Engine Testing || Vibration and Noise

Chapter 9

Vibration and Noise

En

Cop

Chapter Outline
Part 1: Vibration and its
Isolation 20
3
gine

yrig

Vibration
203
Fundamentals: Sources of

Vibration
204
Design of Engine Mountings

and Test-Bed Foundations
207
Practical Considerations in

the Design of Engine

and Test-Bed Mountings
211
Massive Foundations and

Air-Sprung Bedplates
213
Testing

ht � 20

216
Summary of Vibration Section
219
Vibration Notation
219
. DOI: 10.1016/B978-0-08-096949-7.00009-1

12 Elsevier Ltd. All rights reserved.

Part 2: Noise and its

Attenuation in Test Facilities 21
9
Sound Intensity 2
20
Noise Measurements 2
21
Permitted Levels of Noise

in the Workplace 2
22
Noise External to Test Facility

and Planning Regulations 2
23
Noise Reverberation in the

Test Cell Environment 2
23
Noise Notation 2
24
References 224

Useful texts 225

Further Reading 225

PART 1: VIBRATION AND ITS ISOLATION

Note: Vibration is considered in this chapter with particular reference to thedesign and operation of engine test facilities, engine mountings, and theisolation of engine-induced disturbances. Torsional vibration within shaftsystems is covered as a separate subject in Chapter 11.

Vibration

Usually the engine itself is, or should be, the only significant source of vibrationand noise in the engine and powertrain test cell. Secondary sources of noise andvibration within the test cell such as the ventilation system, pumps, and fluidcirculation systems or the dynamometer are usually swamped by the effects ofthe engine.

203

204 Engine Testing

There are several aspects to the problem of engine vibration and noise:

l The engine must be mounted in such a way that neither it, nor connectionsto it, can be damaged by excessive movement or excessive constraint.

l Transmission of engine-induced vibration to and via services connected tothe cell structure or to adjoining buildings must be minimized.

l Noise levels in the cell should be contained as far as possible and the designof alarm systems should take into account ambient noise levels in all relatedwork areas.

Fundamentals: Sources of Vibration

Since the vast majority of engines likely to bemet with are single ormulticylinderin-line vertical engines, we shall being by concentrating on this configuration.

Any unit under test (UUT) may be regarded as having six degrees offreedom of vibration about orthogonal axes through its center of gravity: linearvibrations along each axis and rotations about each axis (see Figure 9.1). Inpractice only three of these modes are usually of importance in engine testing:

l Vertical oscillations on the X axis due to unbalanced vertical forcesl Rotation about the Y axis due to cyclic variations in torquel Rotation about the Z axis due to unbalanced vertical forces in different

transverse planes.

Torque variations will be considered later. In general the rotating masses arecarefully balanced but periodic forces due to the reciprocating masses cannotbe avoided. The crank, connecting rod, and piston assembly shown inFigure 9.2 is subject to a periodic force in the line of action of the piston, givenapproximately by:

f ¼ mpu2cr cos qþ

mpu2cr cos 2q

n(9.1)

Z

Z

X

X

Y

Y

FIGURE 9.1 Internal combustion engine:

principal axes and degrees of freedom.

m

f

I

r

ωc

θ

FIGURE 9.2 Connecting rod crank mechanism: unbalanced forces.

205Chapter | 9 Vibration and Noise

where n ¼ l/r in Figure 9.2. Here mp represents the sum of the mass of thepiston plus, by convention, one-third of the mass of the connecting rod (theremaining two-thirds is usually regarded as being concentrated at the crankpincenter).

The first term of equation (9.1) represents the first-order inertia force. It isequivalent to the component of centrifugal force on the line of action generatedby a mass mp concentrated at the crankpin and rotating at engine speed. Thesecond term arises from the obliquity of the connecting rod and is equivalent tothe component of force in the line of action generated by a mass m/4n at thecrankpin radius but rotating at twice engine speed.

Inertia forces of higher order (3�, 4�, etc., crankshaft speed) are alsogenerated but may usually be ignored.

It is possible to balance any desired proportion of the first-order inertia forceby balance weights on the crankshaft, but these then lead to an equivalentreciprocating force on the Z axis, which may be even more objectionable.

Inertia forces may be represented by vectors rotating at crankshaft speedand twice crankshaft speed. Table 9.1 shows the first- and second-order vectorsfor engines having from one to six cylinders. Note the following features:

l In a single-cylinder engine, both first- and second-order forces areunbalanced.

l For larger numbers of cylinders first-order forces are balanced.l For two- and four-cylinder engines, the second-order forces are unbalanced

and additive.

This last feature is an undesirable characteristic of a four-cylinder engine and insome cases has been eliminated by counter-rotating weights driven at twicecrankshaft speed.

TABLE 9.1 First- and Second-Order Forces in Engines of One to Six

Cylinders

First-orderforces

Second-orderforces

First-ordercouples

Second-ordercouples

206 Engine Testing

The other consequence of reciprocating unbalance is the generation ofrocking couples about the transverse or Z axis and these are also shown inFigure 9.1:

l There are no couples in a single-cylinder engine.l In a two-cylinder engine there is a first-order couple.l In a three-cylinder engine there are first- and second-order couples.l Four- and six-cylinder engines are fully balanced.l In a five-cylinder engine there is a small first-order and a larger second-

order couple.

Six-cylinder engines, which are well known for smooth running, are balancedin all modes.

Variations in engine turning moment are discussed in Chapter 11. Thesevariations lead to equal and opposite reactions on the engine, which tend to


cause rotation of the whole engine about the crankshaft axis. The order ofthese disturbances, i.e. the ratio of the frequency of the disturbance to theengine speed, is a function of the engine cycle and the number of cylinders.For a four-stroke engine the lowest order is equal to half the number ofcylinders: in a single cylinder there is a disturbing couple at half engine speedwhile in a six-cylinder engine the lowest disturbing frequency is at three timesengine speed. In a two-stroke engine the lowest order is equal to the numberof cylinders.

Design of Engine Mountings and Test-Bed Foundations

The main problem in engine mounting design is that of ensuring that themotions of the engine and the forces transmitted to the surroundings, as a resultof the unavoidable forces and couples briefly described above, are kept tomanageable levels. In the case of vehicle engines it is sometimes the practice tomake use of the same flexible mounts and the same location points as in thevehicle; this does not, however, guarantee a satisfactory solution. In the vehiclethe mountings are carried on a comparatively light structure, while in the testcell they may be attached to a comparatively massive pallet or even to a seismicblock. Also, in the test cell the engine may be fitted with additional equipmentand various service connections. All of these factors alter the dynamics of thesystem when compared with the situation of the engine in service and can causefatigue failures of both the engine support brackets and those of auxiliarydevices such as the alternator.

Truck diesel engines usually present less of a problem than small auto-motive engines, as they generally have fairly massive and well-spaced supportsat the flywheel end. Stationary engines will in most cases be carried on four ormore flexible mountings in a single plane at or near the horizontal centerline ofthe crank shaft, as in Figure 9.3.

We shall consider the simplest case, an engine of mass m kg carried onundamped mountings of combined stiffness kN/m (Figure 9.3). The differential

C of G

FIGURE 9.3 Engine carried on four flexible

mountings.

208 Engine Testing

equation defining the motion of the mass equates the force exerted by themounting springs with the acceleration of the mass:

md2x

dt2þ kx ¼ 0 (9.2)

A solution is:

x ¼ constant � sin

ffiffiffiffik

m

r� t

k

m¼ u2

0 natural frequency ¼ h0 ¼ u0

2p¼ 1

2p

ffiffiffiffik

m

r (9.3)

The static deflection under the force of gravity ¼ mg/k, which leads to a veryconvenient expression for the natural frequency of vibration:

h0 ¼ 1

2p

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffig

static deflection

r(9.4a)

or, if static deflection is in millimeters:

h0 ¼ 15:76ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffistatic deflection

p (9.4b)

This relationship is plotted in Figure 9.4.

40

30

20

10

54

3

2

10.2

Nat

ural

freq

uenc

y (H

z)

Static deflection (mm)

0.5 1 2 5 10 20 50 100 300

FIGURE 9.4 Relationship between static deflection and natural frequency.


Next consider the case where the mass m is subjected to an exciting force ofamplitude f and frequency u/2p. The equation of motion now reads:

md2x

dt2þ kx ¼ f sin ut

The solution includes a transient element. For the steady-state condition,amplitude of oscillation is given by:

x ¼ f=k

ð1� u2=u20Þ

(9.5)

Here f/k is the static deflection of the mountings under an applied load f. Thisexpression is plotted in Figure 9.5 in terms of the amplitude ratio, x divided bystatic deflection. It has the well-known feature that the amplitude becomestheoretically infinite at resonance, u ¼ u0.

If the mountings combine springs with an element of viscous damping, theequation of motion becomes:

md2x

dt2þ c

dx

dtþ kx ¼ f sin ut

6

4

2

00

Am

plitu

de r

atio

21

Frequency ratio0

FIGURE 9.5 Relationship between frequency and amplitude ratio (transmissibility), undamped

vibration.

210 Engine Testing

where c is a damping coefficient. The steady-state solution is:

x ¼ f=kffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi�1� u2

u20

�2

þ u2c2

mku20

s sinðut � AÞ (9.6a)

If we define a dimensionless damping ratio:

C2 ¼ c2

4mk

the above equation may be written as:

x ¼ f=kffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi�1� u2

u20

�2

þ 4C2 u2

u20

s sinðut � AÞ (9.6b)

(if C ¼ 1 we have the condition of critical damping when, if the mass is dis-placed and released, it will return eventually to its original position withoutovershoot).

The amplitude of the oscillation is given by the first part of this expression:

Amplitude ¼ f=kffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi�1� u2

u20

�2

þ 4C2 u2

u20

s

This is plotted in Figure 9.6, together with the curve for the undampedcondition, Figure 9.5, and various values of C are shown. The phase angle A isa measure of the angle by which the motion of the mass lags or leads theexciting force. It is given by the expression:

A ¼ tan�1 2Cu0o

u� u

u0

(9.7)

At very low frequencies A is zero and the mass moves in phase with the excitingforce. With increasing frequency, the motion of the mass lags by an increasingangle, reaching 90� at resonance. At high frequencies the sign of A changes andthe mass leads the exciting force by an increasing angle, approaching 180� athigh ratios of u to u0.

Natural rubber flexible mountings have an element of internal (hysteresis)damping, which corresponds approximately to a degree of viscous dampingcorresponding to C ¼ 0.05.

The essential role of damping will be clear from Figure 9.6: it limits thepotentially damaging amplitude of vibration at resonance. The ordinate inFigure 9.6 is often described as the transmissibility of the mounting system: it is

100

0.05

0.1

0.1

0

0.2

0.2

0.5

0.5

0.05

5

2

10.8

0.6

0.4

0.3

0.2

0.10.08

0.06

0.040.03

0.02

0

Tran

smis

sibi

lity

0.1 0.2 0.5 1 2 3 4 5 7 10

Frequency ratio0

C

FIGURE 9.6 Relationship between transmissibility (amplitude ratio) and frequency, damped

oscillations for different values of damping ratio C (logarithmic plot).


a measure of the extent to which the disturbing force f is reduced by theaction of the flexible mounts. It is considered good practice to design thesystem so that the minimum speed at which the machine is to run is not lessthan three times the natural frequency, corresponding to a transmissibility ofabout 0.15. It should be noticed that once the frequency ratio exceeds about 2,the presence of damping actually has an adverse effect on the isolation ofdisturbing forces.

Practical Considerations in the Design of Engineand Test-Bed Mountings

In the above simple treatment we have only considered oscillations in thevertical direction. In practice, as has already been pointed out, an engine carried

212 Engine Testing

on flexible mountings has six degrees of freedom (Figure 9.1). While in manycases a simple analysis of vibrations in the vertical direction will give a satis-factory result, under test cell conditions a complete computer analysis of thevarious modes of vibration and the coupling between them may be advisable.This is particularly the case with tall engines with mounting points at a lowlevel, when cyclic variations in torque may induce transverse rolling of theengine.

Reference [1] lists the design factors to be considered in planning a systemfor the isolation and control of vibration and transmitted noise:

l Specification of force isolation:

l as attenuation (dB)l as transmissibilityl as efficiencyl as noise level in adjacent rooms.

l Natural frequency range to achieve the level of isolation required.l Load distribution of the machine:

l is it equal on each mounting?l is the center of gravity low enough for stability?l exposure to forces arising from connecting services, exhaust system, etc.

l Vibration amplitudesdlow frequency:

l normal operationl fault conditionsl starting and stoppingl is a seismic block or sub-base needed?

l Higher-frequency structure-borne noise (100 Hzþ):

l is there a specification?l details of building structurel sufficient data on engine and associated plant.

l Transient forces:l shocks, earthquakes, machine failures.

l Environment:

l temperaturel humidityl fuel and oil spills.

Detailed design of engine mountings for test-bed installation is a highlyspecialized matter. In general the aim is to avoid “coupled” vibrations, e.g. thegeneration of pitching forces due to unbalanced forces in the vertical direction,or the generation of rolling moments due to the torque reaction forces exertedby the engine. These can create resonances at much higher frequencies than thesimple frequency of vertical oscillation calculated in the following section andlead to consequent trouble, particularly with the engine-to-dynamometer con-necting shaft.


Massive Foundations and Air-Sprung Bedplates

The analysis and prediction of the extent of transmitted vibration to thesurroundings is a highly specialized field. The starting point is the observationthat a heavy block embedded in the earth has a natural frequency of vibrationthat generally lies within the range 1000–2000 c.p.m. There is thus a possibilityof vibration being transmitted to the surroundings if exciting forces, generallyassociated with the reciprocating masses in the engine, lie within this frequencyrange. An example would be a four-cylinder four-stroke engine running at750 rev/min: we see from Table 9.1 that such an engine generates substantialsecond-order forces at twice engine speed, or 1500 c.p.m. Figure 9.7 gives anindication of acceptable levels of transmitted vibration from the point of viewof physical comfort.

Figure 9.8 is a sketch of a typical small seismic block. Reinforced concreteweighs roughly 2500 kg/m3 and this block would weigh about 4500 kg. Notethat the surrounding tread plates must be isolated from the block, also that it isessential to electrically earth (ground) the mounting rails. The block is showncarried on four combined steel spring and rubber isolators, each havinga stiffness of 100 kg/mm (Figure 9.9). From equation (9.4a) the naturalfrequency of vertical oscillation of the bare block would be 4.70 Hz, or

7

2

1

0.7

0.2

0.1

0.07

0.02

0.01

0.007

0.002

Rough

Normal

Smooth

Perceptible

Imperceptible

Frequency c.p.m.

Am

plitu

de m

m

100 200 500 1000 2000 5000

FIGURE 9.7 Perception of vibration. (Redrawn from Ref. [2].)

FIGURE 9.8 Spring-mounted seismic block with embedded box-section rails in top face and

service pipes running down each side of the pit. Pit must have a drainage point or pump sump.

214 Engine Testing

282 c.p.m., so the block would be a suitable base for an engine running at about900 rev/min or faster. If the engine weight were, say, 500 kg the naturalfrequency of blockþ engine would be reduced to 4.46 Hz, a negligible change.An ideal design target for the natural frequency is considered to be 3 Hz.

Resilient pads

Resilient pad

Assembly bolt and snubberSlot to inspectsnubber position

Base casting

Top casting

Rubber spring

Helical steel spring

FIGURE 9.9 Combined spring and rubber flexible mount.


Heavy concrete foundations (seismic blocks) carried on a flexiblemembrane are expensive to construct, calling for deep excavations, complexshuttering, and elaborate arrangements, such as tee-slotted bases, for boltingdown the engines. With the wide range of different types of flexible mountingnow available, it is questionable whether, except in special circumstances, suchas a requirement to install test facilities in proximity to offices, their use iseconomically justified. The trough surrounding the block may be of incidentaluse for installing services; if the gap is small, then there should be means ofdraining out contaminated fluid spills.

It is now common practice for automotive engines to be rigged on vehicle-type engine mounts, then on trolley systems, which are themselves mounted onisolation feet; therefore, less engine vibration is transmitted to the buildingfloor. In these cases a more modern alternative to the deep seismic block isshown in Figure 9.10a and is sometimes used where the soil conditions aresuitable. Here the test bed sits on a thickened and isolated section of the floorcast in situ on the compacted native ground. The gap between the floor andblock is almost filled with expanded polystyrene boards and sealed at floor levelwith a flexible, fuel-resistant, sealant; a damp-proof membrane should beinserted under both floor and pit or block.

Where the subsoil is not suitable for the arrangement shown in Figure 9.10a,then a pit is required, cast to support a concrete block that sits on a mat or pads

(a)

Consolidated subsoil

Cell floor

Seismic block

Isolation gap filled with fuel resistant sealant

Cell floorSeismic block

placed on resilient mat

Concrete pit

(b)

FIGURE 9.10 (a) Isolated foundation block for test stand set onto firm compacted subsoil.

(b) Seismic block onto resilient matting in a shallow concrete pit.

216 Engine Testing

of a material such as cork/nitrile rubber composite, which is resistant to fluidcontamination (Figure 9.10b). The latter design, while still used in someindustries, has the problem of fluids and foreign objects dropping intothe narrow annular space between block and pit unless provision is made toavoid it.


The modern replacement for the cast concrete seismic block is a large cast-iron bedplate supported by air springs, as shown in Figure 9.11. By using airsprings with automatic level control a constant, load-independent horizontalposition of the isolated base plate can be maintained; control accuracy ofbetween �0.1 and �0.01 mm can be achieved depending on the model ofdevice used.

Such precision of surface-level maintenance is completely unnecessary formost automotive engine test cells, but the self-leveling action is valuable whenUUTs of different weights are changed on the bedplate. These systems can betuned from “soft” where the single bedplate may be rocked by operatorsmoving on it, to almost rigid.

The first set-up of a multisection bed can take some time, particularly if thequality of the base floor is of low-quality finish and uneven.

Air-sprung bedplate installations have natural frequencies of around 3 Hz,providing good decoupling between the test bed and the building at lowfrequencies but at considerable expense.

Air-spring bedplate support systems made up of four or more units, such asthat shown in Figure 9.11, connected in pairs within a compressed air circuit are

FIGURE 9.11 A BILZ� air spring, which is part of a self-leveling bedplate support and damping

system. (1) Top plate of air suspension element. (2) Control valve. (3) Pressure safety valve.

(4) Level control. (5) Air hose. The units are usually connected in pairs and supplied with filtered

air. (6) Bar supply pressure. Bedplate control accuracies of down to �0.01 mm are claimed.


now used in large powertrain cells. When the air supply is switched off theblock or plate will settle down and rest on packers; this allows removal formaintenance (item 5 in Figure 9.12). The air springs require a reliable, low-flow, condensate-free air supply that is not always easy to provide to the lowestpoint in the cell system. Sufficient room must be left around the bedplate toallow maintenance access to the air-spring units and access to hatch plates forany centrally located.

In cases where there is no advantage in having the bedplate face at groundlevel (as is required for wheeled trolley systems), the bedplate can be mountedon the flat cell floor. It is possible that plant and engines mounted on rubberviscous mounts or air-spring systems could, unintentionally, become electri-cally isolated from the remainder of the facility because of the rubber elements;it is vital that a common grounding scheme is included in such facility designs(see Chapter 5).

A special application concerns the use of seismic blocks for supportingengines in anechoic cells. It is, in principle, good practice to mount enginesundergoing noise testing as rigidly as possible, since this reduces noiseradiated as the result of movement of the whole engine on its mountings.Lowering of the center of gravity is similarly helpful, since the engines have tobe mounted with a shaft center height of at least 1 meter to allow formicrophone placement.

Except in anechoic cells, it is common practice to mount both engine anddynamometer on a common block; if they are separated the relative movement

4

2

1

53

FIGURE 9.12 Seismic block (or cast-iron bedplate not shown) (4) mounted on air springs

(3) within a shallow concrete pit (2) on consolidated subsoil (1). When the air supply is off, the

self-leveling air springs allow the block to settle on support blocks (5), the rise and fall typically

being 4–6 mm. Maintenance access to the springs is not shown.

+180°

+90°

+180°

+90°

+270°

Displacement

Restoring force180° out of phase with displacement,urges towards center: does no work

VelocityLeads displacement by 90°

AccelerationIn phase with restoring force

Driving forceIn phase with velocity: does work

Damping forceLags displacement by 90°

(or leads it by 270°): absorbs energy

+

Vibrations of a pendulum or spring and mass

–

FIGURE 9.13 Diagram of phase relationship of vibrations.

218 Engine Testing


between the two must be within the capacity of the connecting shaft and itsguard.

Finally, it should be remarked that there is available on the marketa bewildering array of different designs of vibration isolator or flexiblemounting, based on steel springs, air springs, natural or synthetic rubberof widely differing properties used in compression or shear, and com-binations of these materials. For the nonspecialist the manufacturer’sadvice should be sought, with the specific test facility requirement clearlyspecified.

Summary of Vibration Section

This section should be read in conjunction with Chapter 10, which deals withthe associated problem of torsional vibrations of engine and dynamometer. Thetwo aspectsdtorsional vibration and other vibrations of the engine on itsmountingsdcannot be considered in complete isolation. A diagrammaticsummary of the vibrations of a sprung mass that may be found useful is given inFigure 9.13.

Vibration Notation

Mass of piston þ one-third connecting rod mp (kg)Angular velocity of crankshaft uc (s

�1)Crank radius r (m)Crank angle from TDC q

Connecting rod length/r nUnbalanced exciting force f (N)Mass of engine m (kg)Combined stiffness of mountings k (N/m)Amplitude of vibration x (m)Angular velocity of vibration u (rad/s)Angular velocity at resonance u0 (rad/s)Natural frequency n0 (Hz)Phase angle A (rad)Damping coefficient c (N s/m)
Damping ratio CAcceleration due to gravity g (m/s2)
PART 2: NOISE AND ITS ATTENUATION IN TEST FACILITIES

Note: This section deals with some of the fundamentals of noise, its generationand attenuation in powertrain test facilities.

The testing of automotivemodules for noise, vibration, and harshness (NVH)and the design of anechoic cells for this and EMC test facilities are dealt with inmore detail in Chapter 18. Exhaust silencers are covered in Chapter 7. Venti-lation silencers for noise “breakout” and fan noise are covered in Chapter 6.

220 Engine Testing

Sound Intensity

The starting point in the definition of the various quantitative aspects of noisemeasurement is the concept of sound intensity, defined as:

I ¼ p2

rcW=m2

where p2 is the mean square value of the acoustic pressure, i.e. the pressurevariation due to the sound wave, r the density of air, and c the velocity of soundin air.

Intensity is measured in a scale of decibels (dB):

dB ¼ 10 log10

�I

I0

�¼ 20 log10

�p

p0

�

where I0 corresponds to the average lower threshold of audibility, takenby convention as I0 ¼ 10�12 W/m2, an extremely low rate of energypropagation.

From these definitions it is easily shown that a doubling of the soundintensity corresponds to an increase of about 3 dB (log10 2 ¼ 0.301). A tenfoldincrease gives an increase of 10 dB, while an increase of 30 dB corresponds toa factor of 1000 in sound intensity. It will be apparent to the reader that intensityvaries through an enormous range. The value on the decibel scale is oftenreferred to as the sound pressure level (SPL).

In general, sound is propagated spherically from its source and the inversesquare law applies. Doubling the distance results in a reduction in SPL of about6 dB (log10 4 ¼ 0.602).

The human ear is sensitive to frequencies in the range from roughly 16 Hz to20 kHz, but the perceived level of a sound depends heavily on its frequencystructure. The well-known Fletcher–Munson curves (Figure 9.14) wereobtained by averaging the performance of a large number of subjects who wereasked to decide when the apparent loudness of a pure tone was the same as thatof a reference tone of frequency 1 kHz.

Loudness is measured in a scale of phons, which is only identical with thedecibel scale at the reference frequency. The decline in the sensitivity of the earis greatest at low frequencies. Thus, at 50 Hz an SPL of nearly 60 dB is neededto create a sensation of loudness of 30 phons.

Acoustic data are usually specified in frequency bands one octave wide. Thestandard mid-band frequencies are:

31.5 62.5 125 250 500 1000 2000 4000 8000 16,000 Hz

e.g. the second octave spans 44–88 Hz. The two outer octaves are rarely used innoise analysis.

Sou

nd p

ress

ure

leve

l (dB

) (2

0 µP

a)

Equal loudness contours for pure tonesby octave (Fletcher–Munson curves)

110

100

90

80

70

60

50

40

30

20

10

031.5 63 125 250 500 1000 2000 4000 8000 16,000

Octave band center frequency (Hz)

Thresholdof hearing

FIGURE 9.14 Fletcher–Munson curves of equal loudness.


Noise Measurements

Most instruments for measuring sound contain weighted networks thatgive a response to frequency which approximates to the Fletcher–Munsoncurves. In other words, their response to frequency is a reciprocal of theFletcher–Munson relationship (Figure 9.14). For most applications theA-weighting curve (Figure 9.15) gives satisfactory results and the corre-sponding SPL readings are given in dBA. B- and C-weightings are some-times used for high sound levels, while a special D-weighting is usedprimarily for aircraft noise measurements.

C

B

A

A

B and C

50

–5–10

–15

–20

–25–30–35

–40

–45

–50

Frequency (kHz)

Rel

ativ

e re

spon

se (

dB)

0.02 0.05 0.1 0.2 0.5 1 2 5 10 20

FIGURE 9.15 Noise weighting curves.

222 Engine Testing

The dBA value gives a general “feel” for the intensity and discomfort levelof a noise, but for analytical work the unweighted results should be used. Thesimplest type of sound-level meter for diagnostic work is the octave bandanalyzer. This instrument can provide flat or A-weighted indications of SPL foreach octave in the standard range. For more detailed study of noise emissions,an instrument capable of analysis in one-third octave bands is more effective.With such an instrument it may, for example, be possible to pinpoint a partic-ular pair of gears as a noise source.

For serious development work on engines, transmissions, or vehiclebodies, much more detailed analysis of noise emissions is provided by thediscrete Fourier transform (DFT) or fast Fourier transform (FFT) digitalspectrum analyzer. The mathematics on which the operation of these instru-ments is based is somewhat complex, but fortunately they may be usedeffectively without a detailed understanding of the theory involved. It is wellknown that any periodic function, such as the cyclic variation of torque in aninternal combustion engine, may be resolved into a fundamental frequencyand a series of harmonics. General noise from an engine or transmission doesnot repeat in this way and it is accordingly necessary to record a sample of thenoise over a finite interval of time and to process this data to give a spectrum ofSPL against frequency. The Fourier transform algorithm allows this to bedone.

Permitted Levels of Noise in the Workplace

Regulations concerning the maximum and average levels of noise to whichworkers are exposed exist in most countries of the world. In the UK (theControl of Noise at Work Regulations 2005) and Europe the levels of exposure


to noise (peak sound pressure) of workers, averaged over a working day orweek, at which an employer has to take remedial action are:

l Lower exposure action values:l daily or weekly exposure of 80 dBl peak sound pressure of 135 dB.

l Upper exposure action values:l daily or weekly exposure of 85 dBl peak sound pressure of 137 dB.

l There are also absolute levels of noise exposure that must not be exceededwithout the supply and use of personal protection equipment:l daily or weekly exposure of 87 dBl peak sound pressure of 140 dB.

Noise levels actually within an engine test cell nearly always exceed the levelspermitted by statutes, while the control room noise level must be kept underobservation and appropriate measures taken. The employer’s obligations underprevailing regulations include the undertaking of an assessment when anyemployee is likely to be exposed to the first action level. This means thathearing protection equipment must be available to all staff having to entera running cell, where sound levels will often exceed 85 dBA, and that controlroom noise levels should not exceed 80 dBA.

Noise-canceling earphones, in place of the noise-attenuating type, should beconsidered for use in high noise areas where vocal communication betweenstaff is required.

Noise External to Test Facility and Planning Regulations

Almost all planning authorities will impose some form of restriction on thebuilding of a new engine test laboratory that will seek to restrict the noisepollution caused. In the case of an existing industrial site, these restrictions willoften take the form of banning any increase in sound levels at the nominatedboundaries of the site. It is vital, therefore, that in the early planning and designstages a set of boundary noise readings are taken by both the system integratorand the user. The survey should be based on readings taken at important points,such as the boundary position nearest to residential buildings, then marked ona site map with GPS readings. The readings should be taken in as normala working situation as possible, without exceptional occurrences, and at severaltimes throughout the period the facility is to be used. Such a datum set of readingsshould be taken before work starts and agreed as relevant with any interestedparty; theywill then provide a key reference for any dispute post-commissioning.

Noise Reverberation in the Test Cell Environment

The measured value of SPL in an environment such as an engine test cell givesno information as to the power of the source: a noisy machine in a cell having

TABLE 9.2 A Sample of Absorption Coefficients for Materials Found in

Test Cell Construction, at Three Frequencies

Wall Material 125 Hz 500 Hz 4 kHz

Poured concrete, unpainted 0.01 0.04 0.1

Brick, painted 0.01 0.02 0.03

Glass (large cell window) 0.18 0.04 0.02

Perforated metal (13% open)over 50 mm fiberglass

0.25 0.99 0.92

224 Engine Testing

good sound-absorbent surfaces may generate the same SPL as a much quietermachine surrounded by sound-reflective walls. The absorption coefficient isa measure of the sound power absorbed when a sound impinges once upona surface. It is quite strongly dependent on frequency and tends to fall asfrequency falls below about 500 Hz.

Information on absorption coefficients for a wide range of structuralmaterials and sound insulators is given in IHVE Guide B 12. A few approxi-mate values are given in Table 9.2 and indicate the highly reverberatoryproperties of untreated brick and concrete.

It should be remembered that the degree to which sound is absorbed by itssurroundings makes no difference to the intensity of the sound received directlyfrom the engine.

“Crosstalk” between test cell and control room and other adjacent roomscan occur through any openings in the partition walls and through air-conditioning ducts and other service pipes, when there is a common system.

Noise Notation

R.m.s. value of acoustic pressure p (N/m2)Density of air r (kg/m3)Velocity of sound in air c (m/s)

2
Sound intensity I (W/m )Threshold sound intensity I0 (W/m2)
REFERENCES

[1] A.N. Maw, The design of resilient mounting systems to control machinery noise in buildings,

Plastics, Rubber and Composites Processing and Applications 18 (1992) 9–16. ISSN: 0959-

8111.

[2] W. Ker-Wilson, Vibration Engineering, Griffin, London, 1959. ISBN-10: 0852640234 (out of

print in 2011).


USEFUL TEXTS

[1] BS 3425. Method for the Measurement of Noise Emitted by Motor Vehicles.

[2] BS 3539. Specification for Sound Level Meters for the Measurement of Noise Emitted by

Motor Vehicles.

[3] BS 4196 Parts 0 to 8. Sound Power Levels of Noise Sources.

[4] Control of Noise at Work Regulations, Health and Safety Executive, London, 2005.

[5] B. Fader, Industrial Noise Control: Fundamentals and Applications, Wiley, Chichester, 1981.

ISBN-13: 978–0824790288.

[6] I.H.V.E. Guide B12: Sound Control. Chartered Institution of Building Services, London.

[7] W.T. Thompson, Theory of Vibration, Prentice-Hall, London, 1988. ISBN-13:

978-0412783906.

[8] J.D. Turner, A.J. Pretlove, Acoustics for Engineers, Macmillan Education, London, 1991.

ISBN-13: 978-0333521434.

[9] R.H. Warring, Handbook of Noise and Vibration Control, fifth ed., Trade and Technical Press,

Morden, UK, 1983. ISBN-13: 978-0471395997.

FURTHER READING

BS 2475. Specification for Octave and One-third Octave Band-Pass Filters.

BS 3045. Method of Expression of Physical and Subjective Magnitudes of Sound or Noise in Air.

BS 4198. Method for Calculating Loudness.

BS 4675 Parts 1 and 2. Mechanical Vibration in Rotating Machinery.

BS 5330. Method of Testing for Estimating the Risk of Hearing Handicap due to Noise Exposure.

W.T. Thomson, Theory of Vibration with Applications, fourth ed., Taylor & Francis, 1996.

ISBN-13: 978-0412546204.

Documents

Engine Testing || Vibration and Noise