measurements of
Transmission Loss deviate widely even in controlled acoustic laboratories, where
resonance and other elements affect a sound’s behavior.
Table 7: STC Ratings and Their Relationship to Sound Proofing Properties
STC RatingSoundproofing
Properties Speech Comparisons
25-30 Poor Normal speech understood easily anddistinctly through a wall
30-35 Fair Loud speech understood; normal speechaudible but understood with difficulty
35-40 Good Loud speech audible but not understood;normal speech inaudible
40-50 Very Good Loud speech and average radio and TV;only faintly audible
50+ Excellent Very loud noises and hi-fi faint orinaudible
The STC standard applies to frequencies from 125 to 4000 Hz. For this reason, the
standard does not sufficiently consider the importance of low frequency attenuation,
with the result that walls appearing to have adequate STC ratings often fall below what
46
is required. The ASTM also cautions that its system is not intended for use with
external wall structures or barriers.
Openings can also have a significant effect on the TL of a building wall or roof. As an
example, a heavy metal plate with holes over 13% of its surface will transmit 97% of the
sound impinging on it. The PWL or Lw of a sound that will pass through an opening is
approximately determined using the equation:
Lw = Lp + 10 log A
Where:Lp = Sound Pressure Level measured at or near the openingA = the cross-sectional area of the opening in square meters
To reduce the amount of interior noise radiating through apertures, the building must be
made airtight and silencers installed where air is ventilated.
SILENCERS
Silencers or mufflers are widely used to control noise from building openings. There is
no technical distinction between a silencer or muffler, and the terms are used
interchangeably.
Silencer performance is described using the same terms that are applied to acoustic
enclosures or buildings.
1. Insertion Loss (IL) is the difference in sound pressure at the same point before andafter a silencer has been installed. Dynamic insertion loss (DIL) is the reduction inthe sound level under actual operating conditions.
2. Transmission Loss (TL) is the ratio of the sound power impinging upon the silencer(at the source side or silencer entrance) to the sound power transmitted by thesilencer (at the receiving side or the silencer exit).
3. Noise reduction (NR) is defined as the difference between the Sound PressureLevel (SPL) measured at the source side of a muffler and the Sound Pressure Level(SPL) measured at the receiving side.
47
Silencers are of two basic types: 1) absorptive or 2) reactive.
Absorptive silencers contain acoustic materials and rely on the absorptive properties
of these materials to limit noise. They are used to treat noise where large volumes of air
or gas need to be moved at relatively low static pressure, such as on the intake
(suction) and exhaust (discharge) of centrifugal compressors, forced draft fans, gas
turbines, steam or process vents and similar equipment.
The simplest form of an absorptive silencer is a parallel baffle. Parallel baffles look like
a line of furnace filters, each covered by a perforated liner. The “filter” part is a fibrous
material (usually glass or mineral wool). The acoustical performance of baffles
increases with the thickness of the absorbing materials, the narrowness of the spacing
and longer the length. Baffles are placed parallel to the air or gas flow and are
particularly useful in applications where pressure losses need to be kept at a minimum.
Baffles are typically inserted into ducts, stacks, etc. which accommodate inlet or
discharge flows.
Figure 25: Example of Parallel Baffles
Baffles
48
A parallel baffle can be made in tubular form to allow for interfacing with circular inlets
and exhausts. The tube, called an absorptive silencer, consists of straight runs of
acoustically-lined baffles inserted behind perforated metal sheets and wrapped around
in heavy gauge steel. When a silencer is placed at an inlet opening, a thicker baffle is
able to give high attenuation, particularly in the lower frequencies. For exhaust
openings, a thick baffle can actually decrease attenuation. What is done when noise
and flow move in the same direction—as is the case with discharge systems—is to
narrow the space between baffles rather than increase their thickness.
Reactive silencers don’t contain absorptive materials but work on the principle of
reflection and dissipation of sound waves. The reactive (reflective) silencer contains
one or more chambers and perforated tubes inside a casing, but no absorption
materials. A portion of the sound energy entering the silencer is reflected from the
chamber casing back to the sound source. Another portion is dissipated through the
perforations in the tubes. For higher acoustic performance, multiple chambers and
perforated tubes of different sizes are used. The reactive silencer is used primarily for
low frequency control from blowers and compressors.
Higher performance silencers combine both absorptive and reactive principles in their
construction. Custom-made silencer designs with multiple chambers in addition to
acoustically-lined baffles are often required to meet operational requirements. Lagging
of the silencer is also sometimes needed to improve acoustic performance.
49
Figure 26: Example of an Absorptive-Reactive Silencer
Source: Jim R. Cummins and Bill Golden. Silencer Application Handbook. Stoughton, WI:Universal Silencer, 1993, p. 49.
ACOUSTIC PLENUMS
A type of chamber that operates like a reactive silencer is called a plenum. When used
for noise control, plenums are lined with porous materials. Plenums are also used to
slow down high velocity air. As a chamber, acoustic plenums can be found just about
anywhere in industry. For example, acoustic plenums are especially designed for the
inlet and exhaust ends of gas turbines. When required, an entire building can be
designed and acoustically-lined to work as an acoustic plenum.
Multiple Chambers
Absorption Material
50
Figure 27: Example of an Acoustic Plenum
ACOUSTIC LOUVERS
Louvers are designed to eliminate the line-of-sight from the source to the outside. They
can also be acoustically treated to limit noise from air flowing in and out of a building.
Louvers are overlapping slats designed to admit air into a building and exclude rain.
The slats are typically lined with porous materials. Like baffles, the spacing and length
of the slats and thickness of porous material determines acoustical performance.
51
Figure 28: Example of an Acoustic Louver
ACOUSTIC LAGGING
Lagging or wrapping of acoustical material is another method of noise control. Lagging
is often placed around pipes but acoustical wrapping can be applied to noisy equipment
or even silencers. Lagging typically consists of sound absorbing material (fibrous glass,
mineral wool, or polyurethane foam) with an outer layer of dense vinyl or sheet metal.
NNOOIISSEE CCOONNTTRROOLL AAPPPPLLIICCAATTIIOONNSS
ATCO ACOUSTIC ASSEMBLIES
ATCO has developed a line of Noise Management assemblies from sound-absorbing,
barrier, and resonator-type materials and that include vibration isolation and damping.
The assemblies are either whole-wall systems or acoustic panels. Whole walls are
52
erected in layers at the site, starting from the inside. The Noise Management panels
are factory-manufactured and assembled in situ.
Figure 29: Example of a Noise Management Assembly
Each assembly starts with a perforated metal liner. The liner can serve two purposes. It
protects the sound absorbing materials and may act as a resonate type of sound
absorber. Liners can be selected based on the dominant noise frequency. Since
industrial noise is generally broad band with a heavy low frequency component, a liner
that resonates at the lower frequencies may be used.
In colder climates, and where building codes require it, a fire-resistant vapor barrier is
installed next to the liner to control condensation. Next, a layer of acoustic material is
applied. Multiple acoustic layers are used if the wall must achieve very high acoustic
performance. To achieve such performance, a barrier material or septum layer (or
layers) is placed between the acoustic materials. The septum layer is dense and has
53
high Transmission Loss. The outermost layer of the wall structure is a protective, leak
proof facing (e.g., metal cladding, brick, etc.).
ATCO’s acoustic assemblies are applied over structural steel frames rather than affixed
to concrete block walls because the assemblies can be made highly sound absorptive.
A concrete block wall is massive but it is very reflective and even when absorptive
materials are applied to the surface, sound waves passing through the materials are
reflected off the concrete blocks—some, back into the room. In addition, because
ATCO’s assemblies have both high absorption and transmission loss, they can be
significantly lighter than concrete to achieve the same attenuation level. When using
steel framework, damping and vibration isolators are used to reduce flanking (also
called structure borne noise).
ATCO’S BALANCED APPROACH
Acoustically treating the enclosure or building envelope represents one aspect of noise
control. A balanced approach is needed to provide both effective and economical
noise reduction. In a balanced approach, all noise sources are identified, which can be
over 200 in a facility like a power plant. The Sound Power Levels of each source is
entered into an acoustic model. The model generates noise level contours from the
industrial site out to the NSR before acoustic treatment. Many contour maps use
purple and red to display high noise levels, and shades of green to represent lower
noise levels.
54
Figure 30: Noise Contour Levels at a Power Plant Before Acoustic Treatment
55
Figure 31: Noise Contour Levels at a Power Plant After Acoustic Treatment
56
A benefit to using computer modeling is that various acoustic treatments can be applied
to a site “on paper”. This allows a view of the acoustic alternatives before any
commitment is made to the type (and cost) of treatment. The various acoustic
treatment scenarios include one or all of the noise control elements: acoustic envelope,
silencers, plenums, lagging, and so on. In a balanced design, the aim is to select an
acoustical approach that meets the noise requirement at an affordable price. For
example, making the walls and roof of higher attenuation, the acoustical target for the
exhaust silencer could be relaxed — often a cheaper alternative.
Figure 32 depicts ATCO’s balanced approach. Walls with higher STC values are used
to the north and west of the power plant, closest to the affected residences. Less
acoustic (and less expensive) walls are used to the south and east, furthest away from
the community. Silencers are placed at building openings to limit noise. Plus, the DIL
performance of the silencers is balanced with the TL performance of the building’s walls
to achieve the most cost-effective acoustic treatment.
57
Figure 32: Example of ATCO’s Balanced Approach
Northwest View of the Acoustical Treatment of a 110 MW Power Plant
58
Southeast View of the Acoustical Treatment of a 110 MW Power Plant
TESTING AND GUARANTEES
ATCO’s assemblies are tested at certified acoustical laboratories. Tests involve the
determination of the NRC (Noise Reduction Coefficient) and STC (Standard
Transmission Class). Sound pressure measurements are made at all frequencies.
Measurements within the range of 100 to 5000 Hz are conducted in an acoustical
laboratory. To test acoustic performance below the 100 Hz octave band (31.5 to 100
Hz), tests must be conducted in the field. The reason why tests below 100 Hz are not
made is due to the small size of most acoustical laboratories, which do not permit
accurate recording of long low frequency wavelengths.
59
Figure 33: Sample Acoustical Test
60
Because ATCO tests the Noise Management assemblies in the laboratory as well as
in the field, the company can guarantee their acoustic performance. ATCO also
guarantees that the noise target will be met using its balanced approach to the noise
problem.
61
Useful Sources
Bell, Lewis.H. (1973). Fundamentals of Industrial Noise Control. Trumbull, CT: HarmonyPublications.
Bell, Lewis H. (1982). Industrial Noise Control: Fundamentals and Applications. NewYork and Basel: Marcel Dekker, Inc.
Everest, F. Alton. (1994). The Master Handbook of Acoustics. 3rd ed. New York.: TabBooks.
Jensen, Paul; Jokel, Charles R.; Miller, Laymon N. Industrial Noise Control Manual.Rev. ed. Cambridge, MA: Bolt Berank and Newman, Inc., 1984.