Xin Q. Diesel Engine System Design

Подождите немного. Документ загружается.

770 Diesel engine system design

∑ Noise source identication for the contributions from combustion

noise, mechanical noise, and aerodynamic noise (from intake, exhaust,

turbocharger, and cooling fan) in the frequency spectrum

∑ breakdown of the mechanical noise at different engine speeds and loads

in the frequency spectrum (e.g., piston assembly, valvetrain, cranktrain,

block, geartrain, belt, chain, fuel system, pumps, and accessories)

∑ differences between the steady state and the transient noises

∑ noise transfer path and structural attenuation in the frequency

spectrum

∑ sound radiation from the engine surfaces (e.g., engine block, mounts,

crankcase, crank pulley, ywheel cover, front cover, valve cover, oil

pan, heat shields, manifolds, engine mounts, cylinder head, air lter,

exhaust pipe, turbocharger, pumps, alternator, and starter).

11.3.2 Differences between diesel and gasoline engine

noises

It is well recognized that the noise level of the diesel engine is higher than

that of the gasoline engine. The sound quality of the diesel engine also

tends to be worse due to its internal impulsive excitations which induce

high-frequency noises. The heterogeneous combustion in the diesel engine is

generally noisier than the homogeneous combustion in the gasoline engine due

to the high rate of rise of the in-cylinder pressure after the ignition delay. The

so-called diesel knocking refers mainly to the noise in the frequency range

of 500–6000 Hz. It has been recognized that diesel knocking is the major

contributor to poor sound quality perceived by the customer, particularly at

the low-idle condition.

The peak cylinder pressure in the gasoline engine is much lower than

that in the diesel engine. The moving components of the diesel engine are

designed relatively heavier and stronger than those in the gasoline engine

in order to meet the durability requirements at higher cylinder pressures.

Therefore, the diesel engine has higher mechanical impact noise than the

gasoline engine if compared at the same engine speed. However, if compared

at their respective rated speeds, their noise levels are often similar because

the gasoline engine has a higher rated speed. The acoustic similarity of the

combustion noise and the piston related noise in the diesel engine often

make their differentiation and a subjective rating of the piston noise difcult

(Künzel et al., 2000, 2001). Moreover, the diesel engine usually has more

auxiliary components such as the fuel injection pump and the turbocharger.

They are potential sources of additional noise and vibration.

On the other hand, the gasoline engine exhibits some noises that the diesel

engine does not have, for example the piston pin ticking noise at the low-

speed no-load conditions (Werkmann et al., 2005; Moshre et al., 2007),

Diesel-Xin-11.indd 770 5/5/11 12:01:40 PM



771Noise, vibration, and harshness (NVH) in engine system design

the cold start piston clatter noise due to a tertiary motion (Pollack et al.,

2005), and the prominent stick slip piston noise emitted from the excited

crankshaft (Beardmore, 1982; Werner, 1987). These noises are due to the

unique characteristics of the gas loading and the mass of the gasoline engine

(Künzel et al., 2000).

Stucklschwaiger et al. (1999) compared the differences in NVH between

the diesel and gasoline engines. The design congurations related to NVH

features at the vehicle and powertrain levels were discussed. Engine mounting,

engine balance, bank angles, balancing shaft design, and other NVH design

issues were elaborated for light-duty trucks in their study.

11.3.3 Diesel engine noise characteristics

Understanding the contributions to the total engine noise from combustion,

mechanical impacts and aerodynamic effects in the engine speed–load

domain and knowing their behavior in the frequency spectrum are important

for identifying and prioritizing corrective measures to control the noise. For

example, for combustion noise dominated engines design improvement should

be focused on combustion and fuel system design, emissions calibration,

and structural attenuation.

Engine full-load noise characteristics are especially important. Reduction

of the engine noise particularly under full-load operation is often required

in order to comply with vehicle noise regulations. The mechanical noise

usually increases with engine speed. Figure 11.1 shows that there is a large

difference between the ring and motoring noise levels, especially at low

engine speeds. At higher engine speeds the difference becomes smaller due

to the increasing mechanical noise. Figure 11.1 also shows the effect of the

piston on the noise.

Informative measurement results revealing the fundamental characteristics

of diesel engine noises are provided by Badawi et al. (2007). They plotted

the typical NVH events in the time–frequency domain, including valvetrain

operation, fuel injection, combustion, and piston slap. Govindswamy et

al. (2007) illustrate the diesel impulsive noises including diesel knocking,

injector ticking, and gear rattling. Diesel engine brake noise is introduced

in Chapter 6.

11.3.4 Engine noise identiﬁcation

In order to design a low noise engine one must know the sources and paths

of the noise generation and transmission. Several methods have been used

to separate the noise sources. An engine is red and motored respectively

to separate the mechanical sources from the combustion sources. Lead

covering techniques are used to identify the noise radiating characteristics

Diesel-Xin-11.indd 771 5/5/11 12:01:40 PM



772 Diesel engine system design

of any individual surface or component of the engine to be evaluated. The

entire engine is wrapped with a lead sheet of absorbing material (acoustically

masked), and then each component (e.g., the oil pan, the manifold) is exposed

one at a time to determine its contribution to the engine noise. The breakdown

and individual contributions of the mechanical noise can also be obtained

by removing the subject component one at a time during the test.

It is usually difcult to clearly separate the combustion noise from the

mechanical noise because the latter is affected by the gas loading coming

from the in-cylinder pressure. In order to separate these noises, the engine

noise can be measured at different levels of combustion excitation, and then

Sound pressure, overall level.

4-microphone average (1 meter)

Full load (ﬁring)

No load (ﬁring)

Motoring

1000 2000 3000 4000

rpm

1000 2000 3000 4000

Engine speed (rpm)

(a) The effect of engine load on engine sound

pressure level at different engine speeds

(b) The effect of removing the piston on engine

sound pressure level at no load

No load

With pistons (ﬁring)

Without pistons

Sound pressure level dB(A)

110

100

110

100

11.1 Effect of diesel engine operating conditions on noise level.

Diesel-Xin-11.indd 772 5/5/11 12:01:40 PM



773Noise, vibration, and harshness (NVH) in engine system design

the overall and spectral contributions from the combustion and mechanical

sources can be estimated. For example, at a given frequency, if the data

of ‘measured sound pressure level of the engine vs. the decibel level of

the cylinder pressure’ form a line, the slope of the line would indicate the

structural attenuation for the cylinder pressure, and the interception would

approximately reect the sound pressure level of the mechanical noise. The

structural attenuation essentially reects a type of transfer function of the

system, which is measured by the ratio of the output of the system (response)

at a certain location to the input of the system (i.e., the driving force) at

another location. Different transfer functions are obtained for different

input and output locations. Extensive research has been conducted in the

area of separating the combustion noise from the mechanical noise. This

area is important for accurately estimating the noise contribution from the

combustion in order to reduce it.

11.3.5 Transient engine noise

Most engine noise research has been conducted typically based on steady-

state noise measurements because the highly dynamic acoustic test cells are

not commonly available. However, the vehicle noise legislation requires

the pass-by test to be conducted during transient acceleration operation.

Diesel engine transient noise is signicantly different from that at the

corresponding steady-state conditions (i.e., if compared at the same engine

speed and same fueling rate). For example, the transient noise could exceed

the steady-state full-load noise by up to 6 dBA (Dhaenens et al., 2001).

If the vehicle accelerates after a period of idling or light-load running,

the noise during the acceleration transient will be higher than that at the

steady-state conditions due to longer ignition delay and higher cylinder

pressure gradient (rise rate). The longer ignition delay can be caused by

colder intake air temperature, lower boost pressure due to turbocharger

lag, lower combustion chamber wall temperature, and more advanced

dynamic injection timing. In fact, the transient engine noise depends on

the running history before the transient event occurs. Different running

history may produce different conclusions on the difference between the

transient noise and the steady-state noise (Shu et al., 2005b, 2006). The

difference in noise is inuenced largely by the difference in ignition delay.

Diesel engine noise under transient conditions was studied by Head and

Wake (1980), Rust and Thien (1987), Dhaenens et al. (2001), and Shu et

al. (2005b, 2006). Modeling the noise difference between the transient

(especially the acceleration transient as in the drive-by noise test) and the

steady-state conditions is very important.

Diesel-Xin-11.indd 773 5/5/11 12:01:41 PM



774 Diesel engine system design

11.3.6 Diesel engine NVH variation

Diesel engine NVH variation is a very important issue to consider in NVH

development, measurement and overall system design/planning. It includes

test variability and engine-to-engine variability. The statistical variability

may mask the small difference caused by the design changes. The theory of

engine NVH variability and the application in diesel engine noise reduction

were introduced by Reinhart et al. (2003). They illustrated that NVH

design decisions are often made based on a small measured difference that

is statistically insignicant. It is important to design a test plan in order to

achieve statistically signicant results. They used an example of the SAE

J1074 noise test to emphasize that if the data from a single comparison test are

used to evaluate the effect of a design change, the difference must be greater

than 0.8 dB to be more than 90% condent that there is a real difference.

In other words, a measured 0.8 dB difference can indicate a true difference

anywhere between 0 dB (no change) and 1.6 dB (a substantial change)! In

NVH system design, using advanced simulation tools to accurately quantify

the benet of design changes (especially the small changes), applying the

theory of design-for-variability in robust design, and considering the NVH

variability are three key elements to make the design successful.

11.3.7 Engine NVH development process

Good NVH characteristics are a major competitive advantage for an engine

product. Due to their complex nature, NVH failures often cause costly late

design changes. It is important to take NVH into account in diesel engine

system design from the concept stage and to develop practical and advanced

engineering rules that can be used in system design.

The engine NVH development process generally consists of two stages:

(1) the concept design and analysis stage; and (2) the product design and

testing validation stage. In the concept stage, NVH criteria are dened in

target setting/cascading with a top-down approach from the system level

cascaded to the component level. Internal engine excitation forces or noise

sources (e.g., combustion, mechanical, and aerodynamic noises) are simulated.

Usually, NVH problems are analyzed in a chain of input–transfer–response.

The input or excitation sources of the NVH problems can be handled by

multi-body system dynamics in engine system design to identify the force

loading. Since the best approach to control diesel engine NVH is to control

it at the source of excitation, engine system design plays a critical role in the

NVH development process because all the major engine design parameters

and the engine loads are optimized and determined at the system design

level. The transfer path of the NVH problems is usually handled by nite

element analysis (FEA) or statistical energy analysis (SEA). The acoustic

Diesel-Xin-11.indd 774 5/5/11 12:01:41 PM



775Noise, vibration, and harshness (NVH) in engine system design

response can be simulated by the boundary element method (BEM). Once

the prototype design is available, FEA is conducted to calculate the modal

and vibration responses of engine components and assemblies, and BEM is

used to predict the radiated noise levels.

In the product design and testing stage, the engine sound pressure level

(using SAE J1074), sound intensity, and sound power are measured for the

entire engine and certain individual sources (e.g., intake and exhaust noises)

with a bottom-up approach from the component level integrated up to the

system level. Sound quality metrics are also evaluated. Frequency response

function and transfer path analyses are conducted. Engine mount vibration

and torsional vibration are measured. The truck pass-by noise (SAE J366,

Rufnen et al., 1995) and the stationary noise (SAE J1096) are measured,

and the vehicle interior noise is evaluated.

In the detailed design and FEA analysis phase, the structural vibration

response is predicted in terms of natural frequency, mode shape, and

damping. The simulation accuracy of FEA depends on how accurately the

following are specied in the model: physical properties of the structural

material, boundary conditions of the restrained structure, the type and the

location of the dynamic forcing function, the number of modes, and the

frequency range of interest. The FEA forms mass and stiffness matrices

that are used to determine the eigenvalues and eigenvectors (or natural

frequency and mode shapes) of the system. The damping, along with the

natural frequency, determines the dynamic behavior of the system in terms

of amplitude, overshoot, and settling time. FEA can play a major role in the

system integration phase by predicting the overall engine noise after all the

component designs are proposed. The process of low-noise engine design

is discussed in detail by Beidl et al. (2001).

11.3.8 Design measures to reduce engine NVH

Design measures to reduce engine NVH generally fall into four major

categories: (1) reducing the strength of the excitation at the source by

engine design or operation measures; (2) muffling or silencing (for

aerodynamic noise); (3) reducing or isolating sound transmission path by

structural attenuation; and (4) noise insulation and encapsulation. Engine

system design focuses on the rst two categories, and they will be detailed

in later sections. The other two categories are usually component design

work. However, in order to predict the overall engine noise accurately, the

design characteristics in the latter two categories need to be understood by

the system engineer.

In structural attenuation, there are basically four methods to reduce the

noise radiated from the engine surfaces: (1) increasing the stiffness and

resonant frequency of the structure (e.g., adding ribs/ns or increasing wall

Diesel-Xin-11.indd 775 5/5/11 12:01:41 PM



776 Diesel engine system design

thickness); (2) reducing the surface area; (3) increasing the noise transmission

loss (e.g., using absorption or barrier materials for the airborne noise); and

(4) interrupting the noise transfer path (e.g., structural isolation, mounting,

damping by the use of mass dampers for the structure-borne noise). When

the structure stiffness is increased the structural resonances are shifted

toward higher frequencies where the amplitudes of the dynamic excitation

forces become smaller (Wodtke and Bathelt, 2004). The mechanism of noise

damping is to convert the sound energy to another type of energy, normally

heat. Typical examples of noise damping are using the parts made of foam

or porous materials. Light-weight engine designs have been used to increase

the engine power-to-weight ratio, but lowering the engine structure mass will

generally weaken structural attenuation and result in higher noise. Engine

design needs to reach the best compromise between engine performance and

vibro-acoustic comfort.

The mechanism of noise insulation is to block and reect the sound waves.

In addition to the engine block surface, a large amount of noise is radiated

from the accessories or shields that are attached to the basic structure, such

as the oil pan and the turbocharger. These attachments usually have light

structures and their noises are easier to control than for the basic engine

block. The attached components should be fastened at the locations having

low vibration or high stiffness in order to avoid high noise. Local sound

covers (e.g., valve cover, oil pan cover, geartrain cover) can reduce their

radiated noises effectively. Damping materials may be applied on the surface

of the covers or shields to further dissipate the sound. Lastly, total engine

encapsulation is an effective measure to reduce the engine noise. However,

total engine encapsulation is expensive and increases the engine weight and

size. It also requires special ventilation design in order to prevent the engine

surface from overheating.

Heavy-duty diesel engine NVH is summarized by Walker (1999). Light-

duty diesel engine NVH is briey introduced by Schulte (1999) and Wolf

et al. (2003). Reviews on the diesel engine noise are provided by Austen

and Priede (1959), Grover and Lalor (1973), Priede (1980), Hickling and

Kamal (1982), Challen and Croker (1982), Russell (1982), Yawata and

Crocker (1983), Haddad (1984), Cuschieri and Richards (1985), Farnell

and Riding (1999), and Gaikwad et al. (2007). The diesel engine structural

design guidelines for noise control are summarized by Priede et al. (1969),

Jenkins and Kuehner (1973), Anderton and Priede (1982), and Challan

(1982). Overall design strategies of low-noise diesel engines are elaborated

by Anderton (1984), Brandl et al. (1987) and Boesch (1987). Diesel engine

idle noise characteristics has been researched by Miura and Kojima (2003).

A good design example of a modern low-NVH diesel engine is provided

by Kwak et al. (2007). Diesel engine acoustic encapsulation technique is

presented by Nathak et al. (2007).

Diesel-Xin-11.indd 776 5/5/11 12:01:41 PM



777Noise, vibration, and harshness (NVH) in engine system design

11.3.9 The approach of diesel engine system design to

NVH

The term NVH has been widely used to address the noise and vibration

problems together. Indeed, noise and vibration are closely related. They are

both performance attributes of the engine. In fact, vibration is also an attribute

of structural durability because excessive vibration may cause structural

failures. The central theme of diesel engine system design is to conduct a

good performance-based design subject to durability constraints with the

focus at the system level rather than at the component level. Therefore, many

vibration problems at the component durability level are outside the scope

of engine system design (e.g., dynamic stress evaluation of the crankshaft

torsional damper, bending block, rocker cover, and oil pan). However, the

overall planning of designing a low-noise and low-vibration engine does

belong to the scope of engine system design.

The NVH attribute often conicts with other attributes such as packaging,

durability, cost, and other performance attributes. System-level optimization

is required to balance these attributes. Due to the extreme complexity of

NVH issues, experimental work is often needed on the actual parts in order

to make nal judgments. However, analytical methods can be effective in

narrowing down the specication candidates and in shortening the conrmation

cycle.

A ‘top-down’ system design is not equal to the ‘bottom-up’ summation of

all the component designs in NVH. At most the latter can be called ‘system

synthesis’ in a late validation stage of the design cycle. Dening the roles

and distinguishing between the system work and the component work in

the NVH area are important in order to avoid losing focus and unnecessary

repeating between the two. The system design work is characterized by the

following: (1) ownership of planning the system-level NVH loading and

its associated hardware and software; (2) predicting the system-level NVH

operating characteristics; and (3) system optimization. A system engineer

cares less about the local design solutions of a component unless that design

affects the entire system.

In general there are ve advanced simulation tools that are used in the

NVH area for virtual powertrain analysis, namely (1) gas wave dynamics; (2)

engine cycle simulation; (3) multi-body dynamics; (4) nite element analysis

(FEA); and (5) the boundary element method (BEM). The rst three tools

are particularly suitable for engine system NVH analysis. They are used by

a system engineer in his/her daily work to analyze the system behavior of

the gas ows and system vibrations, and he/she has the advantage to take

these tools a step further to apply them in the NVH area. The FEA and BEM

tools are usually very time-consuming and normally more suitable for the

component-level design analysis. In particular, the multi-body dynamics in

Diesel-Xin-11.indd 777 5/5/11 12:01:41 PM



778 Diesel engine system design

engine dynamic simulations includes valvetrain dynamics, piston-assembly

dynamics, cranktrain dynamics, engine balance, and geartrain dynamics.

Three levels of NVH models can be considered in diesel engine system

design to predict the excitation sources, vibrations, and even noises from

various levels of complexity. These models are as follows:

∑ Level-1 system noise model, which does not have crank-angle

resolution.

∑ Level-2 system NVH model, which has instantaneous crank-angle

resolution for all the excitation sources. It does not have to be real-time

capable but needs to be computationally fast for system optimization.

This model is the primary approach in diesel engine system design for

NVH. A real-time capable level-2 model can be used for real-time engine

controls related to noise issues such as fault diagnostics and active noise

control measures.

∑ Level-3 system NVH model, which involves FEA or BEM type of

simulation, and is computationally expensive and cannot be real-time

capable. It predicts structural attenuation (or transfer function for the

noise) and noise radiation. The output of the model includes the detailed

velocity distribution on engine surfaces and the noise emitted. The level-3

model can be used in the system synthesis analysis stage with a detailed

simulation.

A combination of analytical and empirical approaches is required in the

modeling. These models will be detailed in Section 11.12.

11.4 Combustion noise

Combustion noise is an important concern in combustion system design and

performance calibration. Combustion is the primary source of noise in most

naturally aspirated direct injection diesel engines. Although the combustion

noise in turbocharged diesel engines is not dominant at high-speed and high-

load steady-state conditions, it can become dominant at idle, light-load, or

acceleration conditions.

Combustion noise is transferred to the top deck of the cylinder head and

the cylinder wall. It is also transferred to the piston, the connecting rod,

the crankshaft, the main bearings, and the crankcase. Combustion noise is

related to the rate of rise of the in-cylinder pressure (dp/df) in the premixed

combustion phase and also related to the lasting duration of the high rate of

rise, which is directly governed by the duration of ignition delay. It is believed

that the sharply rising cylinder pressure induces highly transient oscillatory

shock waves and excites the cylinder block structure to vibrate and emit noise.

Moreover, the gas waves during the heterogeneous combustion process are

reected by the cylinder wall to form highly oscillatory gas waves to produce

Diesel-Xin-11.indd 778 5/5/11 12:01:41 PM



779Noise, vibration, and harshness (NVH) in engine system design

noise. Because the natural frequencies of the cylinder structural components

are mostly at the medium or high frequency, the structure is excited by the

combustion to generate noise in the medium- to high-frequency range. The

noise is perceived by the ear as impulsive and unpleasant.

Combustion noise is sensitive to the rate of rise of the cylinder pressure.

A small change in the cylinder pressure trace may not affect the output

power signicantly, but it may affect the combustion noise appreciably. A

frequency-domain analysis of the in-cylinder pressure trace is necessary

in order to analyze the sensitivity and identify appropriate solutions to

control the combustion noise. The combustion noise is affected mainly by

the cylinder pressure level (in dB) in the medium-frequency range (e.g.,

100–1000 Hz) that is related to dp/df, and also in the high-frequency range

(e.g., higher than 1000 Hz) that is related to d

p/df

. In the low-frequency

range the pressure levels exhibit an oscillating and decaying pattern that

reects the ring frequency of the engine and its multiple harmonics, just

like a typical periodic forcing function would act. The continuous monotonic

pattern in the medium- to high-frequency range indicates a rising slope of

the cylinder pressure in the time domain. A larger dp/df gives a atter slope

of the pressure level (in dB) in the frequency domain, resulting in higher

combustion noise.

Tung and Crocker (1982) discovered that for turbocharged diesel engines

the frequency content of the combustion gas pressure up to about 300 Hz is

related to the peak cylinder pressure. The frequencies between 300 Hz and

2000 Hz are related to the maximum rate of rise of the cylinder pressure.

The frequencies above 2000 Hz are related to both the amplitude and the

duration of the second-order pressure derivative. They also found that the

frequencies of the cylinder pressure uctuations are closely related to the

cavity resonance frequencies in the combustion chamber.

Because the resonant frequencies of many components in the diesel

engine are in the medium- and high-frequency range (e.g., above 1000

Hz), it is relatively easier to excite the structure vibration and noise in the

medium- to high-frequency range than in the low-frequency range. Therefore,

the large amplitude of the cylinder pressure in the low-frequency range

usually does not translate to high vibration and noise. The term ‘structural

attenuation’ is dened as the difference between the cylinder pressure in dB

and the sound pressure level emitted from the engine surface. Inherently, the

engine has high (strong) structural attenuation in the low-frequency range

and low attenuation in the high-frequency range. When the structure stiffness

of the cylinder bore is very high the resonant frequency of the structure

becomes very high. If the resonant frequency is higher than the frequency

of the combustion excitation, it can effectively attenuate the high-frequency

combustion noise.

Diesel knocking or clatter noise is an impulsive noise. It is most signicantly

Diesel-Xin-11.indd 779 5/5/11 12:01:41 PM

