An Introduction to Sound Level Data for

Mechanical and Electrical Equipment

Course No: M03-027

Credit: 3 PDH

J. Paul Guyer, P.E., R.A., Fellow ASCE, Fellow AEI

Continuing Education and Development, Inc.

9 Greyridge Farm Court
Stony Point, NY 10980

P: (877) 322-5800
F: (877) 322-4774

[email protected]
 An Introduction
to Sound Level
Data for
Mechanical and
Electrical
Equipment


J. Paul Guyer, P.E., R.A.
Paul Guyer is a registered civil engineer,
mechanical engineer, fire protection engineer,
and architect with over 35 years experience
in the design of buildings and related
infrastructure. For an additional 9 years he
was a senior advisor to the California
Legislature on infrastructure and capital
outlay issues. He is a graduate of Stanford
University and has held numerous national,
state and local positions with the American
Society of Civil Engineers and National
Society of Professional Engineers.

J. Paul Guyer 2013 1

 CONTENTS

1. INTRODUCTION
2. SOUND PRESSURE AND SOUND POWER LEVEL DATA
3. PACKAGED CHILLERS WITH RECIPROCATING COMPRESSORS
4. PACKAGED CHILLERS WITH ROTARY-SCREW COMPRESSORS
5. PACKAGED CHILLERS WITH CENTRIFUGAL COMPRESSORS
6. BUILT-UP REFRIGERATION MACHINES
7. ABSORPTION MACHINES
8. BOILERS
9. STEAM VALVES
10. COOLING TOWERS AND EVAPORATIVE CONDENSERS
11. PUMPS
12. FANS
13. AIR COMPRESSORS
14. RECIPROCATING ENGINES
15. GAS TURBINE ENGINES
16. ELECTRIC MOTORS
17. STEAM TURBINES
18. GEARS
19. GENERATORS
20. TRANSFORMERS
21. OPENING IN A WALL
22. GLOSSARY

(This publication is adapted from the Unified Facilities Criteria of the United States government which are
in the public domain, have been authorized for unlimited distribution, and are not copyrighted.)

(Figures, tables and equations in this publication may at times be a little difficult to read, but they are the
best available. DO NOT PURCHASE THIS PUBLICATION IF THIS LIMITATION IS NOT
ACCEPTABLE TO YOU. )

J. Paul Guyer 2013 2

1. INTRODUCTION. This course contains sound pressure and sound power data for
mechanical equipment commonly found in many commercial buildings. Where possible,
the noise data have been correlated with some of the more obvious noise influencing
parameters, such as type, speed, power rating, and flow conditions. The noise levels
quoted in this publication are suggested for design uses; these noise levels represent
approximately the 80 to 90 percentile values. That is, on the basis of these sample
sizes, it would be expected that the noise levels of about 80 to 90 percent of a random
selection of equipment would be equal to or less than the design values quoted in the
manual, or only about 10 to 20 percent of a random selection would exceed these
values. This is judged to be a reasonable choice of design values for typical uses.
Higher percentile coverage, such as 95 percent, would give increased protection in the
acoustic design, but at greater cost in weight and thickness of walls, floors, columns,
and beams. On-site power plants driven by reciprocating and gas turbine engines have
specific sound and vibration problems, which are considered separately elsewhere.

2. SOUND PRESSURE AND SOUND POWER LEVEL DATA. In the collection of data,
most noise levels were measured at relatively close-in distances to minimize the
influence of the acoustic conditions of the room and the noise interference of other
equipment operating in the same area.

2.1 NORMALIZED CONDITIONS FOR SPL DATA. Note: All measurements were
normalized to a common MER condition by selecting a distance of 3 feet and a Room
Constant of 800 ft2 as representative. SPL data measured at other distances and Room
Constants were brought to these normalized conditions by procedures of discussed
elsewhere.

2.2 SOUND POWER LEVEL DATA. For equipment normally located and used
outdoors, outdoor measurements were made and sound power level data are given.
Usually, more measurements and a more detailed estimate of the measurement
conditions were involved in deriving the PWL data, so they are believed to have a
slightly higher confidence level than the normalized SPL data.

J. Paul Guyer 2013 3

2.3 A-WEIGHTED SOUND LEVELS. In the tables and figures that follow, A-weighted
sound levels are also given. Where sound pressure levels are given, the A-weighted
sound level is in pressure; where sound power levels are given, the A-weighted value is
in sound power. A-weighted sound levels are useful for simply comparing the noise
output of competitive equipment. For complete analysis of an indoor or outdoor noise
problem, however, octave band levels should be used.

2.4 MANUFACTURERS NOISE DATA. Whenever possible, and especially for new
types of equipment, the manufacturer should be asked to provide sound level data on
the equipment. If the data show remarkably lower noise output than competitive models
or are significantly lower than the data quoted in this publication, the manufacturer
should be asked to give guarantees of the noise data and to specify the conditions
under which the data were measured and/or computed.

3. PACKAGED CHILLERS WITH RECIPROCATING COMPRESSORS. These units

range in size from 15-ton to 200-ton cooling capacity. The noise levels have been
reduced to the normalized 3 foot distance from the acoustic center of the assembly. In
terms of noise production, the measured compressors are divided into two groups: up to
50 tons and over 50 tons. The suggested 80- to 90-percentile noise level estimates are
given in Figure 1 and in Table 1 for the two size ranges selected. Although major
interest is concentrated here on the compressor component of a refrigeration machine,
an electric motor is usually the drive unit for the compressor. The noise levels attributed
here to the compressor will encompass the drive motor most of the time, so these
values are taken to be applicable to either a reciprocating compressor alone or a motor-
driven packaged chiller containing a reciprocating compressor.

4. PACKAGED CHILLERS WITH ROTARY-SCREW COMPRESSORS. The octave

band sound pressure levels (at 3 foot distance) believed to represent near-maximum
noise levels for rotary-screw compressors are listed in Table 2. These data apply for the
size range of 100- to 300-ton cooling capacity, operating at or near 3600 RPM.

J. Paul Guyer 2013 4

 Figure 1
Sound pressure levels of reciprocating compressors at 3-ft. distance

J. Paul Guyer 2013 5

 Table 1
Sound pressure levels (in db at 3-ft. distance) for packaged chillers
with reciprocating compressors

J. Paul Guyer 2013 6

 Table 2
Sound pressure levels (in db at 3-ft. distance) for
packaged chillers with rotary screw compressors

5. PACKAGED CHILLERS WITH CENTRIFUGAL COMPRESSORS. These

compressors range in size from 100 tons to 4000 tons and represent the leading
manufacturers. The noise levels may be influenced by the motors, gears, or turbines,
but the measurement positions are generally selected to emphasize the compressor
noise. The noise levels given in Figure 2 and Table 3 represent the 80- to 90- percentile
values found when the units were divided into the two size groups: under 500 tons and
500 or more tons. The low-frequency noise levels reflect the increased noise found for
off-peak loads for most centrifugal machines. These data may be used for packaged
chillers, including their drive units. For built-up assemblies, these data should be used

J. Paul Guyer 2013 7

for the centrifugal compressor only and the suggestions below followed for combining
the noise of other components.

Figure 2
Sound pressure levels of centrifugal compressors at 3-ft. distance

J. Paul Guyer 2013 8

 Table 3
Sound pressure levels (in db at 3-ft. distance) for packaged chillers
with centrifugal compressors

6. BUILT-UP REFRIGERATION MACHINES. The noise of packaged chillers, as

presented in the preceding paragraphs, includes the noise of both the compressor and
the drive unit. If a refrigeration system is built up of separate pieces, then the noise level
estimate should include the noise of each component making up the assembly.
Compressor noise levels should be taken from the packaged chiller data. Sound level
data for the drive units (motors, gears, steam turbines) should be taken from the
appropriate tables in the manual or obtained from the manufacturers. Decibel addition
should be used to determine each octave band sum from the octave band levels of the
various components. The acoustic center should be assumed to be at the approximate
geometric center of the assembly, and distances should be extrapolated from that point.
For very close distances (such as 2 to 3 feet) to each component, assume the total
sound levels apply all around the equipment at distances of 3 feet from the approximate

J. Paul Guyer 2013 9

geometric centers of each component, although this assumption will not provide exact
close-in sound levels.

7. ABSORPTION MACHINES. These units are normally masked by other noise in a

mechanical equipment room. The machine usually includes one or two small pumps;
steam flow noise or steam valve noise may also be present. The 3 foot distance SPLs
for most absorption machines used in refrigeration systems for buildings are given in
Table 4.

Table 4
Sound pressure levels (in db at 3-ft. distance) for absorption machines

J. Paul Guyer 2013 10

8. BOILERS.

8.1 NOISE DATA. The estimated noise levels given in Table 5 are believed applicable
for all boilers, although some units will exceed these values and, certainly, many units
will be much lower than these values. These 3 foot noise levels apply to the front of the
boiler, so when other distances are of concern, the distance should always be taken
from the front surface of the boiler. Noise levels are much lower off the side and rear of
the typical boiler. The wise variety of blower assemblies, air and fuel inlet arrangements,
burners, and combustion chambers provides such variability in the noise data that it is
impossible simply to correlate noise with heating capacity.

8.2 BOILER RATING. Heating capacity of boilers may be expressed in different ways:
sq. ft. of heating surface, BTU/hour, lb of steam/hour, or bhp boiler horsepower). To a
first approximation, some of these terms are interrelated as follows:

33,500 BTU/hour = 1 bhp

33 lb of steam/hour = 1 bhp.

In this publication, all ratings have been reduced to equivalent bhp.

9. STEAM VALVES. Estimated noise levels are given in Table 6 for a typical thermally
insulated steam pipe and valve. Even though the noise is generated near the orifice of
the valve, the pipes on either side of the valve radiate a large part of the total noise
energy that is radiated. Hence, the pipe is considered, along with the valve, as a part of
the noise source. Valve noise is largely determined by valve type and design, pressure
and flow conditions, and pipe wall thickness. Some valve manufacturers can provide
valve noise estimated for their products.

10. COOLING TOWERS AND EVAPORATIVE CONDENSERS. The generalizations

drawn here may not apply exactly to all cooling towers and condensers, but the data are

J. Paul Guyer 2013 11

useful for laying out cooling towers and their possible noise control treatments. It is
desirable to obtain from the manufacturer actual measured noise levels for all directions
of interest, but if these data are not forthcoming, it is essential to be able to approximate
the directional pattern of the cooling tower noise. For aid in identification, four general
types of cooling towers are sketched in Figure 3:

A.) The centrifugal-fan blow-through type;

B.) The axial-flow blow-through type (with the fan or fans located on a side wall);
C.) The induced-draft propeller type; and
D.) The underflow forced draft propeller type (with the fan located under the assembly.

10.1 SOUND POWER LEVEL DATA. Sound power level data are given for both
propeller-type and centrigual-fan cooling towers.

10.1.1 PROPELLER-TYPE COOLING TOWER. The approximate overall and A-

weighted sound power levels of propeller-type cooling towers are given by Equations 1
and 2, respectively: for overall PWL (propeller-type),

Lw = 95 + 10 log (fan hp), (Eq. 1)

and for A-weighted PWL,

Lwa = 86 + 10 log (fan hp), (Eq. 2)

where fan hp is the nameplate horsepower rating of the motor that drives the fan.
Octave band PWLs can be obtained by subtracting the values of Table 7 from the
overall PWL.

J. Paul Guyer 2013 12

10.1.2 CENTRIFUGAL FAN COOLING TOWER. The approximate overall and A-
weighted sound power levels of centrifugal-fan cooling towers are given by Equations 3
and 4, respectively: for overall PWL (centrifugal-fan),

Lw = 85 + 10 log (fan hp) (Eq. 3)

for A-weighted PWL,

Lwa = 78 + 10 log (fan hp). (Eq. 4)

Table 5
Sound pressure levels (in db at 3-ft. distance from
the front) for boilers

J. Paul Guyer 2013 13

 Table 6
Sound pressure levels (in db at 3-ft. distance) for
high-pressure thermally insulated steam valves
and nearby piping

J. Paul Guyer 2013 14

 Table 7
Frequency adjustments (in db) for propeller-type
cooling towers

J. Paul Guyer 2013 15

 Table 8
Frequency Adjustments (in dB) for Centrifugal-Fan
Cooling Towers

J. Paul Guyer 2013 16

 Figure 3
Principal types of cooling towers

When more than one fan or cooling tower is used, fan hp should be the total motor-
drive hp of all fans or towers. Octave band PWLs can be obtained by subtracting the
values of table C-8 from the overall PWL.

10.2 SPLS AT A DISTANCE. Cooling towers usually radiate different amounts of

sound in different directions, and the directional corrections of Table 9 should be made
to the average SPL. These corrections apply to the five principal directions from a
cooling tower, i.e., in a direction perpendicular to each of the four sides and to the top of

J. Paul Guyer 2013 17

the tower. If it is necessary to estimate the SPL at some direction other than the
principal directions, it is common practice to interpolate between the values given for the
principal directions.

Table 9
Correction to average SPLs for directional effects of cooling towers

J. Paul Guyer 2013 18

10.3 CLOSE-IN SPLS. Sound power level data usually will not give accurate calculated
SPLs at very close distances to large-size sources, such as cooling towers. The data
may be used where it is required to estimate close-in SPLs at nearby walls, floors, or in
closely confined spaces (at 3- to 5 foot distances from inlet and discharge openings).

10.4 HALF-SPEED OPERATION. When it is practical to do so, the cooling tower fan
can be reduced to half speed in order to reduce noise; such a reduction also reduces
cooling capacity. Half-speed produces approximately two-thirds cooling capacity and
approximately 8- to 10-dB noise reduction in the octave bands that contain most of the
fan-induced noise. For half-speed operation, the octave band SPLs or PWLs of full-
speed cooling tower noise may be reduced by the following amounts, where fB is the
blade passage frequency and is calculated from the relation fB = No. of fan blades x
shaft RPM/GO.

If the blade passage frequency is not known, it may be assumed to fall in the 63-Hz
band for propeller type cooling towers and in the 250-Hz band for centrifugal cooling
towers. Waterfall noise usually dominates in the upper octave bands, and it would not
change significantly with reduced fan speed.

J. Paul Guyer 2013 19

10.5 LIMITATIONS.

10.5.1 DESIGN VARIATIONS. The data given here represent a fairly complete
summary of cooling tower noise, but it must still be expected that noise levels may vary
from manufacturer to manufacturer, and from model to model as specific design
changes take place. Whenever possible, request the manufacturer to supply the specific
noise levels for the specific needs.

10.5.2 ENCLOSED LOCATIONS. Most of the preceding discussion assumes that

cooling towers will be used in outdoor locations. If they are located inside enclosed
mechanical equipment rooms or within courts formed by several solid walls, the sound
patterns will be distorted. In such instances, the PWL of the tower (or appropriate
portions of the total PWL) can be placed in that setting, and the enclosed or partially
enclosed space can be likened to a room having certain estimated amounts of reflecting
and absorbing surfaces. Because of the limitless number of possible arrangements, this
is not handled in a general way, so the problem of partially enclosed cooling towers is
not treated here. In the absence of a detailed analysis of cooling tower noise levels
inside enclosed spaces, it is suggested that the close-in noise levels of Table 10 be
used as approximations.

10.6 EVAPORATIVE CONDENSERS. Evaporative condensers are somewhat similar

to cooling towers in terms of noise generation. A few evaporative condensers have
been included with the cooling towers, but not enough units have been measured to
justify a separate study of evaporative condensers alone. In the absence of noise data
on specific evaporative condensers, it is suggested that noise data be used for the most
nearly similar type and size of cooling tower.

10.7 AIR-COOLED CONDENSERS. For some installations, an outdoor air-cooled

condenser may serve as a substitute for a cooling tower or evaporative condenser. The
noise of an air-cooled condenser is made up almost entirely of fan noise and possibly
air-flow noise through the condenser coil decks. In general, the low-frequency fan noise

J. Paul Guyer 2013 20

dominates. Since most of the low-frequency noise of a typical cooling tower is due to
the fan system, in the absence of specific data on air-cooled condensers, it is suggested
that noise data be used for the most nearly similar type and size of cooling tower.

10.8 EJECTOR-TYPE COOLING TOWER. This is a fanless type cooling tower that
induces air flow through the use of nozzles of high-pressure water spray. Noise levels
are generally lower for the ejector cooling tower than for cooling towers using fans to
produce air flow. Adequate vibration isolation of the drive pump, piping, and tower are
necessary, although the elimination of the fan reduces the severity of tower vibration.

11. PUMPS. The overall and A-weighted 3 foot SPLs given in Table 10. The pump
power rating is taken as the nameplate power of the drive motor. This is easily
determined in field measurements, whereas actual hydraulic power would be unknown
in a field situation. For pump ratings under 100 hp, the radiated noise increases with the
function (10 log hp), but about 100 hp the noise changes more slowly with increasing
power, hence, the function (3 log hp). Octave band SPLs are obtained by subtracting
the values of Table 11 from the overall SPLs of Table 10. Pumps intended for high-
pressure operation have smaller clearances between the blade tips and the cutoff edge
and, as a result, may have higher noise peaks than shown in Tables 11 and 12 (by 5
dB, sometimes 10dB) in the octave bands containing the impeller blade passage
frequency and its first harmonic. These would usually fall in the 1,000 and 2,000 Hz
octave bands. The data of Tables 11 and 12 are summarized in Figure 4.

J. Paul Guyer 2013 21

 Table 10
Overall and A-Weighted Sound Pressure Levels (in dB and dB(A)
at 3-ft. Distance) for Pumps

J. Paul Guyer 2013 22

 Table 11
Frequency Adjustments fin dB) for Pumps

J. Paul Guyer 2013 23

 Figure 4
Sound Pressure Levels of Pumps at 3-ft. Distance

12. FANS

12.1 IN-DUCT NOISE. Recent issues of ASHRAE publications provide updated

methods for estimating the in-duct noise of ventilating fans. Manufacturers also furnish
in-duct PWL data of their fans on request. A current ASHRAE estimation is given by
Equation 5:

Lw = Kw + 10 log Q + 20 log P + BFI + C (Eq. 5)

where Lw the in-duct sound power level of the fan at either the inlet or discharge end of
the fan, Kw the specific sound power level for the particular fan design, Q is the volume
flow rate in cfm (ft.3/min.), and P is the static pressure produced by the fan (inches of
water gage). Values of Kw for the octave bands and for various basic fan blade designs

J. Paul Guyer 2013 24

are given in Part A of Table 12. The blade passage frequency of the fan is obtained
from:

fan RPM x no. of blades/60

and the blade frequency increment BFI (in dB) is added to the octave band sound
power level in the octave in which the blade passage frequency occurs. It is best to
obtain the number of blades and the fan rotational speed from the manufacturer to
calculate the blade passage frequency. In the event this information is not available,
Part B of Table 12 provides the usual blade passage frequency. The estimates given by
this method assume ideal inlet and outlet flow conditions and operation of the fan at its
design condition. The noise is quite critical to these conditions and increases
significantly for deviations from ideal conditions. Part C of Table 12 provides a
correction factor for off-peak fan operation.

12.2 NOISE REDUCTION FROM FAN HOUSING. The fan housing and its nearby
connected ductwork radiate fan noise into the fan room. The amount of noise is
dependent on both internal and external dimensions of the housing and ductwork, the
TL of the sheet metal, and the amount of sound absorption material inside the ductwork.
Because of so many variables, there is no simple analysis procedure for estimating the
PWL of the noise radiated by the housing and ductwork. However, Table 13 offers a
rough estimate of this type of noise. These are simply deductions, in dB, from the induct
fan noise. At low frequency, the housing appears acoustically transparent to the fan
noise, but as frequency increases, the TL of the sheet metal becomes increasingly
effective.

J. Paul Guyer 2013 25

 Table 12
Specific Sound Power Levels Kw (in dB), Blade Frequency Increments (in dB) and Off
Peak Correction for Fans of Various Types, for Use in Equation 5

J. Paul Guyer 2013 26

 Table 13
Approximate Octave-Band Adjustments for Estimating the PWL of Noise Radiated by a
Fan Housing and its Nearby Connected Duct Work

13. AIR COMPRESSORS. Two types of air compressors are frequently found in
buildings: one is a relatively small compressor (usually under 5 hp) used to provide a
high pressure air supply for operating the controls of the ventilation system, and the
other is a medium size compressor (possibly up to 100 hp) used to provide shop air to
maintenance shops, machine shops, and laboratory spaces, or to provide ventilation
system control pressure for large buildings. Larger compressors are used for special
industrial processes or special facilities, but these are not considered within the scope
of this publication. The 3 foot SPLs are given in Figure 5 and Table 14.

J. Paul Guyer 2013 27

 Figure 5
Sound Pressure Levels of Air Compressors at 3-ft. Distance

J. Paul Guyer 2013 28

 Table 14
Sound Pressure Levels (in dB at 3-ft. Distance) for
Air Compressors

14. RECIPROCATING ENGINES. A comprehensive study has been made of the noise
characteristics of reciprocating and turbine engines fueled by natural gas and liquid fuel.
The noise levels of the engines as sound sources are summarized here, because these
engines may be used as power sources in buildings, and their noise should be taken

J. Paul Guyer 2013 29

into account. Typically, each engine type has three sound sources of interest; the
engine casing, the air inlet into the engine, and the exhaust from the engine.

14.1 ENGINE CASING. The PWL of the noise radiated by the casing of a natural-gas
or diesel reciprocating engine is given by Equation 6:

Lw = 93 + 10 log (rated hp) + A +B + C + D , (Eq. 6)

where Lw is the overall sound power level (in dB),rated hp is the engine
manufacturers continuous full-load rating for the engine (in horsepower), and A, B, C,
and D are correction terms (in dB), given in Table 15. Octave band PWLs can be
obtained by subtracting the Table 16 values from the overall PWL given by Equation 6.
The octave band corrections are different for the different engine speed groups. For
small engines (under about 450 hp), the air intake noise is usually radiated close to the
engine casing, so it is not easy or necessary to separate these two sources; and the
engine casing noise may be considered as including air intake noise (from both
naturally aspirated and turbocharged engines).

J. Paul Guyer 2013 30

 Table 15
Correction Terms (in dB) to be Applied to Equation 6 for Estimating the Overall PWL of
the Casing Noise of a Reciprocating Engine

J. Paul Guyer 2013 31

 Table 16
Frequency Adjustments (in dB) for Casing Noise of Reciprocating Engines

14.2 TURBOCHARGED AIR INLET. Most large engines have turbochargers at their
inlet to provide pressurized air into the engine for increased performance. The
turbocharger is a turbine driven by the released exhaust gas of the engine. The turbine
is a high-frequency sound source. Turbine configuration and noise output can vary
appreciably, but an approximation of the PWL, of the turbocharger noise is given by
Equation 7:

Lw = 94 + 5 log (rated hp) -L/6 (Eq. 7)

where Lw and rated hp are already defined and L is the length, in ft., of a ducted inlet
to the turbocharger. For many large engines, the air inlet may be ducted to the engine
from a fresh air supply or a location outside the room or building. The term L/6, in dB,
suggests that each 6 ft. of inlet ductwork, whether or not lined with sound absorption
material, will provide about 1 dB of reduction of the turbocharger noise radiated from the

J. Paul Guyer 2013 32

open end of the duct. This is not an accurate figure for ductwork in general; it merely
represents a simple token value for this estimate. The octave band values given in
Table 17 are subtracted from the overall PWL of Equation 7 to obtain the octave band
PWLs of turbocharged inlet noise.

Table 17
Frequency Adjustments fin dB) for Turbocharger
Air Inlet Noise

14.3 ENGINE EXHAUST. The PWL of the noise radiated from the unmuffled exhaust of
an engine is given by Equation 8:

Lw = 119 + 10 log (rated hp) - T - L/4 (Eq. 8)

J. Paul Guyer 2013 33

where T is the turbocharger correction term (T = 0 dB for an engine without a
turbocharger and T = 6 dB for an engine with a turbocharger) and L is the length, in ft.,
of the exhaust pipe. A turbocharger takes energy out of the discharge gases and results
in an approximately 6-dB reduction in noise. The octave band PWLs of unmuffled
exhaust noise are obtained by subtracting the values of Table 18 from the overall PWL
derived from Equation 8. If the engine is equipped with an exhaust muffler, the final
noise radiated from the end of the tailpipe is the PWL of the unmuffled exhaust minus
the insertion loss, in octave bands, of the muffler.

Table 18
Frequency Adjustments (in dB) for Unmuffled
Engine Exhaust Noise

J. Paul Guyer 2013 34

15. GAS TURBINE ENGINES

15.1 PWL OF THREE SOURCES. As with reciprocating engines, the three principal
sound sources of turbine engines are: the engine casing, the air inlet, and the exhaust.
Most gas turbine manufactures will provide sound power estimates of these sources.
However when these are unavailable the overall PWLs of these three sources, with no
noise reduction treatments, are given in the following equations: for engine casing
noise,

Lw = 120 + 5 log (rated MW) (Eq. 9)

for air inlet noise,

Lw = 127 + 15 log (rated MW) (Eq. 10)

for exhaust noise

Lw = 133 + 10 log (rated MW) (Eq. 11)

where rated MW is the maximum continuous full-load rating of the engine in

megawatts. If the manufacturer lists the rating in effective shaft horsepower (eshp), the
MW rating may be approximated by

MW = eshp/1400

Overall PWLs, obtained from Equations 9 through 11, are tabulated in Table 19 for a
useful range of MW ratings.

15.1.1 TONAL COMPONENTS. For casing and inlet noise, particularly strong high-
frequency sounds may occur at several of the upper octave bands. However which

J. Paul Guyer 2013 35

bands contain the tones will depend on the specific design of the turbine and, as such,
will differ between models and manufacturers. Therefore, the octave band adjustments
of Table 20 allow for these peaks in several different bands, even though they probably
will not occur in all bands. Because of this randomness of peak frequencies, the A-
weighted levels may also vary from the values quoted.

15.1.2 ENGINE COVERS. The engine manufacturer sometimes provides the engine
casing with a protective thermal wrapping or an enclosing cabinet, either of which can
give some noise reduction. Table 19 suggests the approximate noise reduction for
casing noise that can be assigned to different types of engine enclosures. Refer to the
notes of the table for a broad description of the enclosures. The values of Table 19 may
be subtracted from the octave band PWLs of casing noise to obtain the adjusted PWLs
of the covered or enclosed casing. An enclosure specifically designed to control casing
noise can give larger noise reduction values than those in the table. However it should
be noted that the performance of enclosures that are supported on the same structure
as the gas turbine, will be limited by structure borne sound. For this reason care should
be used in applying laboratory data of enclosure performance to the estimation of sound
reduction of gas turbine enclosures.

15.2 EXHAUST AND INTAKE STACK DIRECTIVITY. Frequently, the exhaust of a gas
turbine engine is directed upward. The directivity of the stack provides a degree of noise
control in the horizontal direction. Or, in some installations, it may be beneficial to point
the intake or exhaust opening horizontally in a direction away from a sensitive receiver
area. In either event, the directivity is a factor in noise radiation. Table 21 gives the
approximate directivity effect of a large exhaust opening. This can be used for either a
horizontal or vertical stack exhausting hot gases. Table 21 shows that from
approximately 0 to 60 degrees from the axis of the stack, the stack will yield higher
sound levels than if there was no stack and the sound were emitted by a non-directional
point source. From about 60 to 135 degrees from the axis, there is less sound level than
if there were no stack. In other words, directly ahead of the opening there is an increase

J. Paul Guyer 2013 36

in noise, and off to the side of the opening there is a decrease in noise. Table 21 values
also apply for a large-area intake opening into a gas turbine for the 0 to 60 degree
range; for the 90 to 135 degree range, subtract an addition 3 dB from the already
negative-valued quantities. For horizontal stacks, sound-reflecting obstacles out in front
of the stack opening can alter the directivity pattern. Even irregularities on the ground
surface can cause some backscattering of sound into the 90 to 180 degree regions, for
horizontal stacks serving either as intake or exhaust openings. For small openings in a
wall, such as for ducted connections to a fan intake or discharge, use approximately
one-half the directivity effect of Table 21 (as applied to intake openings) for the 0 to 90
degree region. For angles beyond 90 degrees, estimate the effect of the wall as a
barrier.

16. ELECTRIC MOTORS. Motors cover a range of 1 to 4000 hp and 450 to 3600 RPM.
The data include both drip-proof (DRPR) (splash-proof or weather-protected) and
totally enclosed fan-cooled (TEFC) motors. Noise levels increase with power and
speed.

J. Paul Guyer 2013 37

 Table 19
Overall PWLs of the Principal Noise Components of Gas Turbine Engines Having No
Noise Control Treatments

J. Paul Guyer 2013 38

 Table 20
Frequency Adjustments (in dB) for Gas Turbine Engine Noise Sources

J. Paul Guyer 2013 39

 Table 21
Approximate Directivity Effect fin dB) of a Large Exhaust Stack Compared to a
Nondirectional Source of the Same Power

16.1 TEFC MOTORS. The overall SPLs of TEFC motors, at the normalized 3 foot
condition, follow approximately the relationships of Equations 12 and 13.

For power ratings under 50 hp,

J. Paul Guyer 2013 40

 Lp = 15 + 17 log hp + 15 log RPM (Eq. 12)

and for power ratings above 50 hp,

Lp = 27 + 10 log hp + 15 log RPM (Eq. 13)

where hp is the nameplate motor rating in horsepower and RPM is the motor shaft
speed. For motors above 400 hp, the calculated noise value for a 400-hp motor should
be used. These data are not applicable to large commercial motors in the power range
of 1000 to 5000 hp. The octave band corrections for TEFC motors are given in Table
22. The data of Equations 12 and 13 and Table 22 are summarized in Figure 6, which
gives the SPLs at 3 foot distance for TEFC motors for a working range of speeds and
powers. Some motors produce strong tonal sounds in the 500, 1,000, or 2,000 Hz
octave bands because of the cooling fan blade frequency. Table 22 and Figure 6 allow
for a moderate amount of these tones, but a small percentage of motors may still
exceed these calculated levels by as much as 5 to 8 dB. When specified, motors that
are quieter than these calculated values by 5 to 10 dB can be purchased.

J. Paul Guyer 2013 41

 Table 22
Frequency Adjustments (in dB) for TEFC Electric
Motors.

J. Paul Guyer 2013 42

 Figure 6
Sound pressure levels of TEFC motors at 3-ft. distance.

16.2 DRPR MOTORS. The overall SPLs of DRPR motors, at the normalized 3 foot
condition, follow approximately the relationships of Equations 14 and 15.

For power ratings under 50 hp,

Lp = 10 + 17 log hp + 15 log RPM. (Eq. 14)

and for power ratings above 50 hp,

Lp = 22 + 10 log hp + 15 log RPM. (Eq. 15)

J. Paul Guyer 2013 43

For motors above 400 hp, the calculated noise value for a 400 hp motor should be
used. The octave band corrections for DRPR motors are given in Table 23. The data of
Equations 14 and 15 and Table 23 are summarized in Figure 7, which gives the SPLs at
3 foot distance for DRPR motors over a range of speeds and powers.

Table 23
Frequency Adjustments (in dB) for DRPR Electric Motors.

J. Paul Guyer 2013 44

 Figure 7
Sound Pressure Levels of DRPR Motors at 3 ft. Distance.

17. STEAM TURBINES. Noise levels are found generally to increase with increasing
power rating, but it has not been possible to attribute any specific noise characteristics
with speed or turbine blade passage frequency (because these were not known on the
units measured). The suggested normalized SPLs at 3 foot distance are given in Figure
8 and Table 24.

J. Paul Guyer 2013 45

 Table 24
Sound Pressure Levels (in dB at 3 ft distance) for steam turbines.

J. Paul Guyer 2013 46

 Figure 8
Sound Pressure Levels of Steam Turbines at 3 ft. Distance.

18. GEARS. It is generally true that the noise output increases with increasing speed
and power but it is not possible to predict in which frequency band the gear tooth
contact rate or the ringing frequencies will occur for any unknown gear. The possibility
that these frequency components may occur in any of the upper octave bands is
covered by, Equation 16, which gives the octave band SPL estimate (at the 3 feet
normalized condition) for all bands at and above 125 Hz:

Lp = 78 + 3 log (RPM) + 4 log (hp) (Eq. 16)

where RPM" is the speed of the slower gear shaft and hp is the horsepower rating of
the gear or the power transmitted through the gear. For the 63 Hz band, 3 dB is
deducted; and for the 31 Hz band, 6 dB is deducted from the Equation 16 value. This
estimate may not be highly accurate, but it will provide a reasonable engineering

J. Paul Guyer 2013 47

evaluation of the gear noise. Table 25 gives the estimated SPL in the 125 through 8,000
Hz bands for a variety of speeds and powers, based on Equation 16.

Table 25
Approximate Sound Pressure Levels (in dB at 3-ft. Distance) for Gears, in the 125-
through 8000-Hz Octave Bands, from Equation 16.

19. GENERATORS. The noise of generators, in general, can be quite variable,

depending on speed, the presence or absence of air cooling vanes, clearances of
various rotor parts, etc., but, most of all, on the driver mechanism. When driven by gas
or diesel reciprocating engines, the generator is usually so much quieter than the
engine that it can hardly be measured, much less heard. For gas turbine engines the
high-speed generator may be coupled to the engine through a large gear, and the gear
and the generator may together produce somewhat indistinguishable noise in their
compartment, which frequently is separated by a bulk head from the engine
compartment. Table 26 gives an approximation of the overall PWL of several
generators. It is not claimed that this is an accurate estimate, but it should give

J. Paul Guyer 2013 48

reasonable working values of PWL. It is to be noted that the PWL of the generator is
usually less than that of the drive gear and less than that of the untreated engine casing.
Octave band corrections to the overall PWL are given in Table 27.

Table 26
Approximate Overall PWL (in dB) of Generators, Excluding the Noise of the Driver Unit.

J. Paul Guyer 2013 49

 Table 27
Frequency Adjustments (in dB) for Generators, Without Drive Unit.

J. Paul Guyer 2013 50

20. TRANSFORMERS. The National Electrical Manufacturers Association (NEMA)
provides a means of rating the noise output of transformers. The NEMA audible sound
level, as it is called in the standard, is the average of several A-weighted sound levels
measured at certain specified positions. The NEMA sound level for a transformer can
be provided by the manufacturer. On the basis of field studies of many transformer
installations, the PWL in octave bands has been related to the NEMA rating and the
area of the four side walls of the unit. This relationship is expressed by Equation 17:

Lw = NEMA rating + 10 log A + C, (Eq. 17)

where NEMA rating is the A-weighted sound level of the transformer provided by the
manufacturer, obtained in accordance with current NEMA Standards, A is the total
surface area of the four side walls of the transformer in ft2, and C is an octave band
correction that has different values for different uses, as shown in Table 28. If the exact
dimensions of the transformer are not known, an approximation will suffice. If in doubt,
the area should be estimated on the high side. An error of 25 percent in area will
produce a change of 1 dB in the PWL. The most nearly applicable C value from Table
28 should be used. The Cl value assumes normal radiation of sound. The C2 value
should be used in regular-shaped confined spaces where standing waves will very likely
occur, which typically may produce 6 dB higher sound levels at the transformer
harmonic frequencies of 120, 240, 360, 480, and 600 Hz (for 60-Hz line frequency; or
other sound frequencies for other line frequencies). Actually, the sound power level of
the transformer does not increase in this location, but the sound analysis procedure is
more readily handled by presuming that the sound power is increased. The C3 value is
an approximation of the noise of a transformer that has grown noisier (by about 10 dB)
during its lifetime. This happens occasionally when the laminations or tie-bolts become
loose, and the transformer begins to buzz or rattle. In a highly critical location, it would
be wise to use this value. All of the Table 28 values assume that the transformer initially
meets it quoted NEMA sound level rating. Field measurements have shown that
transformers may actually have A-weighted sound levels that range from a few decibels

J. Paul Guyer 2013 51

(2 or 3 dB) above to as much as 5 or 6 dB below the quoted NEMA value. Quieted
transformers that contain various forms of noise control treatments can be purchased at
as much as 15 to 20 dB below normal NEMA ratings. If a quieter transformer is
purchased and used, the lowered sound level rating should be used in place of the
regular NEMA rating in Equation 17, and the appropriate corrections from Table 28
selected.

Table 28
Octave-Band Corrections (in dB) to be Used in Equation 17 for Obtaining PWL of
Transformers in Different Installation Conditions.

J. Paul Guyer 2013 52

21. OPENING IN A WALL. An opening, such as a door, window, or louvered vent, in
an exterior wall of a noisy room will allow noise to escape from that room and perhaps
be disturbing to neighbors. The PWL of the sound that passes through the opening can
be estimated from equation 18:

Lw = Lp + 10 log A - 10 (Eq. 18)

where Lp is the SPL in the room at the location of the opening and A is the area, in ft2,
of the opening. (Note, the factor of -10 is due to the use of ft2 for A, if m2 had been used
then this factor would be 0). Once the PWL is estimated, the SPL at any neighbor
distance can be estimated. For normal openings (windows or vents) without ducted
connections to the noise source, it may be assumed that the sound radiates freely in all
directions in front of the opening, but to the rear of the wall containing the opening, the
barrier effect of the wall should be taken into account. For ducted connections from a
sound source to an opening in the wall, the sound is somewhat beamed out of the
opening and may be assumed to have a directivity effect of above one-half the amount
given in Table 21 for air intake openings of large stacks.

J. Paul Guyer 2013 53

22. GLOSSARY

J. Paul Guyer 2013 54

 J. Paul Guyer 2013 55