Fan noise

The principle source of noise in any air moving system is the main fan. Rules for determining fan noise and noise-producing mechanisms are covered as well as a review of the sound laws. If the ducting resistance has been incorrectly assessed, the fan noise can be significantly affected.

This Chapter points out some of the pitfalls in the selection of ductwork of the ventilation system which contribute to the addition of unforeseen noise.

A prime source of noise in any air moving system is the main fan. It has the ability to direct its duct-borne noise to the farthest corners of any occupied space and can be a major irritant. The problem can, of course, be magnified by the addition of system generated noise. To the humble fan engineer, it seems remark­able from a noise point-of-view, therefore, that so little apparent attention is given, in the design of a ventilation system, to the correct selection of the fan. To this must be added the often less than ideal ductwork connections to the fan, which can result in an additional unforeseen noise.

It is the intention of this Chapter to point out some of the pitfalls and to suggest that the requisite information be obtained from a reputable manufacturer at the earliest possible time. Unfortu­nately this is not always possible, as the fan supplier will only be chosen late in the building programme when much of the de­sign has been “frozen”. It would be beneficial, however, to con­duct a feasibility study using results obtained from experiments beforehand.

The user’s primary aim is to ensure that the fan will satisfactorily perform its duty. That is to say, it will handle the required volume flowrate at the system pressure and for the stated power. Even more important, however, is what nuisance will be caused, by its noise, to operators of the plant, to neighbours, or to inhabit­ants of the conditioned area. So many misconceptions, half-truths, and errors have been propagated in the field of acoustics, that one might imagine it had replaced alchemy as the “black art” of 20th century man.

This Chapter is not intended to be a textbook of noise measure­ment, and those who wish to know more are referred to the ref­erences in Section 14.15. However, in orderto give meaningful information, it is worth reminding the user of some of the terms employed and their values and underlying concepts.

What is noise?

Noise may simply be defined as:

Sound undesired by the recipient.

What is sound?

Sound may be defined as any pressure variation in a medium — usually air — that can be converted into vibrations by the human eardrum, causing signals to be sent to the brain. As with all other sensations, the result can be pleasant or unpleasant.

Frequency

To vibrate the eardrum it is necessary for the pressure varia­tions in the medium to occur rapidly. The number of variations per second is called the frequency of the sound, measured in cycles per second or Hertz. The human ear can detect sounds from about 20 Hz to 20,000 Hz — the lowest and highest sounds respectively. As a guide, the lowest note on a piano has a fre­quency of 27.5 Hz, whilst the highest note is at 4186 Hz.

Sound power level (SWL)

The noisiness of a fan can be expressed in terms of its sound power (the number of watts of power it converts into noise). It is unusual to do this, however, as the range of values found in practice would be very large. Fan noise can be measured by its sound power level, a ratio which logarithmically compares its sound power with a reference power, the Pico Watt (10-12 watts). The unit of sound power level is the decibel.

Sound power level may be defined as:

SWL =10 log——————————— Equ 14.1

W0

Where:

SWL = sound power level in decibels (re 10-12 watts)

W = sound power of the noise generating equip­

Ment (watts)

W0 = reference power (re 10’12 watts)

Table 14.1 shows how the logarithmic scale compresses the wide range of possible sound powers to sound power levels having a practical range of 30 dBW to 200 dBW.

Sound Power (Watts)

Sound power level dBW

Source

40 000 000

196

Saturn rocket

100 000

170

Ramjet

10 000

160

Turbo jet engine 3200 kg thrust

1 000

150

4 propeller airliner

100

140

10

130

Full orchestra

1

120

Large chipping hammer

0.1

110

Blaring radio

0.01

100

Car on motorway

0.001

90

10 kW ventilating fan

0.0001

80

Voice — shouting

0.00001

70

Voice — conversational level

0.000001

60

0.0000001

50

0.00000001

40

0.000000001

30

Voice — very soft whisper

Table 14.1 Sound powers expressed as sound power levels

Sound pressure level (SPL)

The sound power level of a fan is comparable to the power out­put of a heater. Both measure the energy (in one case — noise energy, the other — heat energy) fed into the environment sur­rounding them. However, neither the sound power level nor the power output will tell us the effect on a human being in the sur­rounding space.

In the case of a heater, the engineer, by considering the volume of the surroundings, the materials of the room, and what other heat sources are present, can determine the resulting tempera­ture at any point. In a similar way, the acoustic engineer, by considering very similar criteria, can calculate the sound pres­sure level at any point. (Remember, it is sound pressure that vi­brates the eardrum membrane and determines how we hear a noise.)

Sound pressure levels are also measured on a logarithmic scale but the unit is the decibel re 2 x 10‘5 Fa. There is another advantage in using the decibel scale. Because the ear is sensi­tive to noise in a logarithmic fashion, the decibel scale more nearly represents how we respond to a noise.

SPL = 20 log — Equ 14.2

Po

Where:

SPL = sound pressure level in decibels (re 2 x 10 5 Fa)

P = sound pressure of the noise (Pa)

P0 = reference pressure (= 2 x10-5 Pa)

It should be realised that in specifying a sound pressure level, the distance from a noise source is implied or stated. In Table

14.1 The position of the observer relative to the source is indi­cated.

Sound pressure Pa

Sound pressure level dB

Typical environment

200.0

140

30 m from military aircraft at take-off

63.0

130

Pneumatic chipping and riveting (operator’s position)

20.0

120

Boiler shop (maximum levels)

6.3

110

Automatic punch press (operator’s position)

2.0

100

Automatic lathe shop

0.63

90

Construction site — pneumatic drilling

0.2

80

Kerbside of busy street

0.063

70

Loud radio (in average domestic room)

0.02

60

Restaurant

0.0063

50

Conversational speech at 1 m

0.002

40

Whispered conversation at 2 m

0.00063

30

0.0002

20

Background in TV and recording studios

0.00002

0

Normal threshold of hearing

Table 14.2 The position of the observer relative to the source

Note: The engineer must clearly distinguish and understand the difference between sound power level and sound pressure level. He must also appreciate that dB re 10’12 watts and dB re 2 x 10’5 Pa are different units.

It is impossible to measure directly the sound power level of a fan. However, the manufacturer can calculate this level after measuring the sound pressure levels in each octave band with the fan working in an accepted standard acoustic test rig.

What he cannot do is unequivocally state what sound pressure levels will result from the use of the fan. This can only be done if details of the way the fan is to be used, together with details of the environment it is serving, are known and a detailed acoustic analysis is carried out.

Octave bands

Noise usually consists of a mixture of notes of different frequen­cies, and because these different frequencies have different characteristics a single sound power level is not sufficient in it­self to describe the intensity and quality of a noise.

Noise is therefore split up into octave bands (bands of fre­quency in which the upper frequency is twice that of the lowest) and a sound pressure level is quoted for each of the bands. The octave band frequencies universally recommended have mid-frequencies of 63, 125, 250, 500, 1000, 2000, 4000, and 8000 Hz.

It is now becoming an increasing requirement for data at 31.5 Hz and 16000 Hz to also be included, although for a number of reasons the former is exceedingly difficult to measure with any degree of certainty.

The noisiness of a fan is specified by a number of sound power levels (in decibels re 10-12 watts), each corresponding to an oc­tave band of frequencies. For research and other purposes it is also possible to measure the noise in more precise bands e. g. y3 octave or at so-called discrete frequencies.

As with sound power levels, sound pressure levels must be quoted for each octave band if a complete picture of the effect of the noise on the human ear is required.

How does sound spread?

The effect of a sound source such as a fan on its environment can be likened to dropping a pebble into a pond. Ripples will spread out uniformly in all directions and will decrease in height as they move from the point where the pebble was dropped. Normally the ripples will be circular in shape unless affected by some barrier. See Figure 14.1

Sound

Source

Fan noise

///

——— T^n————

Absorbed ♦ »Transmitted

Figure 14.1 Sound in a free field (above) and sound incident on a surface (be­low)

It is just the same with a sound source in air. When the distance doubles, the amplitude of the sound halves, and this is a reduc­tion of 6 dB, for using equation 14.2:

Reduction = 20 log — = 20 log 2 = 6 dB

Pi

But the power of the sound source and therefore the SWL is un­changed.

To summarise, if you move from one metre from the source to two metres, the SPL will drop by 6 dB. If you move to four metres it will drop by 12 dB, eight metres by 18 dB, and so on. But this is only true if there are no objects in the path of the sound, which can reflect, or block.

Ideal conditions where the sound can spread unhindered are termed “free field". If there is an object in the way, some of the sound will be reflected, some absorbed, and some transmitted right through. How much is reflected, absorbed, or transmitted depends on the properties of the object, its size, and the partic­ular wavelength of the sound. Generally speaking an object must be larger than one wavelength to have an effect.

I xu Speed of sound =340/s

Wavelength = ———————-

Frequency Hz

For example

340

Sound of 8K Hz : wavelength 340 =- = 0.425 m

8×1000

Sound of 63 Hz: wavelength = = 5.4 m

63

Hence for a high frequency noise even a very small object will disturb the sound field and absorb or isolate it. But low fre­quency noise, whilst less objectionable, is more difficult to block.

Source

/

AA/WvVWVWVWW:

Figure 14.2 Sound in an anechoic chamber

Sound reflecting or reverberation chambers

This is the opposite of the anechoic chamber. All surfaces are made as hard as possible to reflect the noise and all the walls are made at an angle to each other so that there are no parallel surfaces. Thus the sound energy is uniform throughout the room and a “diffuse field” exists. It is therefore possible to mea­sure the SWL, but the SPL measurements in any direction will be meaningless due to the many reflections. Such rooms, see Figure 14.3, are cheaper to build than anechoic chambers and are therefore very popular.

Fan noise

Fan noise

If we wished to make measurements in a free field without any reflections, then the top of a very tall but small cross-section flagpole in the middle of the Sahara desert (after it had been raked flat) would probably be ideal. Obviously there are difficul­ties and an anechoic room is a reasonable alternative. Here the walls, ceiling and floor are covered in a highly sound absorptive material to eliminate any reflections. Thus the SPL in any direction may be measured. See Figure 14.2.

WWWTOWWV Sound 5

подпись: if we wished to make measurements in a free field without any reflections, then the top of a very tall but small cross-section flagpole in the middle of the sahara desert (after it had been raked flat) would probably be ideal. obviously there are difficul-ties and an anechoic room is a reasonable alternative. here the walls, ceiling and floor are covered in a highly sound absorptive material to eliminate any reflections. thus the spl in any direction may be measured. see figure 14.2.
wwwtowwv sound 5
Should be made. Sometimes, however, conditions are so re­verberant or the room so small, that a free field will not be present. Afan in a “real room” is shown diagrammatically in Fig­ure 14.4.

Relationship between sound pressure and sound power levels

SPL = SWL+ 10 log

подпись: spl = swl+ 10 log

Equ 14.3

подпись: equ 14.3

Where:

SPL

SWL

R

Qe

Rc

подпись: where:
spl
swl
r
qe
rc

(m2

подпись: (m2

Fan noise

Figure 14.3 Sound in reverberation chamber

The “real room”

In practice we usually wish to make measurements in a room that is neither anechoic nor reverberant, but somewhere in be­tween. It is then difficult to find a suitable position for measuring the noise from a particular source.

When determining noise from a single fan, several errors are possible. If you measure too closely, the SPL may vary consid­erably with a small change in position when the distance is less than the wavelength of the lowest frequency emitted or less than twice the greatest dimension of the fan, whichever is the greater. This is termed the “near field” and should be avoided.

Other errors arise if measurements are made too far from the fan. Reflections from walls and other objects may be as strong as the direct sound. Readings will be impossible in this rever­berant field. A free field may exist between the reverberant and near field and can be found by seeing if the level drops 6 dB for a doubling in distance from the fan. It is here that measurements

The relationship between SPL and SWL is given as:

‘ R,

47ir

Sound pressure level dB (re 2 x 10’5 Pa)

Sound power level dBW (re lO’12 W)

Distance from the source (m)

Directivity factor of the source in the direction of r

Saa

Room constant =

S = total surface are of the room (m2)

Aav = average absorption coefficient in the room

The first term, within, the brackets is the “direct” sound, whilst the second term is “reflected” sound.

The value of the average absorption coefficient aav can be cal­culated.

If we have an area S, of material in the room having an absorp­tion coefficient a-i, and area S2 with absorption coefficient a2,

And so on, aav = — (S^ +S2a2 +S3a3 +etc)

S

A not only varies with the material, but also differs according to the frequency of the noise. It is therefore necessary to calculate the SPL from the SWL in each frequency. Some typical values of absorption coefficient a can be found in Table 14.3.

— SWL -10 log 2nr

SPL = SWL + 10 log

4nr

Position of source

Directivity factor Qe

Near centre of room

1

At centre of floor

2

Centre of edge between floor and wall

4

Corner between two walls and floor

8

Table 14.4 Values of the directivity factor, assuming fan source in a targe room

 

For special proprietary acoustic materials and all other surface finishes, refer to the manufacturers.

 

Material

Hertz

63

125

250

500

1000

2000

4000

8000

Brickwork

,05

.05

.04

.02

.04

.05

.05

.05

Breezebiock

.1

.2

.45

.6

.4

.45

.4

.4

Concrete

.01

.01

.01

.02

.02

.02

.03

.03

Glazed tiles

0.05

0.05

0.05

0.05

0.05

0.05

0.05

0.05

Plaster

.04

.04

.05

.06

.08

.04

.06

.05

Rubber floor tiles

.05

.05

.05

.1

,1

.05

.05

.05

Table 14.3 Typical values of absorption coefficient a

 

Certain fan manufacturers will quote the sound pressure level of their units at a specified distance — usually 1.5 m or 3 impeller diameters under “free field conditions” and assuming spherical propagation. These would exist if the fan was suspended in space and there were no adjacent floor or walls to either absorb or reflect the noise.

Using the formula in equation 14.3

Qe = 1 and Rc -> oo Thus:

SPL = SWL + 10 log ~ = SWL -10 log 4Ttr2

 

The surface area of a sphere equals Anr2. Thus if the fan is in the geometric centre of the room, its sound will be equally dis­persed over a sphere. If the fan is at the centre of the floor, the sound will be radiated over a half sphere for which the surface are is 2kt2. This is half the previous surface area and thus in­verse of the proportion of the sphere’s surface area. This is known as the directivity factor Qe.

The directivity factor can thus be assessed for all likely fan posi­tions. See Figure 14.5 and Table 14.4.

 

4jtr

 

And

 

If r = 1.5 then SPL = SWL -14.5 dB

Other manufacturers calculate for “hemispherical” propagation under the same free field conditions, i. e. it is assumed that the fan is mounted on a hard reflecting floor. Qe then equals 2.

Thus:

2

 

Fan noise

Tt, = 1

 

And

 

If r = 1.5 then SPL = SWL-11.5 dB

For three diameters, knowing the impeller diameter in metres, the difference in both cases may be calculated. See Figure 14.6.

Whilst these figures may be used as a basis for comparison be­tween different units calculated in the same manner, it must be realised that the SPLs measured on site with a meter may be ei­ther above or below these values. The actual result is as much a function of the room as of the fan characteristics. The analogy of an electric fire in a room with or without heat losses should be remembered.

The internal areas of modem commercial and industrial build­ings have hard boundary surfaces, which cause a high propor­tion of sound energy incident upon them to be reflected and a

 

/

/

/

/

/

° si

A %

F 8 § S S S §

I 1 1

I i

 

Q,, = 2

 

Fan noise Fan noise Fan noise Fan noise

-10

-20

DB

-30

0

High reverberant sound pressure level to be built up. When this
occurs, the sound pressure level readings indicated on a sound
meter are independent of the distance from the noise source.

Understanding the difference between sound power level and
sound pressure level is important, but the engineer must also
know how acceptable levels of sound pressure can be
specified.

It is inconvenient to quote a series of sound values for each ap-
plication. Efforts therefore have been made to express noise
intensity and quality in one single number. The ear reacts differ-
ently according to frequency. All these single figure indices
mathematically weight the sound pressure level values at each
octave band according to the ear’s response at that frequency.

To obtain basic sound pressure level, re 2 x 10 5 Pa under free
field conditions, assuming spherical propagation, measured at
3 fan diameters distance or 1.5 m (whichever is the greater)
from impeller centre, deduct the value indicated by fan diame-
ter from the sound power level (re 10-12 watts).

Weighted sound pressure levels

А, В, C, and D noise levels are an attempt to produce single
number and sound pressure indices. To obtain them, different
values are subtracted from the sound pressure levels in each of
the frequency bands, subtracting most from those bands which
affect the ear least. The results are then added logarithmically
to produce an overall single number sound level. The graphs
(see Figures 14.7 to 14.10), show the different weightings em-
ployed. The resulting noise levels are known respectively as
dBA, dBB, dBC, and dBD.

+10

-50

Figure 14.9 Weighted sound pressure curve С

+10

0

-10

-30

-40

О О о о со о

ООО

ООО

О о ю СО о N

Hz

-50

HP

Figure 14.10 Weighted sound pressure curve D

Theoretically dBA values apply up to levels of 55 dB only, dBB for levels between 55-85 dB only and dBC for higher levels only. dBD is reserved for special noise, e. g., aircraft. However dBA is now used almost exclusively whatever the level. Engineers should check what weighting curves have been used by manu­facturers and, if necessary convert them to a common base be­fore comparisons are made.

A, B, C and D weightings are useful for making initial assess­ments (inexpensive sound level meters are available which measure directly on these scales). Unfortunately too much in­formation is lost in combining all the data into one figure for it to be of use for calculation and design work. Most noise control depends on frequency analysis.

Posted in Fans Ventilation A Practical Guide