Xin Q. Diesel Engine System Design

800 Diesel engine system design

as a part of the exhaust noise, called shell noise. Like the intake noise, the

exhaust noise is usually dominated by low-frequency contents although the

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

range of frequencies. The frequency of the strongest exhaust noise contents

in the low-frequency range can be estimated in a similar way to the intake

noise by using equation 11.13.

Exhaust noise and its resonant amplication are largely affected by the

ow resistance in the aftertreatment system. Sharp resonances can be reduced

by ow resistance. The ow resistance is caused by the viscous force acting

between the exhaust gas and the tube walls or among the turbulent gas ows

of vortices. The ow resistance (i.e., exhaust restriction) results in a loss in

the static pressure and negatively affects engine performance.

The exhaust noise level increases as the engine speed or load increases

mainly because the exhaust gas ow rate, the pressure pulsation amplitude

and frequency increase with engine speed or load. Exhaust system design has

a major inuence on the exhaust noise through gas pressure wave dynamics.

Engine displacement, exhaust valve size, valve opening timing, in-cylinder

pressure level at exhaust valve opening, and exhaust runner and manifold

designs all affect the exhaust noise. Long and thin exhaust tailpipe may

help reduce the exhaust noise. However, excessively thin tailpipe increases

the exhaust restriction and thus hurts the engine breathing performance.

Overall automotive exhaust acoustic system design is discussed by Pang et

al. (2003).

Similar to the intake noise control, there are generally three methods to

control exhaust noise: (1) using silencers; (2) reducing the pulsation magnitude

of the exhaust gas pressure; and (3) reducing the turbulence level of the exhaust

valve gas ow. Exhaust noise control is reviewed by Munjal (1981).

Exhaust noise silencing can be achieved by using reactive (i.e., reective,

sound wave reection, or cancellation) type mufers, reactive-dissipative

perforate type mufers, or absorptive (i.e., dissipative) silencers, or a

combination of them. The constriction type of silencers relies on orice

restrictions and causes excessively high ow restriction; therefore it is not

desirable. The absorptive and cancellation types of silencers are usually

the preferred choice. The absorptive silencer reduces the exhaust noise by

dissipating the sound energy in porous medium, while the cancellation type

silencer cancels the sound pressure waves by out-of-phase waves to reduce

the noise.

The design of the silencer is usually a compromise between acoustic

attenuation and exhaust (or intake) ow restriction (pressure drop). The goal

is to achieve good engine breathing performance at acceptable noise levels.

The overall radiated tailpipe noise is the sum of both the ‘gas pulse’ noise

and the aerodynamic ‘gas ow’ noise (Onorati, 1999). Silencer design needs

to avoid the generation of excessive gas ow noise inside the silencer. The

Diesel-Xin-11.indd 800 5/5/11 12:01:44 PM



801Noise, vibration, and harshness (NVH) in engine system design

mechanism and the theory of the exhaust pulse noise and the ow noise

were explained by Kunz (1999) and Pang et al. (2005). The pulse noise

can be simulated by using nonlinear gas wave dynamics models in engine

cycle simulations. The ow noise usually has to be estimated based on semi-

empirical formulae. Analytical silencer (mufer) design was presented by

Yadav et al. (2007).

Engine cycle and gas wave dynamics simulations play a key role in

exhaust noise control. Engine exhaust system acoustics modeling was

reviewed by Jones (1984) and Onorati (1999). The basic theory of exhaust

acoustic modeling was presented by Ghafouri and Ricci (1993). Engine

exhaust noise modeling was conducted by Onorati (1995), Isshiki et al.

(1996), Morel et al. (1999), and Siano et al. (2005). The details of silencer

simulation with the one-dimensional nonlinear gas wave dynamics models

were provided by Onorati (1999). The simulation of the unsteady ows in

the coupled engine–mufer system can predict the sound pressure level

spectra of the radiated tailpipe noise and the inuence of mufer design on

engine performance. The nonlinear gas dynamics model is more advanced

and reliable than the over-simplied linear acoustic model which is based on

the hypothesis of small pressure perturbations in the pipes. Onorati (1999)

stated in his review article that the silencer simulation technique has reached

a stage where it is capable of giving good predictions of attenuation curves

compared with the measured data over a broad frequency band for a wide

range of mufing pipe systems.

Diesel particulate lter (DPF) has been widely used in modern diesel

engines to control particulate matter. Although they can be separated, the

DPF and the mufer are often combined into one unit in order to save cost

and packaging space. In this case, the design of the combined unit needs

to balance the requirements on particulate ltration and exhaust noise

attenuation. Understanding the noise attenuation characteristics of the DPF

as a mufer is a new challenging and interesting area to the diesel engine

industry. Integrated noise and particulates control is the future direction in

acoustic modeling and design. DPF acoustics was researched by Katari et

al. (2004), Allam and Åbom (2005, 2006), and Feng et al. (2008).

In summary, exhaust noise prediction needs to be conducted in the

following three areas:

∑ nonlinear simulation of gas wave dynamics to predict the gas pulse

noise

∑ using and developing empirical formula to predict the gas ow noise

∑ DPF acoustic simulation.

There are two approaches to simulate the exhaust noise: (1) using a detailed

silencer or DPF model (e.g., obtained from the supplier) in the nonlinear

gas dynamics model; and (2) using a simplied ow restriction model and

Diesel-Xin-11.indd 801 5/5/11 12:01:44 PM



802 Diesel engine system design

available attenuation spectra curves of the silencer (or the DPF) or empirical

correlations to estimate the tailpipe noise.

11.10.3 Turbocharger noise

Turbocharging changes the intake/exhaust noises and the engine NVH

characteristics. The compressor and the turbine reduce the sound propagation

through the intake and the exhaust systems and the intake/exhaust noise is

generally reduced because the compressor housing behaves like a small

reactive silencer element and the turbine usually dissipates the exhaust

ow pulsations. The turbine wheel damps the exhaust pulsations so that the

exhaust noise at the turbine outlet is much lower than the exhaust noise in

naturally aspirated engines.

However, the turbocharger may generate sound quality issues such as

the whooshing noise (Evans and Ward, 2006) due to the high ow and the

whining noise related to the number of blades in the compressor and the

turbine. The turbocharger whining noise in different frequency ranges can

be reduced by proper rotor balance specication or by increasing the gap

between the wheel and the ap of the turbine (Lu and Jen, 2007).

The heat shield of the turbocharger is another source of noise. Lu and

Jen (2007) showed that the turbocharger had 2–4 dB noise reduction after

the heat shield was removed. They pointed out that the heat shield required

for the turbocharger has a negative effect on noise due to its large radiating

surface and weak rigidity. Therefore, the shape and the stiffness of the heat

shield should be optimized to reduce the radiation noise. Lu and Jen (2007)

found that the noise renement by a double-layer heat shield design was

around 1–3 dB compared with a single-layer shield.

Other turbocharger noises include the following: (1) some resonances in

certain frequency range at the exhaust tailpipe; (2) some local resonances

with the turbocharger supporting bracket when the component stiffness is

not high enough; and (3) some high-frequency noise associated with the high

rotating speed, and certain sub-harmonic and imbalance vibrations caused

by rotor dynamics and the blade vibration which is induced by the exhaust

gas pulsations.

Silencers can be used on compressors to alleviate their turbulent intake ow

noise problems. The silencers are usually hole-chamber resonators, interference

resonators (e.g., Herschel–Quincke tube) or absorptive silencers.

Turbocharger noise is reviewed by Rämmal and Åbom (2007). Turbocharger

noise control was researched by Evans and Ward (2005), Trochon (2001),

and Lee et al. (2009a). Turbocharger structural vibration is analyzed by

Inagaki et al. (1993). The acoustics of charge air cooler, which is related to

turbocharging, is analyzed by Knutsson and Åbom (2007).

Diesel-Xin-11.indd 802 5/5/11 12:01:44 PM



803Noise, vibration, and harshness (NVH) in engine system design

11.10.4 Cooling fan noise

The cooling fan is used for engine coolant temperature and heat rejection

controls. The cooling fan noise is a signicant part of vehicle NVH and

important for automotive cabin quietness and comfort. As engine power or

heat rejection increases in low-NO

x

engines, it is increasingly challenging

to design a cooling fan that has a small size, low noise, low power, and low

cost. The cooling fan operates in a wide range of cooling air mass ows, and

its peak efciency and minimum noise operating points do not necessarily

coincide. Optimizing both aerodynamic performance and noise characteristics

is important in selecting the best fan design for a given application.

The cooling fan noise is related to fan size (specic diameter), fan speed

(specic speed), blade design, and airway structures. Design measures such

as decreasing tip clearance and optimizing pitch angle may increase the fan

efciency and reduce the fan noise. Fan blade design can also minimize the fan

vibration and the fan noise. For example, the fan blade passage noise is one

important element of the total fan noise, and is caused by the non-uniformity

of the upstream and downstream ows of the fan rotor. The discrete noise at

the blade passing frequency can cause serious noise problems. Appropriate

blade spacing design may reduce the noise. Optimizing the airway structures

(ow passages) can improve the uniformity of the inow air distribution and

reduce the lift uctuations of the fan blades thus reducing the noise level of

the fan. Moreover, psychoacoustic characteristics (i.e., subjective assessment

of the sound comfort and annoyance) of the cooling fan are also important,

in addition to the requirements of sound pressure level. The engine cooling

fan noise and performance was elaborated by Mellin (1980).

11.11 Engine brake noise

Engine brake performance and design are introduced in Chapter 6. The

compression-release engine brake has unique noise characteristics. Details

on engine brake noise are discussed in Section 6.4.8. The noise of the

compression-release engine brake is an exhaust-ow induced noise related

to the gas wave dynamics in the exhaust manifold and the tailpipe. The

noise is caused by a sudden blow-down event of the in-cylinder charge

near the braking TDC. The noise can be controlled by using better braking

mechanisms that control the engine braking valve ow characteristics, or it

can be controlled by advanced mufers. Engine cycle simulation coupled

with an acoustic simulation model of the exhaust system can be used to

analyze the engine brake noise. More information on engine brake noise can

be found in Reinhart and Wahl (1997) and Wahl and Reinhart (1997).

Diesel-Xin-11.indd 803 5/5/11 12:01:44 PM



804 Diesel engine system design

11.12 Diesel engine system design models of noise,

vibration, and harshness (NVH)

11.12.1 Level-1 system noise model

The level-1 system noise model used in diesel engine system design is a

non-instantaneous (non-crank-angle-resolution) semi-empirical model. The

coef cients or constants in such a model are limited to certain particular

engines from which the formula is derived, and it cannot be broadly applied

to pursue accurate absolute values. This type of model is useful to understand

the basic major parametric dependency of the noise. It can be used to roughly

estimate the relative effect or trend.

The overall engine noise level can be expressed as:

pp

SP

pp

LE

pp

SP

pp

LE

pp

SP

pp

i

SP

Li

SPLiSP

,,

pp

,,

pp

LE

pp

,,

pp

LE

pp

i

,,

i

pp

i

pp

,,

pp

i

pp

SP,,SP

Li,,Li

SPLiSP,,SPLiSP

pp =pp

pp pp

S

ppSpp

11.14

where p

SPL,E

is the overall engine sound pressure level at one meter from the

engine surface. p

SPL,i

represents the contribution from combustion, piston

slap, valvetrain, gears, fuel injection and accessories.

Jenkins (1975) pointed out that the engine noises induced by different

excitation sources can be expressed by the following functions:

pf

NB

p

pf

SP

pf

LC

pf

SP

pf

LC

pf

SP

pf

om

pf

bus

pf

io

pf

n

pf

E

n

E

NB

E

NB

SP

LP

SPLPSP

istonSlap

,

pf

,

pf

LC

pf

,

pf

LC

pf

,

LP,LP

pf= pf

(,

NB(,NB

E

(,

E

NB

E

NB(,NB

E

NB

n

(,

n

NB

n

NB(,NB

n

NB

c

NB(,NB

c

NB

,

NB ,NB

E

,

E

NB

E

NB ,NB

E

NB

)

F )F

=

E

)

E

)

F )F

E

F )F

(,

,

)

(,

2

(,

2

(,

,

5.

8

fN

(,fN(,

Bp

,Bp ,

f

)f )

pf

= pf=

,

pf

,

Nf

(,Nf(,

5.

Nf

5.

(,

5.

(,Nf(,

5.

(,

8

Nf

8

(,

8

(,Nf(,

8

(,

EE

(,

EE

(,

,

EE

,

Bp

EE

Bp

,Bp ,

EE

,Bp ,

P

)

P

)

f

P

f

)f )

P

)f )

pf

SP

pf

LV

pf

,

pf

,LV,

pf

,

pf

SP

pf

LV

pf

SP

pf

alvet

pf

LV

pf

alvet

pf

LV

pf

ra

pf

alvet

pf

ra

pf

alvet

pf

in

pf

EV

Nf

EV

Nf

(,Nf(,

EV

(,Nf(,

TV

EVTVEV

TTVT

EVTEVTVEVTEV

T

TVTTV

njectio

nE

fuel

i

c

TV

c

TV

pf

SP

pf

SP

LF

pf

LF

SPLFSP

pf

SPLFSP

uelI

pf

uelI

njectio

pf

njectio

uelInjectiouelI

pf

uelInjectiouelI

nE

pf

nE

Nm

v

TV

,

TV

)

pf= pf

nE

pf

nE

=

nE

pf

nE

(,

nE

(,

nE

Nm(,Nm

nE

Nm

nE

(,

nE

Nm

nE

,

fuel

,

fuel

Nm ,Nm

,

pf

,

pf

LF

pf

LF,LF

pf

LF

4.

(,

4.

(,

Nm(,Nm

4.

Nm(,Nm

3

(,

3

(,

Nm(,Nm

3

Nm(,Nm

Nm



Nm

nj

nnjn

)

Ï

Ì

Ô

Ï

Ô

Ï

Ô

Ì

Ô

Ì

Ô

Ó

Ô

Ì

Ô

Ì

Ô

Ó

Ô

Ó

Ô

11.15

where N

E

is the engine speed, B

E

is the cylinder bore diameter, F

E

is the

engine load, n

c

is a function of the combustion system, p is the cylinder

pressure, f

P

is a piston design factor, f

VT

is a valvetrain design factor, c

VT

is the valvetrain lash,



m

fuel

is the fuel rate, and v

inj

is the velocity of the

injector at the start of the injection.

The most basic form of the overall engine noise model was deduced based

on the noise measurements on a large number of engines for combustion-

induced noise (Anderton et al., 1970; Challen, 1975; Corcione et al., 2003)

as follows:

p

SPL,E

= C

1

log

10

N

E

+ C

2

log

10

B

E

+ C

3

11.16

where p

SPL,E

is the overall engine sound pressure level in dB(A) at one meter

from the engine surface, N

E

is the engine speed (rpm), B

E

is the cylinder bore

diameter, and C

1

, C

2

and C

3

are empirical constants obtained by  tting the

measurement data. C

1

is related to the cylinder pressure level spectra (i.e.,

Diesel-Xin-11.indd 804 5/5/11 12:01:45 PM



805Noise, vibration, and harshness (NVH) in engine system design

the combustion system), and C

3

re ects the engine structural characteristics.

Note that the two most in uential parameters re ected in equation 11.16 are

engine speed and cylinder bore size. The latter actually re ects the combustion

chamber size. The load dependency of the engine noise was found to be very

weak. For four-stroke turbocharged diesel engines, C

1

= 40, C

2

= 50, C

3

=

–66.5 if B

E

is in inches and C

3

= –136.7 if B

E

is in mm (Anderton et al.,

1970). The implication of equation 11.16 is that it is important to choose a

low rated speed and a small bore size (or using more cylinders for a given

engine displacement) in engine system design for low noise characteristics.

Note that equation 11.16 was obtained from diesel engines having low BMEP

and low fuel injection pressure in the old era.

For a given cylinder bore size and engine speed, there is a large variation

in the engine noise due to the difference in the combustion system. In order

to analyze the effect of the cylinder pressure rise on the combustion noise

and characterize different combustion systems, Hawksley and Anderton

(1978) proposed

n

c

= 4.3 – 0.21 f

rpr

11.17

where n

c

is a combustion index, and f

rpr

is the rate of pressure rise in bar

per degree crank angle. f

rpr

varies greatly with engine load. Based on the

following relationship (Anderton et al., 1970) about noise intensity,

IN

B

s

IN

s

IN

E

n

E

B

E

B

c

IN IN

5

INµIN

11.18

they proposed that the engine noise level for turbocharged diesel engines

followed

pN

Bf

pN

SP

pN

LE

pN

SP

pN

LE

pN

SP

pN

pr

,1

pN

,1

pN

LE

pN

,1

pN

LE

pN

01

EE01EE

pN =pN

pN

,1

pN =pN

,1

pN

30

pN 30pN

pN

,1

pN 30pN

,1

pN

lo

pNlopN

pN

,1

pNlopN

,1

pN

g

pNg pN

,1

g

,1

pN

,1

pNg pN

,1

pN

01

g

01

pN

01

pNg pN

01

pN

EE01EE

g

EE01EE

+ 50

01

+ 50

01

EE01EE

+ 50

EE01EE

lo

01

lo

01

EE01EE

lo

EE01EE

g

Bfg Bf

EE

g

EE

Bf

EE

Bfg Bf

EE

Bf

01

g

01

EE01EE

g

EE01EE

0

g

0

EE0EE

g

EE0EE

Bf+ (Bf

Bf2.1Bf

– 13)

r

Bf

r

Bf

prrpr

lo

llol

g

5455

– 103

10

N

E

11.19

Note that this formula includes the effect of engine load via the parameter

f

pr

.

Anderton (1979) summarized the relationship between the overall engine

noise level and  ve key parameters (i.e., engine speed, bore diameter, cylinder

pressure spectrum at a given frequency, a combustion index, and a shape

factor) based on the noise measurements on a large number of engines as:

pn

N

BC

pn

SP

pn

LE

pn

SP

pn

LE

pn

SP

pn

E

Ep

,1

pn

,1

pn

LE

pn

,1

pn

LE

pn

c,1c

01

0(

BC

0(

BC

Ep0(Ep

BC

Ep

BC

0(

BC

Ep

BC

r0(r

1

pn =pn

pn

,1

pn =pn

,1

pn

10

pn 10pn

pn

,1

pn 10pn

,1

pn

· log

,1

· log

,1

1000

01

1000

01

+ 50

01

+ 50

01

lo

01

lo

01

g

BCg BC

01

g

01

0(

g

0(

BC

0(

BCg BC

0(

BC

Ep0(Ep

g

Ep0(Ep

BC

Ep

BC

0(

BC

Ep

BCg BC

Ep

BC

0(

BC

Ep

BC

BC+ BC

BC

0(

BC+ BC

0(

BC

Ep

BC

0(

BC

Ep

BC+ BC

Ep

BC

0(

BC

Ep

BC

..

0)

2

–

C

11.20

where p

SPL,E

is the overall engine noise level in dB(A), N

E

is the engine

speed (rpm), B

E

is the cylinder bore diameter (mm), C

pr(1.0)

is the value (in

dB) of the simpli ed and reduced cylinder pressure spectrum at 1.0 Hz/rev/

min, n

c

is the combustion index, which is equal to one tenth of the slope

Diesel-Xin-11.indd 805 5/5/11 12:01:46 PM



806 Diesel engine system design

of the simpli ed and reduced cylinder pressure spectrum in dB/decade, and

C

2

is an A-weighted shape factor as a function of the combustion index n

c

.

The average constants in equation 11.20 of different types of engines were

given by Anderton (1979). For example, C

pr(1.0)

= 144 dB and n

c

= 5.1 for

turbocharged two- and four-stroke diesel engines.

Tung and Crocker (1982) used a simpler formula in the following

form:

pC

NC

WC

pC

SP

pC

LE

pC

SP

pC

LE

pC

SP

pC

EE

NC

EE

NC

,1

pC

,1

pC

LE

pC

,1

pC

LE

pC

10

21

EE21EE

03

WC

03

WC

EE03EE

WC

EE

WC

03

WC

EE

WC

pC =pC

pC

,1

pC =pC

,1

pC

l

pC lpC

,1

l

,1

pC

,1

pC lpC

,1

pC

og

NC +NC

NC

EE

NC +NC

EE

NC

l

NC lNC

EE

l

EE

NC

EE

NC lNC

EE

NC

21

l

21

EE21EE

l

EE21EE

og

21

og

21

EE21EE

og

EE21EE

WC +WC

WC

03

WC +WC

03

WC



11.21

where



W

E

W

E

W

is the engine brake power.

The research focus on mechanical noise has been concentrated on the

piston slap noise because it is the most important mechanical noise in diesel

engines. Usami et al. (1975) provided a formula by using measured data of

an indirect injection diesel engine at 3700 rpm as follows:

Ec

pc

c

k,

Ec

k,

Ec

sl

Ec

sl

Ec

ap

Ec

ap

Ec

slapsl

Ec

sl

Ec

ap

Ec

sl

Ec

P

pc

SP

pc

LE

pc

SP

pc

LE

pc

SP

pc

PP

c

PP

c

EcµEc

= 10

pc= 10pc

lo

pclopc

g(

pcg(pc

+ 1.74

PP

+ 1.74

PP

1.

3

,1

pc

,1

pc

LE

pc

,1

pc

LE

pc

pc= 10pc

,1

pc= 10pc

pclopc

,1

pclopc

g(

,1

g(

pcg(pc

,1

pcg(pc

01

pc

01

pc

PP01PP

pcg(pc

01

pcg(pc

1.

3

2

1.

.3

1..31.

)

–

80.4

Ï

Ì

Ô

Ï

Ô

Ï

Ì

Ô

Ì

Ó

Ô

Ì

Ô

Ì

Ó

Ô

Ó

11.22

where E

k,slap

is the impact energy of piston slap, c

P1

and c

P2

are the piston-

to-bore clearances at the skirt top and bottom, respectively. Note that this

equation is independent of engine speed and only includes basically one

parameter, the piston-to-bore clearance c

P

.

Lalor et al. (1980) gave the noise intensity due to the kinetic energy of

piston slap as a function of engine speed N

E

and piston-to-bore clearance

c

P

,

IN

c

sE

IN

sE

IN

P

IN

sE

IN IN

sE

IN

0.667

1.333

INµIN

IN

sE

INµIN

sE

IN

11.23

Priede (1982) developed a formula for the sound intensity of piston slap

as

I

F

t

mc

KB

s

P

F

P

F

PP

mc

PP

mc

s,

KB

s,

KB

liner

KB

liner

KB

E

KB

E

KB

d

·

KB · KB

1/

3

1.333

µ

Ê

Ë

Á

Ê

Á

Ê

Ë

Á

Ë

ˆ

mc

ˆ

mc

¯

mc

¯

mc

PP

¯

PP

mc

PP

mc

¯

mc

PP

mc

˜

mc

ˆ

˜

ˆ

mc

ˆ

mc

˜

mc

ˆ

mc

¯

˜

¯

mc

¯

mc

˜

mc

¯

mc

PP

¯

PP

˜

PP

¯

PP

mc

PP

mc

¯

mc

PP

mc

˜

mc

PP

mc

¯

mc

PP

mc

11.24

where I

s

is the sound intensity, c

P

is the piston-to-bore clearance, m

P

is the

piston mass, K

s,liner

is the cylinder liner stiffness, dF

P

/dt is the variation rate

of the piston lateral force.

Other basic design calculation guidelines for low-noise diesel engines

were provided by Jenkins et al. (1973).

Stout et al. (2005) proposed the following empirical formula for three V6

gasoline engines:

Diesel-Xin-11.indd 806 5/5/11 12:01:47 PM



807Noise, vibration, and harshness (NVH) in engine system design

pN

SB

pN

SP

pN

LE

pN

SP

pN

LE

pN

SP

pN

E

SB

E

SB

,1

pN

,1

pN

LE

pN

,1

pN

LE

pN

01

EE01EE

01

SB

01

SB

0

SB

0

SB

= 50

pN= 50pN

pN

,1

pN= 50pN

,1

pN

lo

pNlopN

pN

,1

pNlopN

,1

pN

g

pNg pN

,1

g

,1

pN

,1

pNg pN

,1

pN

01

g

01

pN

01

pNg pN

01

pN

EE01EE

g

EE01EE

+ 40

01

+ 40

01

EE01EE

+ 40

EE01EE

lo

01

lo

01

EE01EE

lo

EE01EE

g

SBg SB

EE

g

EE

01

g

01

EE01EE

g

EE01EE

01

g

01

SB

01

SBg SB

01

SB

EE01EE

g

EE01EE

SB

EE

SB

01

SB

EE

SBg SB

EE

SB

01

SB

EE

SB

+ 30

SB+ 30SB

SB

01

SB+ 30SB

01

SB

log

SBlogSB

01

log

01

SB

01

SBlogSB

01

SB

– 223.

.5

– 223..5 – 223.

0.9

50

lo

g

+ 1.6

40

lo

g

,1

g

,1

g

01

g

01

g

+ 1.6

01

+ 1.6

40

01

40

lo

01

lo

g

01

g

0

pN

=pN =

0.9 pN 0.9

50pN 50

lopNlo

g pNg

,1

pN

,1

=

,1

=pN =

,1

=

0.9

,1

0.9 pN 0.9

,1

0.9

50

,1

50pN 50

,1

50

lo

,1

lopNlo

,1

lo

g

,1

g pNg

,1

g

01

g pNg

01

g

S

pN

SP

pN

Lm

pN

,1

pN

,1Lm,1

pN

,1

pN

SP

pN

Lm

pN

SP

pN

ech

pN

,1

pN

,1ech,1

pN

,1

g

01

g

E

g

01

g

¥¥

g ¥¥g

+ 1.6 ¥¥+ 1.6

01

¥¥

01

g

01

g ¥¥g

01

g

+ 1.6

01

+ 1.6 ¥¥+ 1.6

01

+ 1.6

pN¥¥pN

50pN 50¥¥ 50pN 50

lopNlo¥¥lopNlo

g pNg ¥¥g pNg

,1

pN

,1

¥¥

,1

pN

,1

50

,1

50pN 50

,1

50¥¥ 50

,1

50pN 50

,1

50

lo

,1

lopNlo

,1

lo¥¥lo

,1

lopNlo

,1

lo

g

,1

g pNg

,1

g ¥¥g

,1

g pNg

,1

g

01

g pNg

01

g ¥¥g

01

g pNg

01

g

01

g

E

g

01

g ¥¥g

01

g

E

g

01

g

EEE

S

E

S

E

S

E

S

E

B

E

B

E

–

0.7

30

log

– 158

10

¥

Ï

Ì

Ô

Ï

Ô

Ï

Ô

Ì

Ô

Ì

Ô

Ó

Ô

Ì

Ô

Ì

Ô

Ó

Ô

Ó

Ô

11.25

where p

SPL,E

is the overall engine noise level, p

SPL,mech

is the mechanical

noise level, and S

E

is the engine stroke.

Zhao and Reinhart (1999) proposed the following empirical formulae

including the in uence of the geartrain noise based on their measurements

on 14 heavy-duty diesel engines of 1994 EPA or Euro II emissions (from 3

liters to 16 liters engine displacement):

pn

i

p

pn

SP

pn

LE

pn

SP

pn

LE

pn

SP

pn

Rate

pn

dG

pn

dG

pn

F

SP

LE

SPLESP

,,

pn

,,

pn

LE

pn

,,

pn

LE

pn

,,

LE,,LE

= 93.9 + 2.6

pn= 93.9 + 2.6pn

pn

dG

pn= 93.9 + 2.6pn

dG

pn

(

pn(pn

pn

dG

pn(pn

dG

pn

– 1)

+ 4.8

PeakTorque

PPeakTorqueP

GF

SP

LE

SPLESP

Lo

w

ni

GF

ni

GF

ni

GF

ni

GF

p

= 91.5 + 2.5

(

– 1)

GF

– 1)

GF

ni – 1)ni

GF

ni

GF

– 1)

GF

ni

GF

ni +ni

GF

ni

GF

+

GF

ni

GF

,,

LE,,LE

Id

IIdI

le

G

SP

LE

SPLESP

Hi

ghI

dl

ghIdlghI

e

n

p

= 85.9 + 0.8

(

– 1)

= 99 +

,,

LE,,LE

1.2

11.21

(

– 1)

n

G

Ï

Ì

Ô

Ï

Ô

Ï

Ô

Ì

Ô

Ì

Ô

Ó

Ô

Ì

Ô

Ì

Ô

Ó

Ô

Ó

Ô

Ó

Ô

Ó

Ô

Ó

Ô

Ó

11.26

where p

SPL,E,Rated

is the overall engine noise level at rated power measured

according to SAE J1074 at 1 m from the engine surface, n

G

is the number

of gear meshes between the crankshaft and the fuel system, and i

F

= 1 for

front geartrains, while i

F

= 0 for rear geartrains. This study suggests that

the engine noise level is more sensitive to the number of gear meshes at

full load than at part load.

It is observed that all the above equations actually provide limited values

to diesel engine system design due to their limited range of parametric

dependency and simpli ed relationships. More physics-based models need to

be developed for engine system design, and they are the level-2 models.

11.12.2 Level-2 system NVH model

The level-2 NVH model used in diesel engine system design possesses the

following features on an instantaneous basis with crank-angle resolution:

∑ Model the excitation source at various complexity levels.

∑ Use a ‘virtual’ combustion noise meter to estimate the combustion

noise by processing the cylinder pressure data generated from cycle

simulations.

∑ Use a piston-assembly-liner dynamics model to estimate the piston slap

noise.

∑ Use a valvetrain dynamics model to estimate the valvetrain vibration

and noise.

∑ Use a gas wave dynamics cycle simulation model coupled with an

Diesel-Xin-11.indd 807 5/5/11 12:01:48 PM



808 Diesel engine system design

acoustic model of the silencer or the DPF to predict the intake and

exhaust noises.

∑ The model can be real-time capable, if simplied, for advanced active

noise control in engine controls and for fault diagnostics.

Note that all the modeling tools involved in these topics are the tools used

by a system design engineer in his/her daily work. The system engineer

has the expertise and a clear incentive to conduct this NVH design/analysis

work at the system level.

Combustion noise meters (e.g., the Lucas CAV and AVL combustion

noise meters) have been widely used in the NVH and engine calibration

areas to conveniently measure the effects of combustion (or essentially the

cylinder pressure changes) on engine noise. The combustion noise meter

was developed based on the following principle about the relationship

between the frequency spectrum of the cylinder pressure and the measured

frequency spectrum of the total engine noise level at one meter from the

engine surface: by subtracting the spectra of the in-cylinder pressure from

the engine noise spectra, the structural attenuation can be determined. Such a

structural attenuation spectrum can be obtained from many engines through

measurement and then an average attenuation spectrum can be calculated

and entered into the combustion noise meter. Then, the meter can be used

to predict the combustion noise once the cylinder pressure data are known

without measuring the overall engine noise level. The noise change due

to the cylinder pressure change can be regarded as the contribution of the

combustion noise change. This technique is acceptable for engines having

similar spectra of structural attenuation.

The combustion noise is equal to the cylinder pressure minus the structural

attenuation (calculated in frequency spectra in the unit of dB). The combustion

chamber pressure signals are commonly measured in engine development in

crank-angle resolution. The combustion meter uses the crank-angle-resolution

signals from a cylinder pressure transducer to calculate the combustion noise

level, which would be the noise level measured 1 m from the engine if only

the ‘combustion related’ noise were included. A combustion meter uses a

frequency analysis (Fast Fourier Transfer) of the cylinder pressure signals to

generate frequency spectra (e.g., one-third octave). The total noise level is the

logarithmic mean value of the spectrum. Then, it subtracts the engine block

structural attenuation function via an attenuation lter (usually a mean free-

eld response lter representing an average engine mass/structure response

based on the experiments conducted on many engines) from the frequency

spectrum of the cylinder pressure to obtain the combustion noise spectrum.

The attenuation lter may also include a function to lter out the combustion

chamber resonance with selectable low-pass lters (e.g., 10 kHz low-pass).

The combustion noise signals are then processed by an A-weighting lter to

Diesel-Xin-11.indd 808 5/5/11 12:01:48 PM



809Noise, vibration, and harshness (NVH) in engine system design

produce the noise sensitive to human hearing in the unit of dB(A). Finally,

a single root-mean-square value of the combustion noise is produced. The

combustion noise can be measured in this way by using cylinder pressure

instead of using a sound meter. Moreover, the mechanical noises are separated

out from the measured combustion noise. The meter can also be used for

recording the combustion noise during transient conditions.

Cylinder pressure derivative with respect to time is an output of the

combustion noise meter. Usually, there is a direct correlation between the

cylinder pressure derivative and the combustion noise. The correlation

between the peak cylinder pressure and the combustion noise is much less

clear although increasing the cylinder pressure sometimes leads to increased

combustion noise for a given turbocharged engine. Note that peak cylinder

pressure affects the mechanical noises such as the piston slap noise and the

gear rattle noise. Details of the combustion noise meters are provided by

Russell (1984), Russell and Young (1985), Reinhart (1987), and Wang et al.

(2007). More accurate accounting for the structural attenuation characteristics

is discussed by Lee M et al. (2009). The limitations of the classical ‘block

attenuation curve’ method and a more advanced alternative approach were

presented by Torregrosa et al. (2007).

A system design engineer uses engine cycle simulations to analyze the

cylinder pressure signals in a virtual world. Although the cylinder pressure

data obtained from the engine cycle simulation may usually contain less

details related to the pressure derivatives (e.g., dp/df or dp/dt) than the real

test data and cannot reect the design changes of the combustion system,

the simulation data are useful to correctly reect certain operating parameter

changes such as the effects of engine speed, fueling rate, air–fuel ratio,

EGR rate, fuel injection timing, ignition delay, and cetane number. The

combustion noise is sensitive to these parameters (Reinhart, 1987) and they

are important system design parameters. Therefore, a virtual combustion

noise meter can be developed to estimate the combustion noise changes in

the entire speed–load domain and during acceleration transients for major

system design parameters. The structural attenuation function can be measured

and developed in the model for a particular engine. Note that the structural

attenuation of a modern diesel engine can be overall much higher than that

used in a combustion noise meter which is based on the old engine structure

data. Even if the structural attenuation response is not very accurate, this

approach is still effective to predict the relative difference in the combustion

noise instead of the absolute values. With further development of advanced

zero- or one-dimensional combustion models the prediction capability for

combustion and cylinder pressure details in the cycle simulation software

will continue to improve. Moving to the virtual world for combustion noise

analysis is the right direction, and diesel engine system design can play a

key role in this process by effectively integrating combustion noise analysis

Diesel-Xin-11.indd 809 5/5/11 12:01:48 PM



Xin Q. Diesel Engine System Design

Подождите немного. Документ загружается.