Xin Q. Diesel Engine System Design

Подождите немного. Документ загружается.

790 Diesel engine system design

period of the impact process. Deformation modeling plays an important role

in piston slap analysis and may enhance the understanding of the real elastic

impact process through modeling the details of the impact force. First, the

deformation of the components affects the clearance between the skirt and

the liner, thus affecting the location and the severity of the piston slap via

the piston impact velocity and the oil lm damping force. Secondly, the

elastic impact between the piston skirt and the liner is a highly complex

transient collision process. During the impact, a part of the impact kinetic

energy is transformed into elastic deformation, and the remaining kinetic

energy becomes lower for the second impact. As an alternative approach,

elastic impact modeling may be more accurate than the right-body impact

modeling in simulating the time history of the impact force. The impact force

is basically equal to the side thrust plus the impulsive slap force. In the rigid-

body impact model, the coefcient of restitution is used to assume the impact

and the rebound processes occur within a very short time interval, and a large

average impulsive impact force during the process is obtained by using the

ratio of the impact impulse to the assumed short time duration. Although the

impulsive force is much larger than the side thrust, the sudden application

of the side thrust acting on the impact surface after the piston travels across

the clearance is also important for the noise, as Lalor et al. (1980) pointed

out. In this approach, the piston impact velocity is the only key parameter

because the impact duration can be uncertain. In an elastic impact model,

the impact force is a more important parameter, which can be equal to the

side thrust superpositioned by a large oscillating/vibratory impulsive force

of the piston–bore impact (i.e., the side thrust is amplied by the vibration

force). The oscillating force consists of the elements such as the stiffness,

the damping, and the piston impact velocity (all in the lateral direction that

is perpendicular to the liner wall) of the piston–lm–liner elastic system.

The oscillating force governs the piston impact and the rebound behavior,

and may provide more details during the impact process than the rigid-body

assumption. Compared with the rigid-body impact model, one advantage of

the elastic impact model is that the impact forcing function can be more

accurately transformed from the time domain to the frequency domain in

order to (1) directly analyze the frequency characteristics of the piston slap

noise, and (2) correlate its harmonic contents with other frequency signals

commonly used in the NVH area (e.g., the liner acceleration). The piston side

thrust has low frequencies (e.g., less than 500 Hz), while the addition of an

impulsive piston slap force can result in dominant force amplitudes at high

frequencies (e.g., 1–3 kHz). The piston impact force and the liner vibration

response can be used to estimate the piston slap noise and the cylinder liner

cavitation.

In the elastic impact modeling area, Lalor et al. (1980) presented a

simplied piston slap model using a mass–spring system to represent the

Diesel-Xin-11.indd 790 5/5/11 12:01:42 PM



791Noise, vibration, and harshness (NVH) in engine system design

piston, the liner, and the engine block. Paranjpe (1998) used a similar simple

spring–damper system to model the elastic impact process and the piston–

liner contact forcing function in the frequency domain. Ohta et al. (1987)

provided an important modeling approach to simulate the transient impact

force during the piston slap process and the dynamic vibration response of

the coupled piston–liner system by using the stiffness and dashpot elements.

They calculated the time history and the frequency spectrum of the impact

forcing function. Their calculated time history and frequency spectrum of the

liner acceleration substantially agreed with measured results. The work by

Ohta et al. (1987) is one of the most important achievements in the piston

slap area.

In the component deformation modeling, early research on the piston

slap simulation with the skirt/bore deformation was conducted by Sander

et al. (1979) without a lubrication model. Kageyama et al. (1994) used a

simpli ed contact deformation (stiffness) model for a barrel skirt without

lubrication. Patel et al. (2007) used a complementarity method to model the

dry-contact piston slap for a non-cylindrical piston with the effects of bore

distortion and skirt deformation to calculate the kinetic energy loss due to

piston slap and to estimate the friction force.

Directly predicting the cylinder liner vibration accurately can be important

for a more direct prediction of the piston slap noise. Ohta et al. (1987)

provided a practical approach to the dynamic modeling of the piston–liner

vibration system by using a set of ordinary differential equations, which is

suitable for the engine system-level design analysis. Kobayashi et al. (2007)

presented a highly complicated method to use  nite element models of the

cylinder block and the piston to simulate the piston slap, the deformations

and the liner vibration. Their computation is time-consuming. Although

acceptable for the component-level design, this approach is not suitable for

an early concept-level engine system design.

In summary, from the perspective of diesel engine system design, a viable

approach for a system engineer to analyze and quantify the piston slap noise

in the NVH area is to adopt a computational approach based on the ordinary

differential equations of the multi-body multi-phase dynamics of the piston

assembly and the cylinder liner. With certain  delity of the impact forcing

details and the recurring impact and rebound events, the time history and the

frequency spectrum of the piston impact forcing function and the cylinder

liner vibration response can be predicted. This requires instantaneous crank-

angle-resolution modeling. The calculated piston impact kinetic energy

loss summed over one engine cycle is the indicator of the sound pressure

level.

The engine noise level (dB) due to piston slap can be estimated by

pW = pW

– 10

l4)l



11.11

Diesel-Xin-11.indd 791 5/5/11 12:01:42 PM



792 Diesel engine system design

where p

SPL

is the sound pressure level,



SWLSW

is the sound power level, l is

the distance from the engine center to the microphone. The sound power of

piston slap is estimated by



WZ = WZ

()

slapsl

()

slapsl

()

11.12

where Z

is the mechanical impedance, Z

is the radiation impedance,

slapsl

is the total kinetic energy loss of all the piston impact events

within one engine cycle, and C is a model calibration constant to offset

the errors in the kinetic energy loss. Z

and Z

are measured or simulated

known inputs.

11.5.3 The effects of piston design and engine operation

on piston slap

Within an entire engine cycle, piston slap usually becomes the most severe

and noisiest during the expansion stroke due to the high in-cylinder  ring

pressure in the expansion stroke (0–180°). Offsetting the piston pin to the

thrust side usually reduces the severity of the piston slap at the more rigid

upper portion of the piston skirt during the expansion stroke. But the piston

slap at the less stiff lower portion of the skirt in other strokes might increase.

Figure 11.2 shows the simulated piston slap and rebound motions of a cold

piston with simpli ed skirt pro le without effective skirt lubrication during

one engine cycle (at 2300 rpm engine speed). Figure 11.3 illustrates the

instantaneous kinetic energy calculated for different piston designs under

lubricated conditions. These simulations provide insights into the complex

physical processes that are dif cult to measure.

11.6 Valvetrain noise

Usually the contribution of the valvetrain noise to the engine sound pressure

level is small. However, the valvetrain noise can be a sound quality issue due to

its high-frequency nature (e.g., above 3 kHz). Valvetrain NVH is an important

topic to consider in overall engine NVH planning and valvetrain design.

The excitation mechanisms of the valvetrain noise include three

sources:

∑ Cam acceleration excitation: The opening and closing cam positive

acceleration peaks (or more fundamentally, the cam jerks – the derivatives

of cam accelerations) excite the valvetrain vibration due to the inertia

force at high engine speeds with high frequencies. Any discontinuity or

non-smoothness in the cam acceleration will produce a ‘hammering’ shock

loading effect to excite the valvetrain to vibrate violently and produce

noise.

Diesel-Xin-11.indd 792 5/5/11 12:01:43 PM



Piston lubrication dynamics simulation with hot piston skirt proﬁle B

Piston lubrication dynamics simulation with hot piston skirt proﬁle D

Piston lubrication dynamics simulation with hot piston skirt proﬁle A

Piston lubrication dynamics simulation with hot piston skirt proﬁle C

Piston 1, 2300 rpm 300 hp, piston skirt friction power 3.057 hp

Piston 2, 2300 rpm 300 hp, piston skirt friction power 2.172 hp

Piston 5, 2300 rpm 300 hp, piston skirt friction power 5.640 hp

Piston 6, 2300 rpm 300 hp, piston skirt friction power 3.870 hp

Piston 3, 2300 rpm 300 hp, piston skirt friction power 3.814 hp

Piston 4, 2300 rpm 300 hp, piston skirt friction power 2.720 hp

Piston 7, 2300 rpm 300 hp, piston skirt friction power 3.460 hp

Piston 8, 2300 rpm 300 hp, piston skirt friction power 2.518 hp

Piston skirt hot barrel-shape proﬁle A: axially symmetric

from skirt top to bottom, very arced

Piston 1: Minimum radial clearance 20 microns, skirt length

88.5 mm, piston pin offset 0.7 mm toward anti-thrust side

Piston 2: Minimum radial clearance 20 microns, skirt length

68.5 mm, no piston pin offset

Piston skirt hot barrel-shape proﬁle C: axially cylindrical

shape

Piston 5: Minimum radial clearance 20 microns, skirt length

88.5 mm, piston pin offset 0.7 mm toward anti-thrust side

Piston 6: Minimum radial clearance 20 microns, skirt length

68.5 mm, no piston pin offset

Piston skirt hot barrel-shape proﬁle B: axially symmetric

but less arced than proﬁle A

Piston 3: Minimum radial clearance 20 microns, skirt length

88.5 mm, piston pin offset 0.7 mm toward anti-thrust side

Piston 4: Minimum radial clearance 20 microns, skirt length

68.5 mm, no piston pin offset

Total kinetic energy of piston secondary

motions, lateral plus tilting (mJ)

Total kinetic energy of piston secondary

motions, lateral plus tilting (mJ)

Total kinetic energy of piston secondary

motions, lateral plus tilting (mJ)

Total kinetic energy of piston secondary

motions, lateral plus tilting (mJ)

0.3

0.25

0.2

0.15

0.1

0.05

0.003

0.0025

0.002

0.0015

0.001

0.0005

0.3

0.25

0.2

0.15

0.1

0.05

0.3

0.25

0.2

0.15

0.1

0.05

0 90 180 270 360 450 540 630 720

Crank angle (degree)

0 90 180 270 360 450 540 630 720

Crank angle (degree)

0 90 180 270 360 450 540 630 720

Crank angle (degree)

0 90 180 270 360 450 540 630 720

Crank angle (degree)

Piston skirt hot barrel-shape proﬁle

D: axially asymmetric, more

recess at skirt bottom, very arced

Piston 7: Minimum radial clearance

20 microns, skirt length 88.5 mm,

piston pin offset 0.7 mm toward

anti-thrust side

Piston 8: Minimum radial clearance

20 microns, skirt length 68.5 mm,

no piston pin offset

11.3 Piston lubrication dynamics simulation on kinetic energy of different piston designs.

Diesel-Xin-11.indd 793 5/5/11 12:01:43 PM



794 Diesel engine system design

∑ Valvetrain impact excitation: These include the impulsive impact loading

at the valve opening between the cam and the follower (for mechanical

lash adjusters), at the valve closing between the valve and the valve

seat, and during the valve event when valvetrain separation and jump/

bounce occurs at high speeds. The most common valvetrain impact is

valve seating. Often the primary source of the noise in the pushrod

valvetrain is the valve seating impact.

∑ Frictional vibration: This occurs at the severe boundary lubrication

conditions where the asperity metal contact occurs between the cam and

the follower, typically at the two locations near the cam nose where the

theoretical lubricant entraining velocity becomes zero. This noise can

be prominent at low speeds.

The experimental work of valvetrain noise identication by Hanaoka and

Fukumura (1973), Dent and Chen (1989), Kalser et al. (1991), and Suh and

Lyon (1999) showed evidence of these valvetrain noises. These noises can

be reected in a measured valve acceleration–cam angle diagram (Hanaoka

and Fukumura, 1973) or a time–frequency diagram (Suh and Lyon, 1999).

There are usually three groups of vibration (highly oscillatory) signals of the

valve acceleration located at the cam opening ank, the cam closing ank,

and the valve closing. They are high frequency signals, especially at valve

seating. Cylinder head vibrates at the similar crank angle timing as the valve

acceleration excitation peaks at the valve opening and closing sides. The

valvetrain noise generally increases with engine speed.

The valvetrain noise is transmitted via both airborne and structure-borne

paths. Anderton and Zheng (1993) found that the airborne valvetrain noise

contributed greatly to the total valvetrain noise, particularly at high engine

speeds above 2000–3300 rpm. The airborne noise inside the valve cover

might be dominated by the valve seating impact. Suh and Lyon (1999)

discovered that there were two transmission paths of the vibration through

the structure for the excitation due to the valve seating impact: (1) directly

transmitted to the cylinder head via the valve seat (being a more dominant

path); and (2) transmitted via the valvetrain components and the camshaft

to the cylinder head. The vibration was transmitted from the cam to the

cylinder head surface for the excitation due to cam acceleration.

Suh and Lyon (1999) found that the contributions to the cylinder head

vibration response from the excitation sources of cam acceleration and

valve seating impact were comparable in magnitude. They concluded that

both sources could be dominant to the noise radiation from the structure

surface in the frequency range up to 6 kHz. They found that valve seating

excited the vibration in a very wide band frequency spectrum and it was the

solely dominant source for the frequency range above 10–20 kHz. Savage

and Matterazzo (1993) conducted an experiment on a 3.3 L gasoline engine

Diesel-Xin-11.indd 794 5/5/11 12:01:43 PM



795Noise, vibration, and harshness (NVH) in engine system design

to investigate the possible inuential factors for the valvetrain noise, such

as cam jerk level, valve spring load, tappet-to-bore clearance, valve stem

clearance and nish, rocker arm bearing clearance, valve overlap, and cylinder

head mass and damping. They discovered that lowering the cam jerk level

had a considerably larger effect on reducing the valvetrain noise (sound

pressure level) than eliminating the valve seating. When the valve seating

was eliminated by using a shimmed tappet (i.e., no seating), the valvetrain

noise did not reduce until the engine speed was higher than 3000 rpm. They

concluded that cam acceleration or jerk level was the most important factor

affecting the valvetrain noise over the entire engine speed range, and valve

seating velocity control only played a signicant role in noise reduction at

high speeds (e.g., above 3500 rpm).

The valvetrain noise can be controlled by reducing the excitation source

or impeding the noise transmission path. The following design measures

provide a few examples:

∑ Use very smooth cam acceleration prole and reduce cam acceleration

level. Match the cam acceleration with the valvetrain stiffness. This

controls the valvetrain inertia excitation related to the valve opening

velocity, separation, and bounce. Smooth acceleration can often be

achieved by advanced cam design techniques and optimization. Reducing

cam acceleration or jerk level often results in smaller valve ow area and

deterioration in engine volumetric efciency. Overall valvetrain system

optimization can minimize the compromise between the valvetrain noise

and the engine fuel economy.

∑ Use high precision to machine the cam prole to make the cam acceleration

and jerk proles of the manufactured cam become smooth. Experimental

data showed that manufacturing precision is very important for valvetrain

noise (Hanaoka and Fukumura, 1973).

∑ Reduce valvetrain vibration by using higher valvetrain stiffness and

lower mass.

∑ Use hydraulic lash adjuster to eliminate valvetrain clearances and the

impact at valve opening.

∑ Reduce valve seating velocity by cam acceleration and vibration control.

An appropriate kinematic cam ramp velocity is also important. Lowering

the valve seating velocity by lowering the cam ramp velocity has been

the traditional design approach to reduce the valvetrain noise. In fact, the

valve seating velocity can be reduced more effectively by using smooth

cam acceleration, and this may reduce the valvetrain inertia excitation

as well.

∑ Use smaller tappet-to-bore clearance to reduce the impact.

∑ Reduce the excitation levels of the cam bearing force.

∑ Reduce or eliminate the valvetrain friction-induced vibration noise by

designing to achieve greater oil lm thickness.

Diesel-Xin-11.indd 795 5/5/11 12:01:43 PM



796 Diesel engine system design

∑ Control the noise transmission path. Use special high-frequency-isolation

materials to reduce the structure-borne noise. Increase the cylinder head

mass, the stiffness, and the material damping to reduce the surface

vibration level of the cylinder head.

∑ Reduce the noise radiation efciency. Valve cover can provide attenuation

of the airborne noise generated under the valve cover. The valve cover

needs to provide a low vibration response to structure-borne noise by

high damping and stiffness or isolation.

Valvetrain noise analysis can be conducted by a system engineer as follows

in a level-2 approach (i.e., with instantaneous crank-angle resolution). At a

given engine speed–load mode, the valvetrain vibration cam force and the

valve seating velocity are calculated with a valvetrain dynamics model. The

time-history of the force and the impulse data are converted to the frequency

spectrum of the excitation sources by using the Fast Fourier Transfer. The

spectrum of the vibration response of the cylinder head surface can be calculated

by multiplying the excitation spectrum with a known (measured) transfer

function characterizing the transmission path. The total response spectral

sum is the superposition of the two excitation mechanisms, the cam force,

and the valve seating. Finally the frequency spectrum of the radiated noise

can be calculated after applying the A-weighting to obtain the A-weighted

spectrum. A practical approach of the valvetrain noise calculation was

provided by Kalser et al. (1991).

11.7 Geartrain noise

The gear rattle noise is a primary NVH issue for automotive drivertains such

as in the transmissions. It is also important in the timing gears and the gears

used to transmit torque to drive the accessories such as the fuel pump in the

diesel engine. The engine transmits non-uniform torques from the cranktrain

to the drivetrain or the geartrain and that causes the gear rattle noise.

Clearances are inevitably required between the idle surfaces of the

engaging gear teeth in geared/toothed connections (e.g., automotive gear

boxes, manual transmissions, clutches, gear couplings) in order to allow for

thermal expansion and manufacturing tolerances. When the gears are lightly

loaded with strongly oscillating torque and/or rotate at low speeds, there is

a high possibility of meshing teeth separation and the resulting vibration

impact between the teeth surfaces. Such a vibratory energy excitation of the

gear teeth within the working clearance generates the annoying rattle noise.

The gear rattle noises in the powertrain and the drivetrain are related to

torsional vibrations. Another annoying gear noise is the whine noise which

is excited by the transmission error at the gear mesh due to manufacturing

errors and tooth deection under load. The gear rattle noise is affected

Diesel-Xin-11.indd 796 5/5/11 12:01:43 PM



797Noise, vibration, and harshness (NVH) in engine system design

by the transmission input speed uctuation, kinematic transmission error,

the backlash in the tooth meshes, and the friction force acting on the

gears.

The ever increasing cylinder pressure and fuel injection pressure

dramatically increase the torsional excitation of the geartrain in the diesel

engine. The gear impact noise may become a signicant noise source at

many operating conditions especially at full load. The geartrain noise level

is a strong function of the size of the geartrain, the number of gear meshes,

the location of the geartrain (front or rear), and the magnitude of torsional

inputs from the crankshaft and the fuel system (Zhao and Reinhart, 1999).

Optimizing gear prole and minimizing the transmission error by reducing

manufacturing and assembly errors may also reduce the noise.

Engine geartrain noise was investigated by Spessert and Ponsa (1990)

and Zhao and Reinhart (1999). Geartrain dynamics and gear motion/rattle

modeling are highly complex. Good references are provided by Croker et

al. (1995), Padmanabhan et al. (1995), Meisner and Campbell (1995), Lahey

et al. (2001), and Wang et al. (2001).

11.8 Cranktrain and engine block noises

The crankshaft torsional vibration, the thin sections in the engine block and

the covers (e.g., the valve cover, the oil pan) bolted to the crankcase and

the cylinder block are important sources of NVH. Sophisticated commercial

software is available (e.g., Ricardo’s ENGDYN) to predict the time-domain

response of the coupled cranktrain and engine block system with nonlinear

oil lm lubrication models (e.g., Offner et al., 2004). The radiated noise can

be calculated from the surface normal velocity using the Rayleigh equation.

The radiated sound power and the noise radiation efciency can be calculated

for each vibrating surface that is approximated as a at plate. The sound

intensity can also be calculated across the surface. Critical modes of vibration

of the powertrain can be identied. Cranktrain dynamics and its interaction

with the cylinder block can be reliably studied in simulation.

Details of crankshaft, crankcase, and engine block vibration and noise

are provided by Russell (1972), Ochiai and Yokota (1982), and Maetani et

al. (1993). Engine mounts were reviewed by Shangguan (2009). Cranktrain

dynamics and engine balance have been extensively covered in the literature

(e.g., Thomson, 1978; Lee et al., 2000).

11.9 Auxiliary noise

The noise contribution from the auxiliary devices is signicant in modern

low-noise diesel engines. The oil pump (tooth pump) noise is excited

mainly by oil pressure peaks. The generator noise comes mainly from the

Diesel-Xin-11.indd 797 5/5/11 12:01:44 PM



798 Diesel engine system design

aerodynamic noise of the fan and the rotor, and the mechanical noise of its

bearings. Fuel system torsional vibration due to high cyclic torque affects

the geartrain noise. Some fuel systems may have very low cyclic torque, thus

contributing little to the gear noise. Haller et al. (1993) reviewed the noise

excitation by auxiliary devices in heavy-duty diesel engines including oil

pumps, generators, reciprocating air compressors, and hydraulic oil pumps.

The hydraulic  uid borne noise was elaborated by Skaistis (1975).

11.10 Aerodynamic noises

11.10.1 Intake noise

The excitation source of the engine intake air noise is the pressure  uctuation

produced by the reciprocating piston motion and the valve timing. The intake

noise is produced by the air pressure pulsation in the inlet duct and the high

air  ow velocity as the turbulent air  ow passes through the intake valve

opening area. The noise due to the mechanical vibration of the intake pipes

can also be associated as a part of the intake noise, or called secondary

airborne noise.

The intake noise is usually dominated by low-frequency contents although

the intake noise associated with the varying valve  ow area contains a wide

range of frequencies. The strongest low-frequency content in the intake noise

is directly related to engine speed and given by the following:

ntak

qi,qi

fi =fi

11.13

where i is the harmonic number, N

is the engine crankshaft speed in rpm,

is the number of cylinders, n

= 2 for four-stroke engines and n

= 1 for

two-stroke engines.

The intake noise increases as the engine speed increases mainly because

the inlet air  ow rate, the air pressure pulsation amplitude and frequency

increase with engine speed. Intake system design has a major in uence on

the intake noise through gas pressure wave dynamics. Engine displacement,

intake valve size, valve opening timing, intake runner and manifold all affect

the intake noise. Moreover, the shape of the air inlet opening (e.g., circular,

or  attened in the aspect ratio) has an effect on the intake air noise, and the

noise may decrease as the circular cross-sectional area of the opening is

decreased or throttled. Excessively  attened inlet design may produce extra

whistling noise.

There are generally three methods to control the intake noise: (1) using

silencers; (2) reducing the pulsation magnitude of the intake air pressure;

and (3) reducing the turbulence level of the intake valve air  ow. Silencers

are the most effective design measures to reduce the intake and exhaust

Diesel-Xin-11.indd 798 5/5/11 12:01:44 PM



799Noise, vibration, and harshness (NVH) in engine system design

noises. The intake air noise can be reduced by two types of silencers, the

absorptive (dissipative) type and the reactive type.

The absorptive silencers use a duct or parallel ducts lined with sound-

absorptive materials to dissipate the acoustic energy to heat. They are

effective in the medium- and high-frequency ranges. They tend to have poor

attenuation at low frequencies and quick erosion in the absorptive lining due

to the high mean ow velocities. The use of adsorptive silencers in internal

combustion engines is generally limited (Onorati, 1999).

The reactive silencers include expansion chambers (e.g., large-volume

air cleaners) and cancellation resonators (e.g., the Helmholtz or quarter

wavelength type). The resonator is tuned to certain frequency or frequencies

with a specic design of volume, hole opening area, effective length and

internal structure (e.g., number of perforates or internal orices along the

passage). When the induction system produces an air pressure wave the

resonator reects a waveform with the same frequency and amplitude but

opposite phase thereby cancelling the induction noise at that frequency. The

resonator is only effective within the narrow operating frequency range it

is tuned for, usually in the low- and medium-frequency ranges. Wide band

resonators can be used to reduce specic noises from the induction system

or the turbocharger.

The silencer performance is characterized by insertion loss and transmission

loss. The insertion loss refers to the difference in the sound pressure level

at a certain location before and after using the silencer. The transmission

loss refers to the difference in the sound power level between the inlet and

outlet of the silencer (Tao and Seybert, 2003).

It should be noted that all the silencing techniques usually conict with

the tight packaging requirement in the engine compartment. Silencers should

be designed to suit the frequency spectrum of the noise source with minimal

pressure losses and good packaging size. Number of stages, volume, pressure

drop, and expansion ratio are the critical design parameters in silencer

selection. The intake silencer is usually designed as a part of the air cleaner

(lter) and it is a standard component in almost all engines.

A review of piston engine intake and exhaust acoustic design is provided

by Davies (1996). Gas wave dynamics simulation of engine intake acoustics

was conducted by Silvestri et al. (1994).

11.10.2 Exhaust noise

The excitation source of the engine exhaust noise is the pressure pulses in

the exhaust duct that are generated by the periodic charging and discharging

processes of the engine. The high-velocity exhaust ow in the ducts generates

signicant turbulence and vortex shedding and the associated noises. The noise

due to the mechanical vibration of the exhaust pipe can also be associated

Diesel-Xin-11.indd 799 5/5/11 12:01:44 PM

