United States Office of Mobil* Source Air Pollution Control EPA 460/3-90-001
Environmental Protection Emission Control Technology Division February 1990
Ager.cy 2565 Plymouth Road
Ann Arbor, Michigan 48105
Air
o EPA Emissions Control Strategies for
Heavy-Duty Diesel Engines
-------
EPA 460/3-90-001
Emissions Control Strategies
for
Heavy-Duty Diesel Engines
by
Christopher S. Weaver
Sierra Research, Inc.
1521 I Street
Sacramento, CA 95814
Contract No. 68-C8-0024
Work Assignment No. 2
EPA Project Officer: Thomas M. Baines
prepared for:
Environmental Protection Agency
Office of Mobile Source Air Pollution Control
Emission Control Technology Division
2565 Plymouth Road
Ann Arbor, Michigan 48105
February 1990
-------
This report is issued by the Environmental Protection Agency to report
technical data of interest to a limited number of readers. Copies are
available free of charge to Federal employees, current contractors and
grantees, and nonprofit organizations - in limited quantities - from
the Library Services Office, Environmental Protection Agency, 2565
Plymouth Road, Ann Arbor, Michigan 48105.
This report was furnished to the Environmental Protection Agency by
Sierra Research, Inc., 1521 I Street, .Sacramento, California, in
fulfillment of Contract No. 68-C8-0024, Work Assignment No. 2. The
contents of this report are reproduced herein as received from Sierra
Research, Inc. The opinions, findings, and conclusions expressed are
those of the author and not necessarily those of the Environmental
Protection Agency. Mention of company product names is not to be
considered as an endorsement by the Environmental Protection Agency.
Publication No. 460/3-90-001
-ii-
-------
ABSTRACT
This report presents basic information on Diesel engine technology,
emissions, and emission controls, and describes a number of options
for reducing emissions from both new vehicles and those which are
already in use. Pollutant formation is determined by the Diesel
combustion process. Factors which affect pollutant emission rates
include the air-fuel ratio in the cylinder, charge temperature and
composition, and the fuel injection characteristics and timing.
Emission rates vary greatly between different operating conditions. As
a result, the choice of a representative test cycle is critical to
establishment of effective emissions regulations. For trucks and
other highway vehicles, the emissions test cycle should include a wide
range of operating conditions, especially transients, in order to
represent real-life operating conditions adequately.
Many different levels of Diesel emissions control are possible,
ranging from simple controls to reduce excessive smoke to the advanced
technologies required to meet U.S. emissions standards in the
mid-'90s. Technology cost and complexity tend to increase with
increasing effectiveness of control, so that different emission
control levels may be suited to different areas at different times.
Simple and inexpensive emission controls can reduce Diesel NOx and
particulate emissions from new vehicles by one-third to nearly one-
half, at a cost of $100 to $200 per vehicle. More sophisticated
engine controls can further reduce emissions, to about 40% of the
uncontrolled level for NOx and 20% of the uncontrolled level for
particulate matter. The cost of these controls is significant,
however - ranging from about $800 to $2,000 per vehicle for heavy
trucks. Still further control of particulate emissions can be
achieved through the addition of aftertreatment devices such as
catalytic converters or catalytic trap-oxidizers, at still higher
cost.
Options for controlling emissions from existing vehicles do not follow
such a neat hierarchy. Depending on the situation, maintenance, fuel,
modifications, transportation controls, and other measures could all
play a role in an integrated emission control strategy. Inspection
and maintenance programs may be useful in themselves, and may also be
required to realize the full benefits of new-vehicle emission
controls.
Although Diesels produce many different types of pollutants, one of
the most significant (as well as visually obvious) pollutants is soot
and other particulate matter. One effective measure for eliminating
these emissions is to substitute an alternative, non-soot-producing
fuel for Diesel fuel. Natural gas and methanol show particular
promise in this regard. This may require substituting another type of
engine as well. Such substitution can create other pollution
problems, however, depending on the fuel and technology employed-- the
fact that emissions are not visible does not necessarily make them
less objectionable.
- iii-
-------
TABLE OF CONTENTS
1.0 Introduction 1
2.0 Background: Factors Affecting Emissions 2
2.1 Diesel Combustion and Pollutant Formation 2
2.2 Influence of Engine Variables 7
2.3 Emissions Testing and Measurement 12
2.4 Fuel Effects 19
3.0 Diesel Engine Technology 23
3.1 Air Motion and Combustion Chamber Design 23
3.2 Fuel Injection 27
3.3 Engine Controls 31
3.4 Turbocharging and Intercooling 33
3.5 Exhaust Gas Recirculation 35
3.6 Lubricating Oil Control 36
3.7 Aftertreatment Systems 37
4.0 Emission Control Strategies for New Diesels 44
4.1 Smoke Controls 45
4.2 First-Level Emissions Control 49
4.3 Existing Technology 52
4.4 Near-Term Technology (U.S. 1991 Standards) 56
4.5 Most Stringent Non-Trap Technology 60
4.6 Maximum Emissions Control: Catalytic Trap-Oxidizers ... 62
4.7 Advanced Technologies 65
5.0 Emission Control Strategies for Diesels Already In Use 68
5.1 Maintenance 68
5.2 Smoke Enforcement 72
5.3 Inspection/Maintenance Programs 75
5.4 Fuel Modification 79
5.5 Retrofitting Emission Controls 82
5.6 Transportation Control Measures 85
6.0 Alternative Fuels 87
6.1 Natural Gas 87
6.2 Liquified Petroleum Gas (LPG) 95
6.3 Methanol 97
6.4 Ethanol 103
7.0 References 105
- iv-
-------
LIST OF FIGURES
Figure
1 Correlation of NOx Emissions Index with Reciprocal
Flame Temperature 5
2 Typical Variation of Emissions with Air-Fuel Ratio
and Load in a Direct-Injection Diesel Engine 8
3 Constant Volume Sampling System for Diesel Emissions
Measurement 13
4 Different Types of Diesel Combustion Chambers 24
5 Typical Diesel Fuel Injection Systems 29
6 Principle of the Ceramic Monolith Trap 38
7 Estimated Impact of Poor Maintenance and Tampering
with Emission Controls on Heavy-Duty Diesel Emissions
in California 1985-2000 73
8 Typical Variation of Emission Levels with Air-Fuel
Ratio X for an Otto-Cycle Engine 91
v-
-------
LIST OF TABLES
Table
1 Effect of Engine Control Strategy - Transient vs.
13-Mode Emissions Test 17
2 Emission Reductions Due to Emission Controls - Engine
Transient vs. Chassis Tests 18
3 Emissions Control Levels Considered 44
4 Estimated Emissions and Cost Impacts of Emissions
Control Level One: Smoke Controls 46
5 Estimated Emissions and Cost Impacts of Emissions
Control Level Two: First Level Controls 50
6 Estimated Emissions and Cost Impacts of Emissions
Control Level Three: Existing Control Technology 53
7 Estimated Emissions and Cost Impacts of
Emissions Control Level Four: Best In-Cylinder
Control Technology 56
8 Estimated Emissions and Cost Impacts of Emissions
Control Level Five: Best Non-Trap Control Technology 61
9 Estimated Emissions and Cost Impacts of Emissions
Control Level Six: Maximum Emissions Control 63
10 Estimated Effects of Tampering and Malfunctions
on Heavy-Duty Diesel Emissions 70
11 Properties of Alternative and Conventional Fuels 88 .
-vi-
-------
FOREWORD
This report was originally drafted in Fall, 1988, as a lengthy chapter
for use in a report on heavy-duty Diesel emissions control issued by
the Organization for Economic Cooperation and Development (OECD). It
was drafted by C.S. Weaver of Sierra Research, under contract to the
U.S. EPA. Thomas Baines was the EPA Project Manager, and his comments
and direction contributed significantly to the results. The OECD
report is presently under review. At EPA's request, meanwhile, the
report has been redrafted as a stand-alone document. This redrafting
involved primarily editorial changes such as section numbering, etc.
With a few exceptions, the technical information was not updated
during the redrafting, and is therefore current as of Fall, 1988.
-vii-
-------
1. INTRODUCTION
This report presents basic information on Diesel engine technology,
emissions, and emission controls, and describes a number of options
for reducing emissions from both new vehicles and those which are
already in use. Section 2 below presents background material on the
factors that affect Diesel emissions, while Section 3 summarizes the
key emissions-related technological features of heavy-duty Diesel
engines. These sections provide necessary background for the
remaining discussion. Options for reducing emissions from new
vehicles are discussed in Section 4; those for existing vehicles in
Section 5. Finally, Section 6 describes alternative fuel options for
both new and existing heavy-duty Diesel engines.
Many different levels of Diesel emissions control are possible,
ranging from simple controls to reduce excessive smoke to the advanced
technologies required to meet U.S. emissions standards in the
mid-'90s. Technology cost and complexity tend to increase with
increasing effectiveness of control, so that different emission
control levels may be suited to different areas at different times.
The emission-control options for new vehicles discussed in Section 4
constitute a graduated sequence of control levels, beginning with the
simplest and least stringent and working up to the most complex,
expensive, and effective levels of technology.
Options for controlling emissions from existing vehicles do not follow
such a neat hierarchy. Depending on the situation, maintenance, fuel
modifications, transportation controls, and other measures could all
play a role in an integrated emission control strategy. Inspection
and maintenance programs may be useful in themselves, and may also be
required to realize the full benefits of new-vehicle emission
controls. Section 5 outlines the potential for reducing emissions
through these measures, as well as the interrelationships between
them, and between them and the various levels of new-vehicle emissions
control.
Although Diesels produce many different types of pollutants, one of
the most significant (as well as visually obvious) pollutants is soot
and other particulate matter. One effective measure for eliminating
these emissions is to substitute an alternative, non-soot-producing
fuel for Diesel fuel. This may require substituting another type of
engine as well. Such substitution can create other pollution
problems, however, depending on the fuel and technology employed--the
fact that emissions are not visible does not necessarily make them
less objectionable. Section 6 discusses the major alternative fuels
available, technologies to utilize them, and the pollution tradeoffs
involved.
-1-
-------
2. BACKGROUND: FACTORS AFFECTING EMISSIONS
In order to discuss the means available for reducing Diesel pollutant
emissions, it is necessary first to understand their sources and the
various engine and fuel-related factors that can affect their
formation and emission. This section presents background information
on the types of pollutants emitted by Diesel engines, their formation
mechanisms, emissions testing and measurement procedures, and the
effects of different engine and fuel variables on emissions levels.
2.1 Diesel Combustion and Pollutant Formation
Combustion
Diesel engine emissions are determined by the combustion process.
This process is central to the operation of the Diesel engine. As
opposed to Otto-cycle engines (which use a more-or-less homogeneous
charge) all Diesel engines rely on heterogeneous combustion. During
the compression stroke, a Diesel engine compresses only air. The
process of compression heats the air to about 700 to 900° C, which is
well above the self- ignition temperature of Diesel fuel. Near the top
of the compression stroke, liquid fuel is injected into the combustion
chamber under tremendous pressure, through a number of small orifices
in the tip of the injection nozzle. The quantity of fuel injected
with each stroke determines the engine power output.
The high-pressure injection atomizes the fuel. As the atomized fuel
is injected into the chamber, the periphery of each jet mixes with the
hot air already present. After a brief period known as the ignition .
delay, this fuel-air mixture ignites. In the premlxed burning phase.
the fuel/air mixture formed during the ignition delay period bums
very rapidly, causing a rapid rise in cylinder pressure. The
subsequent rate of burning is controlled by the rate of mixing between
the remaining fuel and air, with combustion always occurring at the
interface between the two. Most of the fuel injected is burned in
this diffusion burning phase, except under very light loads.
A mixture of fuel and exactly as much air as is required to burn the
fuel completely is called a "stoichiometric mixture". The air-fuel
ratio A is defined as the ratio of the actual amount of air present
per unit of fuel to the stoichiometric amount. In Diesel engines, the
fact that fuel and air must mix before burning means that a
substantial amount of excess air is needed to ensure complete
combustion of the fuel within the limited time allowed by the power
-2-
-------
stroke. Diesel engines, therefore, always operate with overall air-
fuel ratios which are considerably lean of stoichiometric (X greater
than one).
The air-fuel ratio in the cylinder during a given combustion cycle is
determined by the engine power requirements, which govern the amount
of fuel injected. Diesels operate without throttling, so that the
amount of air present in the cylinder is essentially independent of
power output, except in turbocharged engines. The minimum air-fuel
ratio for complete combustion is about equal to 1.5. This ratio is
known as the smoke limit, since smoke increases dramatically at air-
fuel ratios lower than this. The smoke limit establishes the maximum
amount of fuel that can be burned per stroke, and thus the maximum
power output of the engine.
Pollutant formation
The principal pollutants emitted by Diesel engines are oxides of
nitrogen (NOx), sulfur oxides (SOx), particulate matter (PM), and
unburned hydrocarbons (HC). Diesels are also responsible for a small
amount of CO, as well as visible smoke, unpleasant odors, and noise.
In addition, like all engines using hydrocarbon fuel, Diesels emit
significant amounts of C02, which has been implicated in the so-called
"greenhouse effect." With thermal efficiencies typically in excess of
40%, however, Diesels are the most fuel-efficient of all common types
of combustion engines. As a result, they emit less C02 to the
atmosphere than any other type of engine doing the same work.
The NOx, HC, and most of the particulate emissions from Diesels are
formed during the combustion process, and can be controlled by
appropriate modifications to that process, as can most of the
unregulated pollutants. The sulfur oxides, in contrast, are derived
directly from sulfur in the fuel, and the only feasible control
technology is to reduce fuel sulfur content. Most SOx is emitted as
gaseous S02, but a small fraction (typically 2-4 percent) occurs in
the form of particulate sulfates.
Diesel particulate matter consists mostly of three components: soot
formed during combustion, heavy hydrocarbons condensed or adsorbed on
the soot, and sulfates. In older Diesels, soot was typically 40 to 80
percent of the total particulate mass. Developments in in-cylinder
emissions control have reduced the soot contribution to particulate
emissions from modem emission-controlled engines considerably,
however. Most of the remaining particulate mass consists of heavy
hydrocarbons adsorbed or condensed on the soot. This is referred to
as the soluble organic fraction of the particulate matter, or SOF.
The SOF is derived partly from the lubricating oil, partly from
unburned fuel, and partly from compounds formed during combustion.
The relative importance of each of these sources varies from engine to
engine.
In-cylinder emission control techniques have been most effective in
reducing the soot and fuel-derived SOF components of the particulate
-3-
-------
matter. As a result, the relative importance of the lube oil and
sulfate components has increased. In the emission-controlled engines
under development, the lubricating oil accounts for as much as 40
percent of the particulate matter, and the sulfates may account for
another 25 percent. Lube oil emissions can be reduced by reducing oil
consumption, but this may adversely affect engine durability. The
only known way to reduce sulfate emissions is to reduce the sulfur
content of Diesel fuel.
The gaseous hydrocarbons and the SOF component of the particulate
matter emitted by Diesel engines include many known or suspected
carcinogens and other toxic air contaminants. These include
polynuclear aromatic compounds (PNA) and nitro-PNA, formaldehyde and
other aldehydes, and other oxygenated hydrocarbons. The oxygenated
hydrocarbons are also responsible for much of the characteristic
Diesel odor.
Oxides of nitrogen--NOx from Diesels is primarily NO. This gas forms
from nitrogen and free oxygen at high temperatures close to the flame
front. The rate of NO formation in Diesels is a function of oxygen
availability, and is exponentially dependent on the flame temperature.
Figure 1 shows the experimentally-derived relationship between flame
temperature and NOx emissions. In the diffusion burning phase, flame
temperature depends only on the heating value of the fuel, the heat
capacity of the reaction products and any inert gases present, and the
starting temperature of the initial mixture. In the premixed burning
stage, the local fuel-air ratio also affects the flame temperature,
but this ratio varies from place to place in the cylinder and is very
hard to control.
In Diesel engines, most of the NOx emitted is formed early in the
combustion process, when the piston is still near top-dead-center
(TDC). This is when the temperature and pressure of the charge are
greatest. Recent work by several researchers (Wade et al., 1987;
Cartellieri and Wachter, 1987) indicates that most NOx is actually
formed during the premixed burning phase. It has been found that
reducing the amount of fuel burned in this phase can significantly
reduce NOx emissions.
NOx can also be reduced by actions which reduce the flame temperature
during combustion. These actions include delaying combustion past
TDC, cooling the air charge going into the cylinder, reducing the air-
fuel mixing rate near TDC, and exhaust gas recirculation (EGR). Since
combustion always occurs under near-stoichiometric conditions,
reducing the flame temperature by "lean-burn" techniques, as in spark-
ignition engines, is impractical.
Particulate matter--Diesel soot is formed only during the diffusion
burning phase of combustion. Primary soot particles are small spheres
of graphitic carbon, approximately 0.01 fim in diameter. These are
formed by the rapid polymerization of acetylene at moderately high
temperatures under oxygen-deficient conditions. The primary particles
then agglomerate to form chains and clusters of linked particles,
giving the soot its characteristic fluffy appearance. During the
-4-
-------
Figure 1
Correlation of NOx Emissions Index with Reciprocal
Flame Temperature (Source: Plee et al., 1983)
= -34. 300
DI engine
0-> addition N-> addition
- • Present study
a Yu and Shahed
0.01
3.4 3.6 3 8 4.0 4.2 4 4 4.6
104/K
Tf
diffusion burning phase, the local gas composition at the flame front
is close to stoichiometric, with an oxygen-rich region on one side and
a fuel-rich region on the other. The moderately high temperatures and
excess fuel required for soot formation are thus always present.
Most of the soot formed during combustion is subsequently burned
during the later portions of the expansion stroke. Typically, less
than 10% of the soot formed in the cylinder survives to be emitted
into the atmosphere. Soot oxidation is much slower than soot
formation, however, and the amount of soot oxidized is heavily
dependent on the availability of high temperatures and adequate oxygen
during the later stages of combustion. Conditions which reduce the
availability of oxygen (such as poor mixing, or operation at low air-
fuel ratios), or which reduce the time available for soot oxidation
(such as retarding the combustion timing) tend to increase soot
emissions.
The SOF component of Diesel particulate matter consists of heavy
hydrocarbons condensed or adsorbed on the soot. A. significant part of
-5-
-------
this material is unburned lubricating oil, which is vaporized from the
cylinder walls by the hot gases during the power stroke. Some of the
heavier hydrocarbons in the fuel may also come through unburned, and
condense on the soot particles. Finally, heavier hydrocarbons may be
synthesized during combustion, possibly by the same types of processes
which produce soot. Pyrosynthesis of polynuclear aromatic
hydrocarbons during Diesel combustion has been demonstrated by
Williams et al. (1987).
Hydrocarbons --Diesel HC emissions (as well as the unburned-fuel
portions of the particulate SOF) occur primarily at light loads in
most engines. They are due to excessive fuel-air mixing, which
results in some volumes of air-fuel mixture which are too lean to
burn. Other HC sources include fuel deposited on the combustion
chamber walls or in combustion chamber crevices by the injection
process; fuel retained in the orifices of the injector which vaporizes
late in the combustion cycle; and partly reacted mixture which is
subjected to bulk quenching by too-rapid mixing with air. Aldehydes
(as partially-reacted hydrocarbons) and the small amount of CO
produced by Diesels are probably formed in the same processes as the
HC emissions.
The presence of polynuclear aromatic hydrocarbons and their nitro-
derivatives in Diesel exhaust is of special concern, since these
compounds include many known mutagens and suspected carcinogens. A
significant portion of these compounds (especially the smaller two and
three-ring compounds) are apparently derived directly from the fuel.
Typical Diesel fuel contains several percent PNA by volume. Most of
the larger and more dangerous PNA's, on the other hand, appear to form
during the combustion process, possibly via the same acetylene
polymerization reaction that produces soot (Williams et al., 1987).
Indeed, the soot particle itself can be viewed as essentially a very
large PNA molecule.
Visible Smoke--Visible smoke is due primarily to the soot component of
Diesel particulate matter. Under most operating conditions, the
exhaust plume from a properly adjusted Diesel engine is normally
invisible, with a total opacity (absorbtance and reflectance) of two
percent or less. Visible smoke emissions from heavy-duty Diesels are
generally due to operating at air-fuel ratios at or below the smoke
limit, or to poor fuel-air mixing in the cylinder. These conditions
can be prevented by proper design. The particulate reductions
required to comply with the U.S. 1991 emissions standards are expected
to essentially eliminate visible smoke emissions from properly
functioning engines.
Noise--Diesel engine noise is due principally to the rapid combustion
(and resulting rapid pressure rise) in the cylinder during the
premixed burning phase. The greater the ignition delay, and the more
fuel is premixed with the air, the greater this pressure rise and the
resulting noise emissions will be. Noise emissions and NOx emissions
thus tend to be related--reducing the amound of fuel burned in the
premixed burning phase will tend to reduce both. Other noise sources
include those common to all engines, such as mechanical vibration, fan
-6-
-------
noise, and so forth. These can be minimized by appropriate mechanical
design.
Odor--The characteristic Diesel odor is believed to be due primarily
to partially-oxygenated hydrocarbons (aldehydes and similar species)
in the exhaust. These are believed to be due primarily to slow
oxidation reactions in volumes of air-fuel mixture too lean to bum
normally. Unburned aromatic hydrocarbons may also play a significant
role. The most significant aldehyde species are benzaldehyde,
acetaldehyde, and formaldehyde, but other aldehydes such as acrolein
(a powerful irritant) are significant as well. Aldehyde and odor
emissions are closely linked to total HC emissions--experience has
shown that modifications which reduce total HC tend to reduce
aldehydes and odor as well.
2.2 Influence of engine variables
The engine variables having the greatest effects on Diesel emission
rates are the air-fuel ratio, rate of air-fuel mixing, fuel injection
timing, compression ratio, and the temperature and composition of the
charge in the cylinder. Most techniques for in-cylinder emission
control involve manipulating one or more of these variables.
Air-fuel ratio
The ratio of air to fuel in the combustion chamber has an extremely
important effect on emission rates for hydrocarbons and particulate
matter. Figure 2 diagrams the typical relationship between air-fuel
ratio X and emissions in a Diesel engine. As discussed above, the
power output of the engine is determined by the amount of fuel
injected at the beginning of each power stroke. At very high air-fuel
ratios (corresponding to very light load), the temperature in the
cylinder after combustion is too low to burn out residual
hydrocarbons, so emissions of gaseous HC and particulate SOF are high.
At lower air-fuel ratios, less oxygen is available for soot oxidation,
so soot emissions increase. As long as A remains above about 1.6,
this increase is relatively gradual. Soot and visible smoke emissions
show a strong non-linear increase below the smoke limit at about X -
1.5, however.
In naturally-aspirated engines (those without a turbocharger), the
amount of air in the cylinder is independent of the power output.
Maximum power output for these engines is normally "smoke-limited"--
that is, limited by the amount of fuel that can be injected without
exceeding the smoke limit. Maximum fuel settings on these engines
represent a compromise between smoke emissions and power output.
Where Diesel smoke is regulated (as in the U.S. and EEC), this
compromise must result in smoke opacity below the regulated limit.
Otherwise, opacity is limited by the manufacturer's judgement of
commercially acceptable smoke emissions.
-7-
-------
Figure 2
Typical Variation of Emissions with Air-Fuel Ratio and Load
•in a Direct-Injection Diesel Engine
cr
X
d
4
a
CO
z
Q
CO
•' \
A \
« \
/ /
* /
» /
* y
•
PM
—
1 1 !
2 4 8
AIR-FUEL RATIO
- INCREASING ENGINE LOAD -
16
In turbocharged engines, increasing the fuel injected per stroke
increases the energy in the exhaust gas, causing the turbocharger to
spin more rapidly and pump more air into the combustion chamber. For
this reason, power output from turbocharged engines is not usually
smoke-limited. Instead, it is limited by design limits on variables
such as turbocharger speed and engine mechanical stresses.
Turbocharged engines do not normally experience low air-fuel ratios
during steady-state operation. Low air-fuel ratios can occur during
transient accelerations, since the inertia of the turbocharger rotor
keeps it from responding instantly to an increase in fuel input.
Thus, the air supply during the first few seconds of a full-power
acceleration is less than the air supply in steady-state operation.
To overcome this problem, turbocharged engines in highway vehicles
commonly incorporate an acceleration smoke limiter. This device
limits the fuel flow to the engine until the turbocharger has time to
respond. The setting on this device must compromise between
acceleration performance (drivability) and low smoke emissions. In
the U.S., acceleration smoke emissions are limited by regulation;
elsewhere, they are limited by the manufacturer's judgement of
commercial acceptability.
-8-
-------
Air-fuel mixing
The rate of mixing between the compressed charge in the cylinder and
the injected fuel is among the most important factors in determining
Diesel performance and emissions. During the ignition delay period,
the mixing rate determines the amount of fuel that mixes with the air,
and is thus available for combustion during the premixed burning
phase. The higher the mixing rate, the greater the amount of fuel
burning in premixed mode, and the higher the noise and NOx emissions
will tend to be.
In the diffusion burning phase, the rate of combustion is limited by
the rate at which air and fuel can mix. The more rapid and complete
this mixing, the greater the amount of fuel that burns near piston
top-dead-center, the higher the fuel efficiency, and the lower the
soot emissions. Too-rapid mixing, however, can increase hydrocarbon
emissions--especially at light loads--as small volumes of air-fuel
mixture are diluted below the combustible level before they have a
chance to bum. Unnecessarily intense mixing also dissipates energy
through turbulence, increasing fuel consumption.
In engine design practice, it is necessary to strike a balance between
the rapid and complete mixing required for low soot emissions and best
fuel economy, and too-rapid mixing leading to high NOx and HC
emissions. The primary factors affecting the mixing rate are the fuel
injection pressure, the number and size of injection orifices, any
swirling motion imparted to the air as it enters the cylinder during
the intake stroke, and air motions generated by combustion chamber
geometry during compression. Much of the progress in in-cylinder
emissions control over the last decade has come through improved
understanding of the interactions between these different variables
and emissions, leading to improved designs.
Air-fuel mixing rates in present emission-controlled engines are the
product of extensive optimization to assure rapid and complete mixing
under nearly all operating conditions. Poor mixing may still occur
during "lug-down"--high-torque operation at low engine speeds.
Turbocharger boost, air swirl level, and fuel injection pressure are
typically poorer in these "off-design" conditions. Maintenance
problems such as injector tip deposits can also degrade air-fuel
mixing, and result in greatly increased emissions.
Injection timing
The timing relationship between the beginning of fuel injection and
the top of the compression stroke has an important effect on Diesel
engine emissions and fuel economy. For best fuel economy, it is
preferable that combustion begin at or somewhat before top-dead-
center. Since there is a finite delay between the beginning of
injection and the start of combustion, it is necessary to inject the
fuel somewhat before this point (generally 5 to 15 degrees of
crankshaft rotation before).
-9-
-------
Since fuel is injected before Che piston reaches top-dead-center, the
charge temperature is still increasing as the charge is compressed.
The earlier fuel is injected, the cooler the charge will be, and the
longer the ignition delay. The longer ignition delay provides more
time for air and fuel to mix, increasing the amount of fuel that burns
in the premixed combustion phase. In addition, more fuel burning at
or just before top dead center increases the maximum temperature and
pressure attained in the cylinder. Both of these effects tend to
increase NOx emissions.
On the other hand, earlier injection timing tends to reduce
particulate and light-load HC emissions. Fuel burning in premixed
combustion forms little soot, while the soot formed in diffusion
combustion near TDC experiences a relatively long period of high
temperatures and intense mixing, and is thus mostly oxidized. The
end-of-injection timing also has a major effect upon soot emissions--
fuel injected more than a few degrees after TDC burns more slowly, and
at a lower temperature, so that less of the resulting soot has time to
oxidize during the power stroke. For a fixed injection pressure,
orifice size, and fuel quantity, the end-of-injection timing is
determined by the timing of the beginning of injection.
The result of these effects is that injection timing must compromise
between PM emissions and fuel economy on the one hand and noise, NOx
emissions, and maximum cylinder pressure on the other. The terms of
the compromise can be improved to a considerable extent by increasing
injection pressure, which increases mixing and advances the end-of-
injection timing. Another approach under development is split
iniection. in which a small amount of fuel is injected early in order
to ignite the main fuel quantity which is injected near TDC.
Compared to uncontrolled engines, modem emission-controlled engines
generally exhibit moderately retarded timing to reduce NOx, in
conjunction with high injection pressures to limit the effects of
retarded timing of PM emissions and fuel economy. The response of
fuel economy and PM emissions to retarded timing is not linear--up to
a point, the effects are relatively small, but beyond that point
deterioration is rapid. Great precision in injection timing is
necessary--a change of one degree crank angle can have a significant
impact on emissions. The optimal injection timing is a complex
function of engine design, engine speed and load, and the relative
stringency of emissions standards for the different pollutants. To
attain the required flexibility and precision of injection timing has
posed a major challenge to fuel injection manufacturers.
Compression ratio
Diesel engines rely on compression heating to ignite the fuel, so the
engine's compression ratio has an important effect on combustion. A
higher compression ratio results in a higher temperature for the
compressed charge, and thus in a shorter ignition delay and higher
flame temperature. The effect of a shorter ignition delay is to
reduce NOx emissions, while the higher flame temperature would be
-10-
-------
expected Co increase them. In practice, these two effects nearly
cancel, so that changes in compression ratio have little effect on
NOx.
Emissions of gaseous HC and of the SOF fraction of the particulate
matter are reduced at higher compression ratios, as the higher
cylinder temperature increases the burn-out of hydrocarbons. Soot
emissions may increase at higher compression ratios, however. Since
the higher compression is achieved by reducing the volume of the
combustion chamber, this results in a larger fraction of the air
charge being sequestered in "crevice volumes" such as the top and
sides of the piston, where it is not available for combustion early in
the power stroke. Thus, the effective air-fuel ratio in the
combustion chamber decreases, and soot emissions go up. This effect
can be limited (and overall air utilization and power output improved)
by reducing crevice volumes to the maximum extent possible.
Engine fuel economy, cold starting, and maximum cylinder pressures are
also affected by the compression ratio. For an idealized Diesel
cycle, the thermodynamic efficiency is an increasing function of
compression ratio. In a real engine, however, the increased
thermodynamic efficiency is offset after some point by increasing
friction, so that a point of maximum efficiency is reached. With most
heavy-duty Diesel engine designs, this optimal compression ratio is
about 12 to 15. To ensure adequate starting ability under cold
conditions, however, most Diesel engine designs require a somewhat
higher compression ratio--in the range of 15 to 20 or more.
Generally, higher-speed engines with smaller cylinders require higher
compression ratios for adequate cold starting.
Charge temperature
Reducing the temperature of the air charge going into the cylinder has
benefits for both PM and NOx emissions. Reducing the charge
temperature directly reduces the flame temperature during combustion,
and thus helps to reduce NOx emissions. In addition, the colder air •
is denser, so that (at the same pressure) a greater mass of air can be
contained in the same fixed cylinder volume. This increases the air-
fuel ratio in the cylinder and thus helps to reduce soot emissions.
By increasing the air available while decreasing piston temperatures,
charge-air cooling can also make possible a significant increase in
power output. Excessively cold charge air can reduce the burnout of
hydrocarbons, and thus increase light-load HC emissions, however.
This can be counteracted by advancing injection timing, or by reducing
charge air cooling at light loads.
Charge composition
NOx emissions are heavily dependent on flame temperature. By altering
the composition of the air charge to increase its specific heat and
the concentration of inert gases, it is possible to decrease the flame
temperature significantly. The most common way of accomplishing this
-11-
-------
is through exhaust gas recirculation (EGR). At moderate loads, EGR
has been shown to be capable of reducing NOx emissions by a factor of
two or more, with little effect on particulate emissions. Although
soot emissions are increased by the reduced oxygen concentration,
particulate SOF and gaseous HC emissions are reduced, due to the
higher in-cylinder temperature caused by the hot exhaust gas. EGR
cannot be used at high loads, however, since the displacement of air
by exhaust gas would result in an air-fuel ratio below the smoke
limit--and thus very high soot and PM emissions.
Emissions Tradeoffs
It is apparent from the foregoing discussion that there is an inherent
conflict between some of the most powerful Diesel NOx control
techniques and particulate emissions. This is the basis for the much-
discussed "tradeoff" relationship between Diesel NOx and particulate
emissions. This "tradeoff" is not absolute--various NOx control
techniques have varying effects on soot and HC emissions, and the
importance of these effects varies as a function of engine speed and
load. These tradeoffs do place limits on the extent to which any one
of these pollutants can be reduced, however. To minimize emissions of
all three pollutants simultaneously requires careful optimization of
the fuel injection, fuel-air mixing, and combustion processes over the
full range of engine operating conditions.
2.3 Emissions Testing and Measurement
The pollutant emissions measured for a given Diesel engine generally
depend greatly on how the emissions are measured. Emission rates for
Diesels vary as functions of speed, load, and other operating
conditions, and transient emissions may be different from those
measured in steady-state. Any meaningful discussion of achievable
emissions standards and/or emissions levels in use must therefore
consider the emissions testing and measurement methods used.
Preferably, the engine operating conditions during the test should
accurately reflect those in the real world.
This section describes the emissions measurement and testing
procedures used in the U.S., Europe, and Japan, discusses the
characteristics of an ideal test procedure, and evaluates the extent
to which the currently used methods fulfill this ideal.
Emission Test Procedures
Measuring gaseous emissions from an engine in steady-state operation
is a straightforward task--it is only necessary to measure the exhaust
flowrate and the pollutant concentration in the exhaust. Although
direct measurement of exhaust flowrates is difficult, this quantity is
readily approximated through measurements of the fuel flowrate, along
with either inlet air flowrate or carbon concentration in the exhaust.
-12-
-------
The situation becomes considerably more complicated where measurements
are needed during transient engine operating conditions. For these
conditions, the approach specified by regulation in the U.S. (and used
in nearly all research elsewhere) is known as constant volume sampling
or CVS. With this technique, the exhaust from the engine or vehicle
being measured is led to one end of a dilution tunnel, where it is
mixed with atmospheric air. The air/exhaust mixture is pumped from
the other end of the tunnel at a constant, known rate. Thus, the sum
of the exhaust gas flowrate and the dilution air flowrate remains
constant--no matter how the exhaust flowrate varies, the dilution air
flowrate will automatically adjust to compensate. As a result, it can
readily be shown that the pollutant concentration in the diluted
exhaust stream is proportional to the pollutant flowrate in the raw
exhaust. Average pollutant emissions over a given test cycle can then
be determined easily by sampling from the diluted stream at a constant
rate. Figure 3 is a diagram showing the principles involved.
Figure 3
Constant Volume Sampling System for Diesel Emissions Measurement
(Source: U.S. Code of Federal Regulation, 86-140.82)
nptioMAi roo
nPAIIDClfl All
UACHUIOUMO HEAOfNO
COUNltHS
10 UACRCIIOUM) SAMPlI OAc.
AMBlSNT AtA INLET
1AOSOIUIC
rncssane
TIlAHSOUCCn
zero am
4
DlSQlAttui
>«C SPAN OAS
METER
HEAIED 6AMPU line CQ
MEAlEO
FtO
TO OUTSIOC VENT
3 WAY
VALVE
OHUItON All* FILTER
1 SAMPLING train
lit At CD PRO0E
rOEUtAUS! SAM PIE UAQ
nEAO OACXCnOUNO
DILUTION TUNNEL
IICATEO Pftooe
iff At EiaUMTJU
PARTICULATE PfTOOK
I uunNOoiwics
Mdbbt^c
VCIOCU EMIAUST 1NLIT
ClUICAi FtOW
VENlURI
CYCLOMC
SEi'AftAlOR
OPTIONAL
FILTER (CttAS! I)
OAOt-UP FILTER
PIIASC 2)
PRIMARY FILTER (PIIASS I AM) 3)
BACK-UP FILTER (PIIAS1 I AND 3)
MOTe TlfKCI rlLTCR HOLDERS
[ONC FOR
ARC ALSO
(ONC FOR EACH PtlASE)
ACCCP1A8L8
CVS
COMPRE SSOR
UNI I
-13-
-------
Measurement of Diesel particulate emissions adds additional
complications to the measurement procedure. Particulate emissions are
measured by drawing dilute exhaust through a micropore filter at a
constant rate, and then measuring the weight gained by the filter.
However, Diesel particulate matter consists, in part, of condensed
hydrocarbons. If the temperature at the filter were high enough,
these could pass through the filter as hydrocarbon vapor, rather than
being collected. For this reason, EPA regulations specify a maximum
filter temperature of 50 °C. To ensure that this temperature is not
exceeded even at full power, it may be necessary to use a very high
dilution airflow, or to further dilute the gas stream to the
particulate filter by means of a second dilution tunnel. This latter
approach, which is more commonly used, is also shown in Figure 3.
The above discussion of the dilution tunnel technique emphasizes
measurement issues. It should also be mentioned that another goal in
sampling and analyzing air-diluted exhaust is to have the measurements
relate to actual in-use human exposures to the various pollutants.
Air-diluted exhaust is, therefore, a more realistic condition to
analyze than raw exhaust because of the unique chemistry that occurs
during the dilution process. This also was a reason for selecting the
50°C filter face temperature mentioned above.
Emissions Test Procedures
Mass emissions--In measuring vehicle emissions, it is desirable to
relate the emissions measured to some measure of vehicle output. In
the case of light-duty vehicles, the measure of output that has been
adopted is vehicle-miles travelled (in the U.S.) or vehicle test
cycles completed (in the European Economic Community). In the case of
heavy-duty vehicles, a different unit of output is needed. Due to the
variety of heavy-duty truck models, equipment options, and duty
cycles, it would be impractical to specify heavy-duty emissions limits
in terms of pollution per unit of distance travelled. Instead, heavy-
duty emissions are measured for the engine alone, rather than a
vehicle, and are expressed in terms of units of pollution per unit of
work output (brake horsepower-hours in the case of the U.S. cycle, and
kilowatt hours in the case of the European). As with light-duty
vehicles, these emissions are measured while operating over a fixed
test cycle, but in this case the test cycle is specified in terms of
engine, rather than vehicle, parameters.
Three emissions test procedures for heavy-duty Diesel engines are in
use around the world. The current procedure used in the U.S. is the
Federal Heavy-Duty Transient Test Procedure. In this procedure,
engine speed and load are continually varied according to a fixed
schedule, in order to simulate a typical urban driving pattern. The
particular speed-load schedule used is a composite developed from
measurement and statistical analysis of actual speed-load histories
for a large number of heavy-duty trucks driven in urban areas in the
1970's.
The Transient Test replaced an earlier 13-mode, steady-state emissions
test procedure, a variation of which is still used as the basis for
-14-
-------
European emissions regulations. The 13-mode test involves "mapping"
the engine emissions by operating at 2, 25, 50, 75, and 100% of
maximum torque at two speeds: rated speed for maximum power, and the
maximum torque speed. Three periods of idle operation complete the 13
modes. Emissions (in g/hr) and power output (in HP or KW) from each
mode are then combined according to a weighting scheme to arrive at a
composite value for each. Dividing the composite emission rate by the
composite horsepower gives composite emissions in grams per
horsepower-hour or grams per kilowatt-hour.
In the original U.S. version of the 13-mode test, all of the non-idle
modes were weighted equally in the calculation. The European version
uses a different weighting scheme which gives much more emphasis to
the maximum-torque operating modes. This is intended to reflect
driving patterns in Europe, where operation near the maximum torque
point (for best fuel economy) is said to be much more common than in
the U.S.
Japanese emissions test procedures for gaseous pollutants from heavy-
duty Diesel engines also involve a steady-state test, using a six-mode
test cycle. Unlike the U.S. and European emissions standards,
Japanese emissions standards are expressed in terms of pollutant
concentration in the exhaust (thus giving an inherent advantage to
turbocharged engines).
Smoke opacity testing--In addition to regulating mass emissions rates,
both the U.S. and EEC also regulate Diesel smoke opacity. In the
U.S., these regulations date back to 1972, and the numerical standards
been the same since 1973. With the advent of increasingly stringent
limits on PM emissions in 1988, 1991, and 1994, these standards are
becoming increasingly irrelevant.
The smoke opacity test procedure used in the U.S. simulates an
acceleration from stop, followed by a gear change and continued
acceleration, followed by "lugging down" from full engine power to the
maximum torque point. The degree of opacity of the exhaust plume is
determined by a light-transmission opacimeter. The European test
procedure is similar, but involves only the lug-down mode. These
procedures measure only the occurrence of offensively high visible
smoke levels--not particulate emissions. The correlation between the
smoke measurements and average particulate mass emissions in new
engines is poor.
Considerations in Choosing a Test Cycle
Diesel particulate and hydrocarbon emissions are fairly sensitive to
the exact test cycle used, and especially to the presence of transient
conditions. In tests of engines calibrated for U.S. emissions
standards using the older 13-mode cycle, it was found that PM and HC
emissions on the Transient Test were only moderately correlated with
those measured in the 13-mode procedure (Barsic, 1984). Indices of
correlation (R2) values for transient vs 13-mode PM and HC emissions
were only 0.55 and 0.59, respectively. Generally, PM and HC emissions
-15-
-------
on Che transient test were found to be higher, but this was very
engine-dependent. A large part of the excess emissions due to the
transient test is believed to be due to turbocharger lag, as discussed
in Section 2.2. NOx emissions showed better correlation--the index of
correlation between 13-mode and transient test results was 0.80.
Modern emission-controlled engines have been designed to maintain
effective control even during transient conditions. For these
engines, some manufacturers and consultants have even found it
practical to do their development testing using a steady-state test
cycle, with results which generally track the transient values (some
U.S. manufacturers have found that only transient testing gives them
the reliable data required, however). The fact that good correlation
can be demonstrated between transient and steady-state results in some
engines does not imply, however, that it would now be practical to
substitute a steady-state test for the transient test as the basis for
regulation. The close control of transient emission levels which
produced this correspondence was made necessary by the need to pass a
transient test. Were the transient test element to be deleted, the
transient emission controls would be unneeded, and would doubtless be
eliminated.
This observation can be generalized as follows: regulations based on a
specific emissions test procedure tend to control emissions only in
the operating modes experienced during that test procedure. Since, in
the real world (especially in urban areas), vehicles operate under a
wide variety of speed-load conditions, including transients, it is
important that the test procedure reflect these conditions. A
procedure which measures emissions only at a limited number of
specific operating conditions is vulnerable to circumvention by
emission control strategies aimed specifically at those operating
conditions.
With the advent of computer-controlled fuel injection systems, engine
control strategies can become almost arbitrarily complex. In the case
of the 13-mode, steady-state test, it would be easy to develop an
engine control strategy which optimized for emissions only in a smalL
area in the speed-load plane around each test point, with the strategy
in the remainder of the speed-load plane being optimized for fuel
economy and performance, instead. Nor is this simply a theoretical
possibility--some manufacturers of light-duty vehicles in the U.S.
have pursued a very similar strategy in calibrating their electronic
engine control systems for the light-duty test procedure.
A recent set of tests in the EPA Motor Vehicle Emission Laboratory
confirmed that such a defeat strategy is, in fact, very possible. A
series of tests were performed on a Detroit Diesel Corporation
6V-92TAD engine with full electronic control. The first test was done
with the electronic controller optimized for 1989 U.S. Federal
standards. This means no NOx control and 0.60 g/BHP-hr particulate.
The tests performed were the U.S. heavy-duty transient cycle and the
old U.S. 13-mode test. The electronic controller was then reprogramed
to a retarded timing condition only at the peak torque and rated speed
points, which are the speeds at which all of the non-idle U.S. 13-mode
-16-
-------
testing is performed. Results of this testing are summarized in
Table 1.
From the 13-mode results in Table 1, it would appear that the change
in engine calibration dramatically reduced NOx emissions. The NOx
emissions were reduced by almost half with only a slight increase in
the particulate levels. (The HC and CO emissions increased, but since
they are very low in an absolute sense they are not considered
further.) In considering the transient test data, however, it is
apparent that the reductions in NOx and particulate were minimal--
only about three percent. Since the transient test results are
considered representative of actual use, it can be concluded that
minimal if any actual control of in-use emissions would have resulted
under such a strategy.
Table 1
Effect of Engine Control Strategy
Transient vs. 13-Mode Emissions Test
HC
CO
NOx
Part.
Transient Cvcle
Current Calibration
0.45
2.62
9.33
0.28
Modified Calibration
0.47
3.05
9.05
0.27
Difference
4%
16%
-3%
-4%
US 13 Mode
Current Calibration
0.52
0.45
9.79
0.13
Modified Calibration
0.63
0.81
5.20
0.15
Difference
21%
80%
-47%
15%
Detroit Diesel 6V-92 TA
DDEC II electronic controls
Current Calibration - 1989 Federal U.S. Standards
Modified Calibration - same as "Current Calibration" except retarded
timing at rated and peak torque speeds
As further evidence that emission reductions in the transient test are
comparable to those in actual use, Table 2 compares transient test
data with chassis dynamometer emissions measurements for a number of
controlled and uncontrolled bus engines. Several pre-control bus
engines were tested over the HD-FTP transient test in the EPA heavy
duty emissions factors program. Also, a current (controlled) bus
engine was tested at the EPA facility over the HD-FTP. The same types
of engines were also tested on a chassis dynamometer by the City of
New York Department of Environmental Protection. For our analysis,
the chassis test is considered somewhat more representative of in-use
emissions than the engine test. The results in Table 2 show that the
same general levels of emissions reduction were seen as a result of
going to emission controls when the tests were performed both with the
HD-FTP Transient cycle test and the chassis test.
-17-
-------
Table 2
Emission Reductions due to Emission Controls
Engine Transient vs. Chassis Tests
HC
CO
NOx
Part.
Transient Test. g/BHP-hr
Pre-Control
Control
Difference
0.98 5.92 7.67 0.97
0.67 1.52 7.50 0.24
-32% -74% -2% -75%
Chassis test for buses, g/mile
Pre-Control
Control
Difference
9.19 32.90 42.28 4.40
2.44 3.70 41.26 0.68
-73% -89% -2% -85%
Another important consideration in test cycle selection is that
engine calibrations to reduce emissions over the selected test cycle
should result in a proportional reduction in emissions over typical
in-use driving cycles. While there is presently little information
available to show whether this is the case with the U.S. transient
cycle, the engine/chassis data cited above are one indication that
the transient cycle is resulting in effective control under other
driving cycles as well.
As pointed out by Cornetti et al. (1988), most of the work produced
during the U.S. test cycle is produced at near-rated speed and load.
While EPA considers this to be typical of truck operations in U.S.
cities, it is not typical of European driving patterns or--to an
increasing extent—present long-distance driving patterns in the
U.S. The development of high torque-rise engines, overdrive
transmissions, road-speed governors, and the increasing concern for
fuel economy are resulting in U.S. trucks spending much more operating
time near peak torque speed. This is occurring to some extent even in
urban areas. Since the transient test contains little operation at
this speed, this will allow manufacturers to calibrate their engines
for best performance and economy, rather than best emissions, in that
area of the speed-load plane. To the extent that truckers actually
run in that area, rather than near rated speed, actual in-use
emissions may be increased. The extent of this potential increase
will be unknown, however, until such time as tests are conducted.
Vehicle Testing
Experience with light-duty vehicles has shown that to obtain realistic
estimates of actual emissions from vehicles in use it is necessary to
test a representative sample of in-use vehicles. Extrapolation from
the manufacturer's certification data is generally misleading. Since
it would be impractical to remove the engines from any significant
number of heavy-duty vehicles in order to measure their emissions,
such tests would need to be made with the engine in the vehicle, using
a chassis dynamometer. To simulate the effects of transient operating
-18-
-------
conditions, the dynamometer used must be capable of simulating the
inertia of the moving vehicle, just like the chassis dynamometers used
in present light-duty emissions testing. Since the inertia of a
heavy-duty vehicle is much greater, however, the flywheels and the
dynamometer itself must be correspondingly larger.
At present, only a very limited number of heavy-duty chassis
dynamometers with transient emissions capability exist. Four such
facilities have been built in the U.S., and one in Chile (another U.S.
facility is under construction). Data from these facilities are
presently much too limited to draw any firm conclusions about the
relationship between in-use emissions and those projected from engine -
dynamometer tests. The available data are not reassuring, however--
they suggest that actual in-use emissions of HC and PM (including the
effects of emissions deterioration) may be several times those
projected from manufacturer's data. A fuller analysis of this issue
for the case of California is given by Weaver and Klausmeier (1988) .
2.4 Fuel Effects
The quality and composition of Diesel fuel can have important effects
on pollutant emissions. The area of fuel effects on Diesel emissions
has seen a great deal of study in the last few years, and a large
amount of new information has become available. These data indicate
that the fuel variables having the most important effects on emissions
are the sulfur content and the fraction of aromatic hydrocarbons
contained in the fuel. Other fuel properties may also affect
emissions, but generally to a much lesser extent. Finally, the use of
fuel additives may have a significant impact on emissions.
Sulfur Content
Diesel fuel for highway use normally contains between 0.1 and 0.5
percent by weight sulfur, although some (mostly less-developed)
nations permit 1% or even higher sulfur concentrations. Sulfur in
Diesel fuel contributes to environmental deterioration both directly
and indirectly. Most of the sulfur in the fuel burns to S02, which is
emitted to the atmosphere in the Diesel exhaust. Because of this,
Diesels are significant contributors to ambient S02 levels in some
areas. This makes them indirect contributors to ambient particulate
levels and acid deposition as well. In the U.S., Diesel fuel
accounted for about 620 thousand tons of S02 in 1984, or about 3% of
all S02 emissions during the same period.
Most of the fuel sulfur which is not emitted as S02 is converted to
various metal sulfates and to sulfuric acid during the combustion
process or immediately afterward. Both of these materials are emitted
in particulate form. The typical rate of conversion in a heavy-duty
Diesel engine is about 2 to 3 percent of the fuel sulfur; and about 3
to 5 percent in a light-duty engine. The effect of the sulfate
particles is increased by their hygroscopic nature--they tend to
absorb significant quantities of water from the air. Sulfate and
-19-
-------
associated water particles typically account for 0.05 to 0.10 g/BHP-hr
of particulates in heavy-duty engines.
Certain precious-metal catalysts can oxidize S02 to S03, which
combines with water in the exhaust to form sulfuric acid. The rate of
conversion with the catalyst is dependent on the temperature, space
velocity, and oxygen content of the exhaust, and on the activity of
the catalyst--generally, catalyst formulations which are most
effective in oxidizing hydrocarbons and CO are also most effective at
oxidizing S02. The presence of significant quantities of sulfur in
Diesel fuel thus limits the potential for catalytic converters or
catalytic trap-oxidizers for use in controlling PM and HC emissions.
Sulfur dioxide in the atmosphere oxidizes to form sulfate particles,
in a reaction similar to that which occurs with the precious-metal
catalyst. Viewed in another way, the presence of the catalyst merely
speeds up a reaction which would occur anyway (although this can have
a significant effect on human exposure to the reaction products).
According to analysis by the California Air Resources Board staff
(1984), roughly 1.20 lb of secondary particulate is formed per pound
of S02 emitted in the South Coast Air Basin. For a Diesel engine
burning fuel of 0.29 weight percent sulfur at 0.42 lb of fuel per
horsepower-hr, this is equivalent to 0.85 grams per horsepower-hour.
For comparison, the average rate of primary or directly-emitted
particulate emissions from heavy-duty Diesel engines in use was about
0.8 grams/BHP-hr in a recent study (EMA, 1985).
Quite aside from its particulate-forming tendencies, sulfur dioxide is
recognized as a hazardous pollutant in its own right. The health and
welfare effects of S02 emissions from Diesel vehicles are probably
much greater than those of an equivalent quantity emitted from a
utility stack or industrial boiler, since Diesel exhaust is emitted
close to the ground level in the vicinity of roads, buildings, and
concentrations of people.
Aromatic Hydrocarbon Content
Aromatic hydrocarbons are hydrocarbon compounds containing one or more
"benzene-like" ring structures. They are distinguished from paraffins
and napthenes, the other major hydrocarbon constituents of Diesel
fuel, which lack such structures. Compared to these other components,
aromatic hydrocarbons are denser, have poorer self-ignition qualities,
and produce more soot in burning. Ordinarily, "straight-man" Diesel
fuel produced by simple distillation of crude oil is fairly low in
aromatic hydrocarbons. Catalytic cracking of residual oil to increase
gasoline and Diesel production results in increased aromatic content,
however. A typical straight-run Diesel might contain 20 to 25%
aromatics by volume, while a Diesel blended from catalytically cracked
stocks could have 40-50% aromatics.
Aromatic hydrocarbons have poor self-ignition qualities, so that
Diesel fuels containing a high fraction of aromatics tend to have low
cetane numbers. Typical cetane values for straight run Diesel are in
-20-
-------
the range of 50-55; those for highly aromatic Diesel fuels are
typically 40 to 45, and may be even lower. This produces more
difficulty in cold starting, and increased combustion noise, HC, and
NOx due to the increased ignition delay.
Increased aromatic content is also correlated with higher particulate
emissions. Aromatic hydrocarbons have a greater tendency to form soot
in burning, and the poorer combustion quality also appears to increase
particulate SOF emissions. Increased aromatic content may also be
correlated with increased SOF mutagenicity, possibly due to increased
PNA and nitro-PNA emissions. There is also some evidence that more
highly aromatic fuels have a greater tendency to form deposits on fuel
Injectors and other critical components. Such deposits can interfere
with proper fuel-air mixing, greatly increasing PM and HC emissions.
Other Fuel Properties
Diesel fuel consists of a mixture of hydrocarbons having different
molecular weights and boiling points. As a result, as some of it
boils away on heating, the boiling point of the remainder increases.
This fact is used to characterize the range of hydrocarbons in the
fuel in the form of a "distillation curve" specifying the temperature
at which 10%, 20%, etc. of the hydrocarbons have boiled away. A low
10% boiling point is associated with a significant content of
relatively volatile hydrocarbons. Fuels with this characteristic tend
to exhibit somewhat higher HC emissions than others. Formerly, a
relatively high 90% boiling point was considered to be associated with
higher particulate emissions. More recent studies (Wall and Hoekman,
1984) have shown that this effect is spurious--the apparent
statistical linkage was due to the higher sulfur content of these
high-boiling fuels.
Other fuel properties may also have an effect on emissions. Fuel
density, for instance, may affect the mass of fuel injected into the
combustion chamber, and thus the air-fuel ratio. This is because fuel
injection pumps meter fuel by volume, not by mass, and the denser fuel
contains a greater mass in the same volume. Fuel viscosity can also
affect the fuel injection characteristics, and thus the mixing rate.
The corrosiveness, cleanliness, and lubricating properties of the fuel
can all affect the service life of the fuel injection equipment--
possibly contributing to excessive in-use emissions if the equipment
is worn out prematurely.
Fuel Additives
Several generic types of Diesel fuel additives can have a significant
effect on emissions. These include cetane enhancers, smoke
suppressants, and detergent additives. In addition, some additive
research has been directed specifically at emissions reduction in
recent years. Although some moderate emission reductions have been
demonstrated, there is yet no consensus on the widespread
applicability or desirability of such products.
-21-
-------
Cetane enhancers are used to enhance the self-ignition qualities of
Diesel fuel. These compounds (generally organic nitrates) are
generally added to reduce the adverse impact of high-aromatic fuels on
cold starting and combustion noise. These compounds also appear to
reduce the aromatic hydrocarbons' adverse impacts on HC and PM
emissions, although PM emissions with the cetane-improver are
generally still somewhat higher than those from a higher-quality fuel
able to attain the same cetane rating without the additive.
Smoke suppressant additives are organic compounds of calcium, barium,
or (sometimes) magnesium. Added to Diesel fuel, these compounds
inhibit soot formation during the combustion process, and thus greatly
reduce emissions of visible smoke. Their effects on the particulate
SOF are not fully documented, but one study (Draper et al, 1988) has
shown a significant increase in the PAH content and mutagenicity of
the SOF with a barium additive. Particulate sulfate emissions are
greatly increased with these additives, since all of them readily form
stable solid metal sulfates, which are emitted in the exhaust. The
overall effect of reducing soot and increasing metal sulfate emissions
may be either an increase or decrease in the total particulate mass,
depending on the soot emissions level at the beginning and the amount
of additive used.
Detergent additives (often packaged in combination with a cetane-
enhancer) help to prevent and remove coke deposits on fuel injector
tips and other vulnerable locations. By thus maintaining new-engine
injection and mixing characteristics, these deposits can help to
decrease in-use PM and HC emissions. A recent study for the
California Air Resources Board (Weaver and Klausmeier, 1988) estimated
the increase in PM emissions due to fuel injector problems from trucks
in use as being more than 50% of new-vehicle emissions levels. A
significant fraction of this excess is unquestionably due to fuel
injector deposits.
-22-
-------
3. DIESEL ENGINE TECHNOLOGY
Diesel engine emissions are determined by Che combustion process
within the cylinder. This process is central to the operation of the
Diesel engine. Virtually every characteristic of the engine affects
combustion in some way, and thus has some direct or indirect effect on
emissions. Some of the engine systems affecting Diesel emissions are
the fuel injection system, the engine control system, air intake and
combustion chamber, and the air charging system. Actions to reduce
lubricating oil consumption can also impact HC and PH emissions.
Finally, exhaust aftertreatment systems such as trap-oxidizers and
catalytic converters may play a significant role in reducing emissions
in the future.
This section is intended as background information for the discussion
of emissions control programs and scenarios in Sections 4 and 5. It
briefly outlines the functions and basic types of each of the
technologies and systems listed above, and describes their current
status and any ongoing developments.
3.1 Air Motion and Combustion Chamber Design
The geometries of the combustion chamber and the air intake port
control the air motion in the Diesel combustion chamber, and thus play
an important role in air-fuel mixing and emissions. A number of
different combustion chamber designs, corresponding to different basic
combustion systems, are in use in heavy-duty Diesel engines at
present. This section outlines the basic combustion systems in use,
their advantages and disadvantages, and the effects of changes in
combustion chamber design and air motion on emissions.
Combustion Systems
Diesel engines used in heavy-duty vehicles use several different types
of combustion systems. The most fundamental difference is between
direct injection (DI) engines and indirect injection (IDI) engines.
In an indirect-injection engine, fuel is injected into a separate
"prechamber," where it mixes and partly burns before jetting into the
main combustion chamber above the piston. In the more common direct-
injection engine, fuel is injected directly into a combustion chamber
hollowed out of the top of the piston. DI engines can be further
divided into high-swirl, low-swirl (quiescent chamber), and wall-
wetting designs. The latter, used by some German manufacturers, has
many characteristics in common with indirect injection systems.
Figure 4 shows a typical combustion chamber of each type.
-23-
-------
Figure 4
Different Types of Diesel Combustion Chambers
Indirect Injection
direct injection
low swirl
y/////////ArVmm v////////z, 4mm
direct injection
high swirl
direct injection
wall-wetting
-24-
-------
Fuel-air mixing in the direct-injection engine is limited by the fuel
injection pressure and any motion imparted to the air in the chamber
as it entered. In high-swirl DI engines, a strong swirling motion is
imparted to the air entering the combustion chamber by the design of
the intake port. These engines typically use moderate-to-high
injection pressures, and three to five spray holes per nozzle. Low-
swirl engines rely primarily on the fuel injection process to supply
the mixing. They typically have very high fuel injection pressures
and six to nine spray holes per nozzle. Wall-wetting DI engines also
have fairly high swirl, but the injection system is designed to
deposit the fuel on the combustion chamber wall, where it vaporizes
and burns relatively slowly.
In the indirect-injection engine, much of the fuel-air mixing is due
to the air swirl induced in the prechamber as air is forced into it
during compression, and to the turbulence induced by the expansion out
of the pre-chamber during combustion. These engines typically have
better high-speed performance than direct-injected engines, and can
use cheaper fuel-injection systems. Historically, uncontrolled IDI
Diesel engines have also exhibited lower emission levels than
uncontrolled DI engines. With recent developments in DI engine
emission controls, however, this is no longer the case. Disadvantages
of the IDI engine are the extra heat and frictional losses due to the
prechamber. These result in a 5-10 percent reduction in fuel
efficiency compared to a DI engine.
The wall-wetting DI design also shows a disadvantage in fuel economy
compared to other DI engine designs, and the further disadvantage of
high HC emissions. Its advantages include low combustion noise, and
the ability to use a cheaper fuel injection system.
Presently, all light-duty and most light-heavy-duty Diesels in the
U.S. use IDI engines, but all medium-heavy and heavy-heavy engines are
direct-injected. The same pattern is found in the rest of the world,
although the incidence of DI engines in the light-heavy-duty class
tends to be greater. Most European and Japanese truck engines, and
most medium-heavy U.S. truck engines are of the high swirl type, while
most heavy-heavy U.S. engines are low-swirl designs. A number of
advanced, low-emitting and fuel-efficient high-swirl DI engines have
recently been introduced in the light-heavy duty category as well, and
it appears that these engines will completely displace the existing
IDI designs. Small, low-emitting, high-speed DI engines (of the high-
swirl design) are also being developed for light-duty trucks and
passenger cars, but their penetration into this market segment is
still very small.
DI Combustion Chamber Design
Changes in the engine combustion chamber and related areas have
demonstrated a major potential for emission control. Design changes
to reduce the crevice volume in DI Diesel cylinders increase the
amount of air available in the combustion chamber. Changes in
combustion chamber geometry--such as the use of a reentrant lip on the
-------
piston bowl--can markedly reduce emissions by improving air-fuel
mixing and minimizing wall impingement by the fuel jet. Optimizing
the intake port shape for best swirl characteristics has also yielded
significant benefits. Several firms are considering variable swirl
intake ports, to optimize swirl characteristics across a broader range
of engine speeds.
Crevice volume--The crevice volume is that part of the compression
volume which lies outside the combustion chamber. This includes the
clearance between the top of the piston and the cylinder head, and the
"top land"--the space between the side of the piston and the cylinder
wall above the top compression ring. The air in these volumes
contributes little to the combustion process. The smaller the crevice
volume, the larger the combustion chamber volume can be for a given
compression ratio. Thus, reducing the crevice volume effectively
increases the amount of air available for combustion.
The major approaches to reducing the crevice volume are to reduce the
clearance between the piston and cylinder head through tighter
production tolerances, and to move the top compression ring toward the
top of the piston. This increases the working temperature of the top
ring, and poses mechanical design problems for the piston top and
cooling system as well. These problems have been addressed through
redesign and the use of more expensive materials. The higher piston
ring temperature may also make additional demands on the oil.
Combustion chamber shape--Numerous test results indicate that, for
high swirl DI engines, a reentrant combustion chamber shape (in which
the lip of the combustion chamber protrudes beyond the walls of the
bowl) provides a substantial improvement in performance and emissions
over the previous straight-sided bowl designs. Researchers at AVL
(Cartellieri and Wachter, 1987) found that the use of a reentrant bowl
gave a 20 percent reduction in PM emissions from those measured with a
straight-sided bowl at the same compression ratio. NOx emissions were
increased 3 percent, but the reentrant bowl combustion chamber is also
more tolerant of retarded injection timing than the straight-sided
bowl.
Because of the superiority of the reentrant bowl design for high-swirl
engines, most manufacturers of such engines are developing or already
using this approach. Similar improvements in the performance of low-
swirl DI engines may also be possible through modifications to
combustion chamber geometry, but there is much less unanimity as to
what the optimal shape may be. A number of different variations on
the classic "Mexican hat" combustion chamber shape have been tried,
with some success.
Intake air swirl--Optimal matching of intake air swirl ratio with
combustion chamber shape and other variables is critical for emissions
control in high-swirl engines. The swirl ratio is the ratio of the
rotational speed of the air charge in the cylinder to the rotational
speed of the engine, which is determined by the design of the air
intake port. Unfortunately, the selection of a fixed swirl ratio
involves some tradeoffs between low-speed and high-speed performance.
-26-
-------
At low speeds, a higher swirl ratio provides better mixing, permitting
more fuel to be injected and thus greater torque output at the same
smoke level. However, this can result in too high a swirl ratio at
higher speeds, impairing the airflow to the cylinder. Too high a
swirl ratio can also increase HC emissions, especially at light loads.
Attaining an optimal swirl ratio is most difficult in smaller light-
heavy and medium-heavy DI engines, as these experience a wider range
of engine speeds than do heavy-heavy engines. One solution to this
problem is to vary the swirl ratio as a function of engine speed. A
two-position variable swirl system has been developed and applied to
some Diesel engines in Japan (Shimada et al, 1986). This system is
being considered for engines used in the U.S. as well. Test data
using this system show a noticeable reduction in PK and NOx emissions
due to optimization of the swirl ratio at different speeds.
3.2 Fuel infection
The fuel injection system in a Diesel engine includes the machinery by
which the fuel is transferred from the fuel tank to the engine, then
injected into the cylinders at the right time for optimal combustion,
and in the correct amount to provide the desired power output. The
quality, quantity, and timing of fuel injection determine the engine's
power, fuel economy, and emissions characteristics, so that the fuel
injection system is one of the most important components of the
engine.
The fuel injection system normally consists of a low-pressure pump to
transfer fuel from the tank to the system, one or more high-pressure
fuel pumps to create the pressure pulses that actually send the fuel
into the cylinder, the injection nozzles through which fuel is
injected into the cylinder, and a governor and fuel metering system.
These determine how much fuel is to be injected on each stroke, and
thus the power output of the engine.
The major areas of concentration in fuel injection system development,
have been on increased injection pressure, increasingly flexible
control of injection timing, and more precise governing of the fuel
quantity injected. Systems offering electronic control of these
quantities, as well as fuel injection rate, have been introduced.
Some manufacturers are also pursuing technology to vary the rate of
fuel injection over the injection period, in order to reduce the
amount of fuel burning in the premixed combustion phase. Reductions
in NOx and noise emissions and maximum cylinder pressures have been
demonstrated using this approach (Gill, 1988). Other changes have
been made to the injection nozzles themselves, to reduce or eliminate
sac volume and to optimize the nozzle hole size and shape, number of
holes, and spray angle for minimum emissions.
-------
Injection System Types
Fuel injection systems used in heavy-duty Diesel vehicles can be
divided into two basic types. The most common type consists of a
single fuel pump (typically mounted at the side of the engine) which
is driven by gears from the crankshaft, and connected to individual
injection nozzles at the top of each cylinder by special high-pressure
fuel lines. These pump-line-nozzle (PLN) injection systems can be
further divided into two subclasses: "distributor" fuel pumps, in
which a single pumping element is mechanically switched to connect to
the high-pressure fuel lines for each cylinder in turn; and "in-line"
pumps having one pumping element per cylinder, each connected to its
own high-pressure fuel line. The latter type is much more common in
heavy-duty trucks.
The most common alternative to the pump-line nozzle injection systems
are systems using unit injectors, in which the individual fuel
metering and pumping element for each cylinder is combined in the same
unit with the injection nozzle at the top of the cylinder. The
pumping elements in a unit injector system are generally driven by the
engine camshaft. Figure 5 shows some typical examples of pump-line
nozzle and unit injector fuel injector systems.
Worldwide, many more engines are made with pump-line-nozzle injection
systems than with unit injectors. This is primarily due to the higher
cost of unit injector systems. Presently, three U.S. engine
manufacturers (accounting for more than half of U.S. heavy-heavy-duty
engine production) produce unit-injector-equipped truck engines. Due
to the absence of high-pressure fuel lines, however, unit injectors
are capable of higher injection pressures than pump-line-nozzle
systems. With improvements in electronic control, these systems offer
better fuel economy at low emission levels than the pump-line-nozzle
systems. For this reason, many heavy-duty engine models sold in the
U.S. will be equipped with unit injectors for the 1991 model year.
Fuel injection pressure and injection rate--High fuel injection
pressures are desirable in order to improve fuel atomization and fuel-
air mixing, and to offset the effects of retarded injection timing by
increasing the injection rate. A number of workers have published
data on the effects of higher injection pressures on PM and/or smoke
emissions. All show marked reductions as injection pressure is
increased. High injection pressures are most important in low-swirl,
direct-injection engines, since the fuel injection system is
responsible for most of the fuel-air mixing in these systems. For
this reason, low-swirl engines tend to use unit injector systems,
which can achieve peak injection pressures in excess of 1,500 bar.
The injection pressures achievable in pump-line-nozzle fuel injection
systems are limited by the mechanical strength of the pumps and fuel
lines, as well as by pressure wave effects, to about 800 bar.
Improvements in system design to minimize pressure wave effects, and
increases in the size and mechanical strength of the lines and pumping
elements have increased the injection pressures achievable in pump-
line nozzle systems substantially from those achievable a few years
-28-
-------
Figure 5
Typical Diesel Fuel Injection Systems
(Source: Lilly, 1984)
Pump-Line-Nozzle System
(only one of six injectors shown)
Unit Injector System
-29-
-------
ago. It now appears that a point of diminishing returns may have been
reached in this area--further increases in injection pressure in some
experimental systems have not greatly improved emissions.
The pumping elements in all current fuel injection systems are driven
through a fixed mechanical linkage from the engine crankshaft. This
means that the pumping rate, and thus the injection pressure, are
strong functions of engine speed. At high speeds, the pumping element
moves rapidly, and injection pressures and injection rates are high.
At lower speeds, however, the injection rate is proportionately lower,
and injection pressure drops off rapidly. This can result in poor
atomization and mixing at low speeds, and is a major cause of high
smoke emissions during lugdown. Increasing the pumping rate to
provide adequate pressure at low speeds is impractical, as this would
exceed the system pressure limits at high speed.
A new type of in-line injection pump has recently been developed which
provides a partial solution to this problem (Ishida et al., 1986).
The cam driving the pumping elements in this pump has a non-uniform
rise rate, so that pumping rate at any given time is a function of the
cam angle. By electronically adjusting a spill sleeve, it is possible
to select the portion of the cam's rotation during which fuel is
injected, and thus to vary the injection rate. Injection timing
varies at the same time, but the system is designed so that desired
injection rate and injection timing correspond fairly well. Ishida
and coworkers obtained a 25 percent reduction in PM emissions and a 10
percent reduction in HC using this system, with virtually no increase
in NOx. The same approach could easily be applied to a unit injector
system, using an electronically controlled spill valve.
Another approach to increasing injection pressure at low engine speeds
is the use of electro-hydraulic actuators for injection instead of
mechanically-driven pumping elements. Through appropriate design and
control schemes, such systems can control and maintain fuel injection
pressures nearly independently of engine speed. A number of such
systems have been described in the technical literature, but--to
date--none has actually been implemented on commercial engines. At
least one major engine manufacturer plans to introduce such systems in
the U.S. in 1991, however.
Initial injection rate and oremixed burning--Reducing the amount of
fuel burned in the premixed combustion phase can significantly reduce
total NOx emissions. This can be achieved by reducing the initial
rate of injection, while keeping the subequent rate of injection high
to avoid high PM emissions due to late burning. This requires varying
the rate of injection during the injection stroke. This represents a
difficult design problem for mechanical injections systems, but should
be possible using electro-hydraulic injectors. Another approach to
the same end is split injection, in which a small amount of fuel is
injected in a separate event ahead of the main fuel injection period.
Data published by a U.S. manufacturer (Gill, 1988) show a marked
beneficial effect from reducing the initial rate of injection. Based
on these data, it appears likely that a 30 to 40 percent reduction in
-------
NOx emissions could be achieved through this technique, without
significant adverse impacts on fuel consumption, HC, or FM emissions.
As a side benefit, engine noise and maximum cylinder pressures (for a
given power output) are also reduced.
Low sac/sacless nozzles--The nozzle sac is a small internal space in
the tip of the injection nozzle. The nozzle orifices open into the
sac, so that fuel flowing past the needle valve first enters the sac,
and then sprays out the orifices. The small amount of fuel remaining
in the sac tends to bum or evaporate late in the combustion cycle,
resulting in significant PM and HC emissions. The sac volume can be
minimized or even eliminated by redesigning the injector nozzle. One
manufacturer reported nearly a 30 percent reduction in PM emissions
through elimination of the nozzle sac. It is also possible to retain
some of the sac while designing the injector nozzle so that the tip of
the needle valve covers the injection orifices when it is closed.
This valve-covers-orifice or VCO injector design is used in some
production engines, and in many engines being developed for compliance
with the U.S. 1991 emissions standards.
3.3 Engine controls
Traditionally, Diesel engine control systems have been closely
integrated with the fuel injection system, and the two systems are
often discussed together. These earlier control systems (still in use
on most engines) are entirely mechanical. The last few years have
seen the introduction of an increasing number of computerized
electronic control systems for Diesel engines. With the introduction
of these systems, the scope of the engine control system has been
greatly expanded.
Mechanical Controls
Most current Diesel engines still rely on mechanical engine control
systems. The basic functions of these systems include basic fuel
metering, engine speed governing, maximum power limitation, torque
curve "shaping", limiting smoke emissions during transient
acceleration, and (sometimes) limited control of fuel injection
timing. Engine speed governing is accomplished through a spring-and
flyweight system which progressively (and quickly) reduces the maximum
fuel quantity as engine speed exceeds the rated value. The maximum
fuel quantity itself is generally set through a simple mechanical stop
on the rack controlling injection quantity. More sophisticated
systems allow some "shaping" of the torque curve to change the maximum
fuel quantity as a function of engine speed.
Acceleration smoke limiters are needed to prevent excessive black
smoke emissions during transient acceleration of turbocharged engines.
Most are designed to limit the maximum fuel quantity injected as a
function of turbocharger boost, so that full engine power is developed
only after the turbocharger comes up to speed.
-31-
-------
Many pump-line-nozzle fuel injection systems incorporate mechanical
injection timing controls. Since the injection pump is driven by a
special shaft geared to the crankshaft, injection timing can be
adjusted within a limited range by varying the phase angle between the
two shafts, using a sliding spline coupling. A mechanical or
hydraulic linkage slides the coupling back and forth in response to
engine speed and/or load signals.
In mechanical unit injector systems, the injectors are driven by a
direct mechanical linkage from the camshaft, making it very difficult
to vary the injection timing. Cummins, in its California engines, has
introduced a mechanical timing control which operates by moving the
injector cam followers back and forth with respect to the cam.
Although effective in limiting light-load HC and PM emissions under
the stringent California NOx standards, these systems have proven very
troublesome and unpopular among users.
Computerized Electronic Controls
The advent of computerized electronic engine control systems has
greatly increased the potential flexibility and precision of fuel
metering and injection timing controls. In addition, it has made
possible whole new classes of control functions, such as road-speed
governing, alterations in control strategy during transients,
synchronous idle speed control, and adaptive learning--including
strategies to identify and compensate for the effects of wear and
component-to-component variation in the fuel injection system.
By continuously adjusting the fuel injection timing to match a stored
"map" of optimal timing vs. speed and load, an electronic timing
control system can significantly improve on the NOx/particulate and
NOx/fuel-economy tradeoffs possible with static or mechanically-
variable injection timing. Most electronic control systems also
incorporate the functions of the engine governor and the transient
smoke limiter. This helps to reduce excess particulate emissions due
to mechanical friction and lag-time during engine transients, while
simultaneously improving engine performance. Potential reductions in
PM emissions of up to 40% been documented with this approach by Wade
and coworkers (1983).
A potential drawback of the increasing use of electronic controls is
the demand that it places on the emissions test procedure. As
discussed in Section 2.3, such a control system could easily be
programmed to defeat a simple steady-state test procedure such as the
13-mode by maintaining optimal emission control only near the test
points, not across the entire range. A similar problem is possible
in the Federal Transient Test, given the apparent mismatch between the
operating conditions emphasized in the test and those found in long-
distance truck operation. The earlier mechanical controls were
incapable of such sophisticated control strategies.
Other electronic control features, such as cruise control, upshift
indication, and communication with an electronically controlled
-32-
-------
transmission will also help Co reduce fuel consumption, and will thus
likely reduce in-use emissions. Since the effect of these
technologies is to reduce the number of BHP-hrs per mile, rather than
the amount of pollution per BHP-hr, their effects will not be
reflected in dynamometer emissions test results, however.
3.4 Turbochareine and Intercooling
A turbocharger consists of a centrifugal air compressor feeding the
intake manifold, mounted on the same shaft as an exhaust gas turbine
in the exhaust stream. By increasing the mass of air in the cylinder
prior to compression, turbocharging correspondingly increases the
amount of fuel that can be burned without excessive smoke, and thus
increases the potential maximum power output. The fuel efficiency of
the engine is improved as well. The process of compressing the air,
however, increases its temperature, increasing the thermal load on
critical engine components. By cooling the compressed air in an
intercooler before it enters the cylinder, the adverse thermal effects
can be reduced. This also increases the density of the air, allowing
an even greater mass of air to be confined within the cylinder, and
thus further increasing the maximum power potential.
Increasing the air mass in the cylinder and reducing its temperature
can reduce both NOx and particulate emissions, as well as permitting
greater fuel economy and more power output from a given engine
displacement. Most heavy-duty Diesel engines are presently equipped
with turbochargers, and most of these have intercoolers. In the U.S.,
virtually all engines will be equipped with these systems by 1991.
Recent developments in air charging systems for Diesel engines have
been primarily concerned with increasing the turbocharger efficiency,
operating range, and transient response characteristics; and with
improved intercoolers to further reduce the temperature of the intake
charge. Tuned intake air manifolds (including some with variable
tuning) have also been developed, to maximize air intake efficiency in
a given speed range.
Turbocharger refinements
Turbochargers for heavy-duty Diesel engines are already highly
developed, but efforts to improve their performance continue. The
major areas of emphasis are improved matching of turbocharger response
characteristics to engine requirements, Improved transient response,
and higher efficiencies. Engine/turbocharger matching is especially
critical, because of the inherent conflict between the response
characteristics of the two types of machines. Engine boost pressure
requirements are greatest near the maximum torque speed, and most
turbochargers are matched to give near-optimal performance at that
point. At higher speeds, however, the exhaust flowrate is greater,
and the turbine power output is correspondingly higher. Boost
pressure under these circumstances can exceed the engine's design
limits, and the excessive turbine backpressure increases fuel
-33-
-------
consumption. Thus, some compromise between adequate low-speed boost
and excessive high-speed boost must be made.
Variable geometry turbochargers
Because of the inherent mismatch between engine response
characteristics and those of a fixed-geometry turbocharger, a number
of engine manufacturers are considering the use of variable geometry
turbines instead (Wallace et al., 1986). In these systems, the
turbine nozzles can be adjusted to vary the turbine pressure drop and
power level in order to match the engine's boost pressure
requirements. Thus, high boost pressures can be achieved at low
engine speeds, without wasteful overboosting at high speed. The
result is a substantial improvement in low-speed torque, transient
response, and fuel economy, and a reduction in smoke, NOx, and PM
emissions.
Prototype variable geometry turbochargers (VGT) have been available
for some time, but they have not been used in production vehicles up
to this point. The major reasons for this are their cost (which could
be 50% more than a comparable fixed-geometry turbocharger),
reliability concerns, and the need for a sophisticated electronic
control system to manage them. Vith the forthcoming deployment of
electronic engine controls on virtually all vehicles in the U.S.,
these latter arguments have lost much of their force, and the fuel
economy and performance advantages of the VGT are great enough to
outweigh the costs in many applications. As a result, variable
geometry turbochargers should be available on a number of production
heavy-duty Diesel engines in the relatively near future.
Other types of superchargers
A number of alternative forms of supercharging have been considered,
with a view to overcoming the mismatch between turbocharger and engine
response characteristics. The two leading candidates at present are
the Sulzer Comprex (tm) gas-dynamic supercharger, and mechanically-
assisted turbochargers such as the "three-wheel" turbocharger
developed by General Motors. The major advantages of these systems
are superior low-speed performance and improved transient response.
These advantages would be expected to yield some improvement in PM
emissions, as well as driveability and torque rise.
Intercoolers
As discussed in Section 2.2, charge air cooling helps to reduce both
NOx and PM emissions, while increasing maximum power and decreasing
fuel consumption and the thermal loading on engine components.
Because of these advantages, intercoolers are almost universally used
on highly-rated turbocharged engines. Presently, most intercoolers
rely on the engine cooling water as a heat sink, since this minimizes
the components required. The relatively high temperature of this
-34-
-------
water (about 90° C) limits the benefits available, however. For this
reason, an increasing number of heavy-duty Diesel engines are being
equipped with low-temperature charge-air cooling systems.
The most common type of low-temperature charge-air cooler rejects heat
directly to the atmosphere through an air-to-air heat exchanger
mounted on the truck chassis in front of the radiator. Although bulky
and expensive, these charge-air coolers are able to achieve the lowest
charge-air temperatures--in many cases, only ten or 15 degrees C above
ambient. An alternative approach is low-temperature air-to-water
intercooling, which has been pursued by Cummins Engine in the U.S.
Cummins has chosen to retain the basic water-air intercooler, but with
drastically reduced radiator flowrates to reduce the water temperature
coming from the radiator. This water is then passed through the
intercooler before it is used for cooling the rest of the engine.
Intake manifold tuning
Tuned intake manifolds have been used for many years to enhance
airflow rates on high-performance gasoline engines, and are being
considered for some heavy-duty Diesel engines. A tuned manifold
provides improved airflow and volumetric efficiency at speeds near its
resonant frequency, at the cost of reduced volumetric efficiency at
other speeds. At least one medium-heavy-duty manufacturer is
considering a variable-resonance manifold, in order to improve airflow
characteristics at both low and high speeds.
3.5 Exhaust Gas Recirculation
EGR is a time-proven NOx control technique for light-duty gasoline and
Diesel vehicles, but has been little used in heavy-duty Diesel
engines. In heavy-duty Diesel engines, EGR has been shown to increase
wear rates and oil contamination, resulting in higher maintenance
expenses and shorter engine life (Cadman and Johnson, 1986). For this
reason, engine manufacturers have avoided it, and little research on
its effects has been performed. In the past, a few California-model
medium-heavy engines used EGR to meet the California NOx standard.
Considerable adverse experience with these engines has reinforced the
existing prejudice against EGR use in heavy-duty Diesels.
Another reason for avoiding EGR is that it was considered to have
little advantage over other NOx control techniques such as retarding
injection timing, at least in DI engines. Yu and Shahed (1981) found
little difference in the NOx/smoke tradeoff curves for EGR and for
injection timing. EGR has a lesser impact on fuel economy than
retarded timing (moderate EGR actually improves fuel economy
slightly), but this has been outweighed by its perceived adverse
effects on durability.
Some recent research results suggest that a re-evaluation of this
technique may be in order, however. This research indicates that
properly modulated EGR does not necessarily increase PM emissions
-35-
-------
significantly, even though NOx may be dramatically reduced. EGR often
(but not always) increases soot emissions, but gaseous HC and
particulate SOF are generally reduced. In some cases, soot emissions
may be reduced by EGR as well (Shiga et al., 1985). The effect of EGR
on overall PM emissions may thus be positive or negative, depending on
the specific operating mode.
Data from a number of manufacturers (e.g. Wade et al., 1985) have
shown that exhaust gas recirculation can reduce NOx emissions from
light-duty Diesel vehicles by 50% or more, without adverse effects on
PM emissions or fuel economy. These results cannot be translated
directly to heavy-duty engines, however, due to the differences in
emission test cycles between the light-duty FTP and the heavy-duty
test procedures. Compared to the light-duty cycle, the heavy-duty
procedures involve much more high-power operation, which would limit
the amount of EGR that could be tolerated. Nonetheless, these data
_strongly suggest that the properly modulated EGR could result in a
major reduction in NOx emissions, with minimal impacts on PM, fuel
economy, or driveability.
To obtain a rough quantification of the potential impact of EGR in a
heavy-duty engine, Weaver and Klausmeier (1987) developed a rough
model of EGR effects, using published modal emissions data for a
modern low-emission heavy-duty engine. For this engine, good
correlation had been demonstrated between transient cycle emissions
and emissions in a specialized steady-state cycle constructed to mimic
the operating conditions in the transient cycle (Cartellieri and
Wachter, 1987). The effects of EGR on emissions in this steady-state
cycle were estimated by assigning an EGR rate to each operating mode,
then estimating the resultant effects on soot, SOF, and NOx emissions.
The results of this calculation showed a 27% reduction in NOx
emissions (from 6.0 to 4.3 g/BHP-hr), at the cost of a 14% increase in
PM (from 0.242 to 0.276 g/BHP-hr). Although very rough, these results
may be considered as an approximate indicator of EGR's potential for
heavy-duty Diesel engines. Compared to the effects of injection
timing retardation at similar NOx levels, the tradeoff ratio of 1.7
g/BHP-hr NOx reduction to 0.034 g/BHP-hr PM increase is an extremely
favorable one.
3.6 Lubricating Oil Control
A significant fraction of Diesel particulate matter consists of oil-
derived hydrocarbons and related solid matter. The long-chain
(typically 20-carbon) hydrocarbons in the oil readily condense to form
liquid particles in the dilute exhaust, and are subsequently
collected on the particulate filter. A number of researchers have
estimated the oil contribution to particulate emissions by assigning
all of the heavy-hydrocarbon fraction of the SOF to the oil.
Estimates of the oil contribution to overall PM emissions by this
means range from 10 to 50%. More accurate measurements using a
radioactive tracer technique show a somewhat smaller--but still very
significant--contribution by the lube oil to total SOF. The
-36-
-------
difference between the two methods may be due to pyrosynthesis of high
molecular weight hydrocarbons from the lower molecular weight
hydrocarbons in the fuel.
Reduced oil consumption has been a design goal of heavy-duty Diesel
engine manufacturers for some time, and the current generation of
Diesel engines already uses fairly little oil compared to their
predecessors. Further reductions in oil consumption are possible
through careful attention to cylinder bore roundness and surface
finish, optimization of piston ring tension and shape, and attention
to valve stem seals, turbocharger oil seals, and other possible
sources of oil loss. Some oil consumption in the cylinder is
required with present technology, however, in order for the oil to
perform its lubricating and corrosion-protective functions.
Advances in piston/cylinder tribology could potentially eliminate or
greatly .>;duce oil consumption in the cylinder. Areas such as
boundary lubrication and development of low-friction ceramic coatings
are presently the subjects of much research. The potential for
transforming this research into durable and reliable engines on the
road remains to be demonstrated, however.
Some manufacturers have measured emissions from engines modified for
ultra-low oil consumption. Oil consumption in each case was reduced
to an extent that was expected to result in unacceptable long-term
durability. These tests have shown a reduction of 0.07 to 0.10 g/BHP-
hr in particulate matter due to the reduced oil consumption.
3.7 Aftertreatment Systems
The preceding discussion has dealt entirely with measures to reduce
pollutant concentrations in the exhaust before it leaves the engine.
An alternative approach is to use a separate processing system to
eliminate pollutants from the exhaust after it leaves the engine, but
before it is emitted from the tailpipe into the ambient air.
Possibilities for such aftertreatment systems include particulate
trap-oxidizers, Diesel catalytic converters, electrostatic
agglomerators, and various techniques for reducing NOx to oxygen and
nitrogen gases.
Trap-Oxidizers
A trap-oxidizer system consists of a durable particulate filter (the
"trap") positioned in the engine exhaust stream, along with some means
for cleaning the filter by burning off ("oxidizing") the collected
particulate matter. The construction of a filter capable of
collecting Diesel soot and other particulate matter from the exhaust
stream is a straightforward task, and a number of effective trapping
media have been developed and demonstrated. The great problem of
trap-oxidizer system development has always been with the process of
"regenerating" the filter by burning off the accumulated particulate
matter.
-37-
-------
As discussed in Section 2.1, Diesel particulate matter consists
primarily of a mixture of solid carbon coated with heavy hydrocarbons.
The ignition temperature of this mixture is about 500-600° C, which is
above the normal range of Diesel engine exhaust temperatures. Thus,
special means are needed to assure regeneration. Once ignited,
however, this material bums to produce very high temperatures, which
can easily melt or crack the particulate filter. Initiating and
controlling the regeneration process to ensure reliable regeneration
without damage to the trap is the central engineering problem of trap-
oxidizer development.
Compared to the in-cylinder emission control technologies discussed in
the preceding sections, trap-oxidizer technology has progressed more
slowly and in a more predictable manner. This is partly due to the
simpler and more predictable physical and chemical phenomena involved,
and partly due to the much lower priority accorded it in most
manufacturers' research and development efforts. Despite the
relatively slow general rate of progress, several manufacturers have
fielded successful prototype trap-oxidizer systems.
Traps - -Presently. most of the trap-oxidizer systems under development
are based on the cellular cordierite ceramic monolith traps produced
by Corning Glassworks and NGK-Locke. Figure 6 shows the principal of
operation and a typical example of such a trap. These traps can be
formulated to be highly efficient (collecting essentially all of the
soot, and a large fraction of the particulate SOF), and they are
relatively compact, having a large surface area per unit of volume.
Because of their relatively simple production process, they could also
be produced fairly inexpensively. They can also be coated or
impregnated with catalyst material to assist regeneration.
Figure 6
Principle of the Ceramic Monolith Trap
(Source: Gulati and Merry, 1984)
INTERAM MAT /SSCAN
3E
ENTRY
EXIT
CERAMIC WALL-FLOW FILTER
-38-
-------
The high concentration of soot per unit of volume with the ceramic
monolith makes these traps rather sensitive to regeneration
conditions. Trap loading, temperature, and gas flow rates must be
maintained within a fairly narrow window. Otherwise, the trap fails
to regenerate fully, or cracks or melts due to overheating.
An alternative trap technology is provided by the ceramic fiber coil
traps developed by Mann and Hummel and Daimler Benz in West Germany.
These traps are composed of a number of individual filtering elements,
each of which consists of a number of thicknesses of silica-fiber yarn
wound on a punched metal support. A number of these filtering
elements are suspended inside a large metal can to make up a trap.
Daimler-Benz has deployed prototype traps of this design in more than
50 city buses in West Germany (Hardenberg, 1987).
The advantages of the ceramic fiber coil trap include high filtering
efficiency and immunity to thermal cracking. In addition, its low
ratio of filtering area to volume (which results in a low volumetric
soot loading) and the heat capacity and thermal conductivity of the
materials make this trap nearly impossible to melt. Its primary
disadvantages are the relatively large volume required and a fairly
rapid increase in backpressure with increasing particulate loading.
The silica yarn coils can also be mechanically cut or frayed by sharp
objects, or loosened by repeated thermal cycling (Hardenberg et al.,
1987a). The trap is also relatively complex structurally-- implying
that it could be expensive to manufacture.
Numerous other trapping media have been tested or proposed. These
include ceramic foams, corrugated mullite fiber felts, and catalyst
coated stainless steel wire mesh. Traps based on the latter
technology were demonstrated in a number of U.S. and European
programs, most of which showed rather poor performance. Presently,
all of the most successful trap-oxidizer systems under development are
based on either the ceramic monolith or the silica fiber coil traps.
Regeneration--Numerous techniques for regenerating particulate trap-
oxidizers have been proposed, and a great deal of development work has
been invested in many of these. These approaches can generally be
divided into two groups: passive systems and active systems. Passive
systems rely on attaining the conditions required for regeneration as
a result of the normal operation of the vehicle. This requires the
use of a catalyst (either as a coating on the trap or as a fuel
additive) in order to reduce the ignition temperature of the collected
particulate matter. Regeneration temperatures as low as 420° C have
been reported with catalytic coatings, and even lower temperatures are
achieveable with fuel additives.
Active systems, on the other hand, monitor the buildup of particulate
matter in the trap and trigger specific actions intended to regenerate
it when needed. A wide variety of approaches to triggering
regeneration have been proposed, from Diesel fuel burners and electric
heaters to catalyst injection systems.
-39-
-------
Passive regeneration systems face special problems on heavy-duty
vehicles. Regeneration temperatures must be attained in normal
operation, even under lightly loaded conditions. Because of the
variability in their loading and use patterns, trucks may sometimes
operate for long periods at very light loads. Exhaust temperatures
from heavy-duty Diesel engines are already fairly low, and recent
developments such as charge-air cooling and increased turbocharger
efficiency are reducing them still further. Under some conditions,
therefore, it would be possible for a truck to drive for many hours
without exceeding the exhaust temperature (around 400-450° C) required
to trigger regeneration.
Presently, no purely passive systems appear to be under serious
consideration for heavy-duty applications. However, some
manufacturers are working on quasi-passive systems, in which the
system will usually regenerate passively without intervention, but the
active regeneration system remains as a backup.
Active regeneration systems can be classified as either in-line or
bypass-type systems. In the in-line system, exhaust continues to flow
through the trap during regeneration, while with the bypass system the
exhaust is bypassed around the trap. The exhaust stream from a
vehicle engine varies rapidly and unpredictably in both temperature
and flowrate, depending on the demands of the driving cycle. This
variability would pose impossible control problems for systems such as
Diesel fuel burners and electric heaters. The need to heat the entire
exhaust stream would also be very wasteful of energy, and would be
well beyond the capacity of a truck's electrical system. For these
reasons, burner and electric heater-based regeneration systems usually
bypass the exhaust around the trap during regeneration.
Engine and catalyst manufacturers have experimented with a wide
variety of catalytic material and treatments to assist in trap
regeneration. Good results have been obtained both with precious
metals (platinum, palladium, rhodium, silver) and with base metal
catalysts such as vanadium and copper. Precious metal catalysts are
effective in oxidizing gaseous HC and CO, as well as the particulate
SOF, but are relatively ineffective at promoting soot oxidation.
Unfortunately, these metals also promote the oxidation of NO in the
exhaust to the more toxic N02, and of S02 to particulate sulfates such
as sulfuric acid (H2S04). The base-metal catalysts, in contrast, are
effective in promoting soot oxidation, but have little effect on HC,
CO, NO, or S02.
To date, no catalytic coating has sufficiently reduced the trap
regeneration temperature to permit reliable passive regeneration in
heavy-duty Diesel service. Catalyst coatings also have a number of
advantages in active systems, however. The reduced ignition
temperature and increased combustion rate due to the catalyst mean
that less energy is needed from the regeneration system. Regeneration
will also occur spontaneously under most duty cycles, greatly reducing
the number of times the regeneration system must operate. The
spontaneous regeneration capability also provides some insurance
-40-
-------
against a regeneration system failure. Finally, the use of a catalyst
may make possible a simpler regeneration system.
Although normal heavy-duty Diesel exhaust temperatures are not high
enough under all operating conditions to provide reliable regeneration
for a catalyst-coated trap, the exhaust temperature can readily be
increased by changes in engine operating parameters. Retarding the
injection timing, bypassing the intercooler, throttling the intake air
(or cutting back on a variable geometry turbocharger), and/or
increasing the EGR rate all markedly increase the exhaust temperature.
Applying these measures all the time would seriously degrade fuel
economy, engine durability, and performance. The presence of an
electronic control system, however, would make it possible to apply
them briefly, and only when needed to regenerate the trap. Since they
would be needed only at light loads, the effects on durability and
performance would be imperceptible. One engine manufacturer has
successfully accumulated more than 145,000 miles on a prototype system
of this type.
Catalytic converters
Recent dramatic progress in in-cylinder particulate control has
greatly reduced engine-out particulate levels. This progress has been
most effective in reducing the solid soot fraction of the particulate,
so that the soluble organic fraction (SOF) of the particulate matter
now accounts for a much larger share than previously. Depending on
the engine and operating conditions, the SOF may account for from 30
to more than 70 percent of the engine-out particulate matter.
Like a catalytic trap, a Diesel catalytic converter oxidizes a large
part of the hydrocarbon constituents of the SOF, as well as gaseous
HC, CO, odor, and mutagen emissions. Unlike a catalytic trap however,
a flow-through catalytic converter does not collect any of the solid
particulate matter, which simply passes through in the exhaust. This
eliminates the need for a regeneration system, with its attendant
technical difficulties and costs. The particulate control efficiency
of the catalytic converter is, of course, much less than that of a
trap. However, a particulate control efficiency of even 25 to 35
percent is enough to bring many current development engines within the
target range for the U.S. 1991 emissions standard.
Diesel catalytic converters have a number of advantages. In addition
to reducing particulate emissions enough to comply with the 1991
standard, the oxidation catalyst greatly reduces HC, CO, and odor
emissions. The catalyst is also very efficient in reducing emissions
of gaseous and particle-bound toxic air contaminants such as
aldehydes, PNA, and nitro-PNA. While a precious-metal catalyzed
particulate trap would have the same advantages, the catalytic
converter is much less complex, bulky, and expensive. Unlike the
trap, the catalytic converter has little impact on fuel economy or
safety, and it will probably not require replacement as often. Also,
unlike the trap-oxidizer, the catalytic converter is a relatively
mature technology--millions of catalytic converters are in use on
-41-
-------
gasoline vehicles, and Englehard Corporation PTX (tm) Diesel catalytic
converters have been used in underground mining applications for more
than 20 years.
The disadvantages of the catalytic converter are the same as those of
the precious-metal catalyzed particulate trap: sulfate emissions and
conversion of NO to the more toxic N02. The NO to N02 conversion
occurs naturally in the atmosphere, so the only differences in N02
exposure would occur where people are exposed to relatively fresh
exhaust. The increase in the toxic effects of N02 under these
circumstances should be more than counterbalanced by the decrease in
CO, aldehydes, PNA, and nitro-PNA. However, the tendency of the
precious-metal catalyst to convert S02 to particulate sulfates
requires the use of low-sulfur fuel: otherwise, the increase in
sulfate emissions would more than counterbalance the decrease in SOF.
As discussed in Section 5.4, however, low-sulfur fuel appears to be a
cost-effective emissions control measure in any case. Regulations
mandating low-sulfur fuel (0.05 wt. percent sulfur maximum) in the
U.S. have been proposed (in early 1989) and will probably be finalized
in 1990. The implementation date is proposed to be October 1993.
Electrostatic Agglomerator/Precipitators
Electrostatic precipitators have been used in a number of novel
approaches to particulate emissions control. An electrostatic
agglomerator has been used as the front end in an experimental system
for removing Diesel particles by cyclone collection developed by
Robert Bosch AG (Polach and Hagele, 1984), and a similar system has
been proposed by Kittelson and coworkers (1986).
The Bosch researchers developed a fairly compact agglomerator system
using serrated "spray disks" to" increase the corona discharge and
charging rate of the particles. This system was developed as a "pre-
agglomerator" for a cyclone collection system. A prototype of the
Bosch system was tested on a light-duty vehicle using the U.S. Federal
Test Procedure. The particulate mass collection efficiency on this
test was measured at 58 percent. The system was reported to give the
same muffling capability as a muffler, while occupying the space
required by 1 1/2 mufflers. It was also indicated that the system
resulted in a 3 percent increase in fuel consumption.
A similar system has been proposed by Kittelson and coworkers (1986) .
In this approach, particles are collected and agglomerated by a
multiplate electrostatic precipitator. Kittelson et al. discovered
that Diesel particles have a significant charge as they leave the
engine; therefore, no separate charging system is needed. After
sufficient particulate matter has built up on the plates, the
agglomerated particles begin to be reintrained by the exhaust gas.
They can then be collected downstream by a cyclone or other inertial
filter.
The key problem with both the Bosch system and the approach of
Kittelson et al. lies in disposing of the collected particulate
-42-
-------
matter. In the Bosch system, a high-particulate gas stream is
recycled into the engine and burned. This would be infeasible in a
turbocharged engine, due to the potential fouling of the compressor
and intercooler with particulate matter. All of the engine
manufacturers' reservations about EGR systems would also apply to this
system as well. Unless this problem can be resolved, the
electrostatic collection approach to particulate control is unlikely
to proceed beyond the laboratory.
NOx Reduction Techniques
Under appropriate conditions, NOx can be chemically reduced to form
oxygen and nitrogen gases. This process is used in modern closed-
loop, three-way catalyst equipped gasoline vehicles to control NOx
emissions. However, the process of catalytic NOx reduction used on
gasoline vehicles is inapplicable to Diesels. Because of their
heterogeneous combustion process, Diesel engines require substantial
excess air, and their exhaust thus inherently contains significant
excess oxygen. The three-way catalysts used on automobiles require a
precise stoichiometric mixture in the exhaust in order to function--in
the presence of excess oxygen, their NOx conversion efficiency rapidly
approaches zero.
A number of aftertreatment NOx reduction techniques which will work in
an oxidizing exhaust stream are currently available or under
development for stationary pollution sources. These include selective
catalytic reduction (SCR), selective non-catalytic reduction (Thermal
Denox(tm)), and reaction with cyanuric acid (RapReNox(tm)). However,
each of these systems requires a continuous supply of some reducing
agent such as ammonia or cyanuric acid to react with the NOx. Because
of the need for frequent replenishment of this agent, and the
difficulty of ensuring that the replenishment is performed when
needed, such systems are considered impractical for vehicular use.
Even if the replenishment problems could be resolved, these systems
would raise serious questions about crash safety and possible
emissions of toxic air contaminants. They will not be considered
further in this report, therefore.
-43-
-------
4. EMISSION CONTROL STRATEGIES FOR NEW DIESELS
Emission control requirements for heavy-duty Diesel engines have grown
increasingly stringent over the last 20 years, and will continue to do
so for at least the next decade. Technology for Diesel emissions
control, in consequence, has grown ever more sophisticated and
expensive, and emissions requirements have come to play an increasing
role in overall engine design. As a result one can define a number of
different levels of Diesel emission control, arranged in ascending
order of effectiveness, complexity, and cost.
Due to differences in local air-quality requirements, economic
capacities, and politically-determined tradeoffs between environmental
and economic goals, different levels of emissions control may be
appropriate for different jurisdictions. The focus of this section is
therefore not to lay out one preferred level of emissions control, but
to present a menu of differing control levels, in ascending order of
complexity, cost, and effectiveness. These range from simple limits
on excessive smoke emissions through sophisticated in-cylinder control
strategies to trap-oxidizer systems. Table 3 summarizes the emission
control levels considered.
Each of
these
is discussed
separately
below.
Table 3
Emissions
Control
Levels
Considered
Emissions
(g/BHP-hr)1
Smoke
Emissions Level
HC
CO
NOx
PM
A
B
C
1 Smoke Controls
25%
10%
40%
2 First-level controls
1.02
4.02
8.02
0.502
15%
7%
25"%
3 Current Technology
1.0
4.0
6.0
0.50
15%
7%
25%
4 Best engine-out technology 0.5
3.0
5.0
0.25
5 Best non-trap technology
0.2
0.5
5.0
0.15
6 Maximum Control
0.2
0.5
4.0
0.08
1 U.S. Heavy-duty transient test or similar, except as noted.
2 13-mode, steady state test.
The costs of controlling emissions from Diesel engines can include one
or more of the following:
Research and development costs;
Emissions certification and other costs of demonstrating
compliance with regulations;
-44-
-------
- Additional manufacturing cost for the engine;
- Additional maintenance costs; and
Increased fuel consumption.
In this section, we develop rough estimates of each of these
quantities (where applicable) for each of the emissions control levels
considered. Because of country-to-country variations (e.g. in duty
cycles, fuel costs, and interest rates) no attempt has been made to
combine these separate cost estimates into an overall lifecycle cost
or cost-effectiveness. These calculations are left to be carried out
by national authorities, using economic parameters appropriate to each
nation.
The costs of research and development, certification, and
manufacturing are paid directly by the engine manufacturer. Depending
on competitive conditions, the manufacturer may or may not be able to
recover these costs by increasing the price paid by the buyer. Due to
the variation in competitive conditions, we have not attepted to
estimate the effect of these cost elements on the ultimate price paid
-- the cost estimates shown are the added costs to the manufacturer.
The development cost estimates also reflect the cost to a manufacturer
undertaking such development now, with the benefit of the substantial
body of research that has already been done in the area. This is
probably less than the cost to the first manufacturers undertaking
development in this field.
The cost of additional maintenance and fuel consumption are paid by
the vehicle owner, not the manufacturer. The costs given for
maintenance are the estimated average annual cost to the vehicle
owner. In addition, we have estimated the percentage increase or
decrease in fuel consumption due to each emissions control level. Due
to the wide country-to-country variation in fuel consumption and fuel
prices, no attempt has been made to attach a monetary value to the
change in fuel consumption.
4.1 Smoke Controls
The first Diesel emissions regulations were established to limit
emissions of visible smoke. These regulations are still in force in
both the U.S. and the EEC. The advent of strict Diesel PM regulations
in 1991 and 1994 will render the U.S. standards more or less
irrelevant, however. As the first step in Diesel PM emissions
control, smoke limits are both inexpensive and very cost-effective.
Table 4 summarizes the emissions characteristics and costs of this
level of emissions control.
-45-
-------
Table 4
Estimated Emissions and Cost Impacts of
Emissions Control Level One: Smoke Controls
Emissions Standards
Estimated Emissions1
uncontrolled
with smoke control
reduction
% reduction
Emissions (g/BHP-hr)
HC CO NOx PM
1.0
1.0
0.0
0%
4.0
3.0
1.0
25%
11.0
11.0
0.0
0%
1.2
0.8
0.4
33%
A
20% 10%
Smoke
B C
35%
Economic Effects
Change in fuel rongumption f%)
Change in cost per vehicle (S)
Initial cost
Maintenance cost/year
Light-
Heavy
0%
0-100
0-20
Medium-
Heaw
0%
0-150
0-20
Additional cost per engine family (1000s of
Development/testing 0-100 0-100
Certification 10 10
Heavy-
Heavy
0%
0-150
0-20
0-100
10
1 U.S. Heavy-duty transient test or similar.
Technology
Technology for controlling excessive visible smoke emissions is well-,
developed. Excessive smoke is the result of injecting too much fuel
for the existing air supply and fuel-air mixing conditions. This
generally occurs under one of three conditions: full-load operation in
naturally-aspirated engines, transient acceleration in turbocharged
engines, and "lug-down"--operation at low engine speeds and high
loads.
Excessive full-load smoke in naturally-aspirated engines can be
prevented by limiting the maximum fuel quantity to a level which does
not exceed the smoke limit. This will normally mean a reduction in
engine power output. Alternatively, steps to improve engine
"breathing" and air-fuel mixing (e.g. increasing injection pressure)
may permit a reduction in smoke at the same power output.
Excessive lug-down smoke occurs as a result of poor mixing. In
present production fuel injection systems, the fuel injection pump is
geared to the crankshaft. A reduction in engine speed thus reduces
-46-
-------
the injection rate and thus (since injector orifice area is constant)
the injection pressure. At sufficiently low speeds, this impairs
atomization and mixing enough to cause heavy smoke at high loads.
This can be prevented by "shaping" the engine governor curve to reduce
the maximum fuel quantity permitted at low speeds. Although this
reduces emissions, it also impairs low*speed torque, and may thus
require more frequent gear shifting—or a larger engine--in some
applications.
Transient smoke emissions from turbocharged engines can be controlled
through the use of an acceleration smoke limiter. These are of two
types: simple time-delays and boost-pressure sensors, which limit the
maximum fuel quantity permitted as a function of turbocharger boost.
Both have the effect of reducing the maximum fuel quantity during
transient accelerations--thus limiting smoke from overfueling, but
also impairing acceleration performance to some degree. The boost
sensors are preferable, since they automatically adjust for variations
in turbocharger performance, and generally allow better acceleration
performance for the same smoke level.
Reliable measurements of the impact of Diesel smoke control measures
on total emissions are scarce. However, Weaver and Klausmeier (1988)
evaluated the impact of tampering with maximum fuel settings and
acceleration smoke limiters on emission-controlled engines. For a
heavy-duty truck engine, measurements by an engine manufacturer showed
that disabling the acceleration smoke limiter increased PM emissions
by roughly 50%. The increase in PM emissions from a naturally-
aspirated engine on which the maximum fuel limit had been increased
was also estimated at 50%. Assuming that these emissions were
characteristic of uncontrolled levels, the institution of smoke
controls should reduce emissions by about one-third. The effects of
smoke control will be heavily dependent on the specific duty cycle.
Where (as in bus operation) the duty cycle includes mostly full-power
accelerations, uncontrolled emissions are likely to be much higher,
and the effect of smoke controls will be proportionately greater.
Emissions Standards and Test Procedures
Smoke emissions standards can be defined in terms either of smoke
opacity (as in the U.S.) or in terms of a smoke "number" such as the
Bosch or Hartridge number, which measures the optical absorbtion per
unit volume. Since, for a given smoke concentration, opacity varies
as a function of the optical path length, the U.S. procedure requires
that the path length be specified. In the U.S. procedure, longer path
lengths (corresponding to greater exhaust pipe diameters) are used for
more powerful engines. This recognizes the fact that--at the same
soot concentration--an engine producing a greater volume of exhaust
creats a darker and thus more offensive smoke plume.
The smoke emissions test procedure used in the U.S. is performed on an
engine dynamometer, equipped with either a flywheel or dynamometer
controls to simulate vehicle inertia. The engine is operated through
a test cycle consisting of two accelerations (simulating acceleration
-47-
-------
from stop with a gear change) followed by "lugging down" at full load
from the rated power point to the maximum torque point. Average smoke
opacity is computed for each half-second interval during the cycle,
and the cycle itself is repeated three times.
Three smoke opacity values are computed: "A" or average acceleration;
"Bn, or lug-down peak; and "C" or acceleration peak. The A smoke
opacity is the average of the 15 highest opacity values in the
acceleration phases of each test cycle. The B opacity value is the
average of the 5 highest values from the lug-down phase of each cycle.
The C opacity, finally, is the average of the 1/2 second intervals
with the highest opacity in each test cycle. The maximum smoke
opacity levels permitted in the U.S. are A smoke, 20%; B smoke, 15%;
and C smoke, 50% opacity. Smoke of 15% opacity is clearly visible,
but "thin", and can still be seen through easily. Thi9 opacity level
is about the threshold at which smoke becomes noticeable to many
people. Smoke of 50% opacity appears "thick" and dark, and most
people would consider it visually offensive at this level.
The U.S. smoke standards have not changed since 1973, and appear to be
substantially more lenient than necessary. In reviewing smoke
certification data for the last 15 years, average A and B smoke for
naturally aspirated engines are both around 10%, while for
turbocharged engines average A smoke is around 15-18%, and B smoke
around 5-9%. Average peak smoke is about 25%, and seldom exceeds 35%
opacity. Therefore, smoke opacity standards of 20%, 10%, and 35% for
A, B, and C smoke, respectively, appear readily attainable.
Costs
The costs of establishing smoke emission limits would be relatively
small. Since such limitations are already in place in the largest
Diesel engine markets, nearly all engine manufacturers have already
developed the requisite technology and calibrations. The costs of
substituting a smoke-controlled model for a non-smoke-controlled model
would then consist of only the additional manufacturing costs, plus
the costs of testing and filing a certificate of compliance.
Manufacturing costs would be nil for naturally-aspirated engines, and
would range from around $50 to $150 for turbocharged engines. In the
case where the manufacturer chooses to upgrade technology in response
to the smoke limits (eg. by installing a turbocharger), the additional
manufacturing costs would be higher, but in this case, these costs
would be ascribable to improved engine performance, rather than the
smoke standard per se.
The costs of enforcing smoke standards for new engines would also be
fairly low. The major expense would be in setting up and maintaining
an office to keep track of emissions certification records.
Occasional spot-checking of actual compliance would also be desirable
(a smoke enforcement or inspection/maintenance program, if one were
established, would be useful in suggesting specific engine models to
be checked). This would require an engine dynamometer and associated
facilities, at a capital cost of $100,000 to $300,000, depending on
-48-
-------
land, building, and labor costs. Operating expenses for a moderate
spot-checking program for smoke emissions would probably be about
$100,000 per year.
Other Effects
The steps taken to limit smoke emissions will reduce acceleration
performance in both turbocharged and naturally-aspirated engines. In
addition, maximum steady-state power output from naturally aspirated
engines would be reduced somewhat (perhaps as much as 10-15%). To
recover the same vehicle performance, engine manufacturers would need
to increase engine displacement and/or improve engine breathing and
fuel injection to lower smoke at the same power level.
In-Use Maintenance and Enforcement Requirements
Changes in maximum fuel settings to limit smoke would not increase
maintenance requirements. Acceleration smoke limiters do require some
periodic maintenance, but this is minor. The cost of this added
maintenance is estimated at $20 per year. To the extent that
manufacturers upgrade technology to meet smoke limits (e.g. by
substituting in-line for distributor-type injection pumps), the
maintenance costs for the new technology may be higher.
The reduced engine performance resulting from smoke limitations will
create an incentive for vehicle owners and drivers to tamper with the
maximum fuel settings. To counteract this, manufacturers should be
required to make these settings tamper-resistant to the degree
practical (bearing in mind that there is a legitimate need to adjust
them in some cases, such as when recalibrating an injection pump).
Presently, manufacturers and injection service shops generally seal
these adjustments, in order to protect themselves against warranty
claims caused by tampering. Tampering is widespread nevertheless. A
smoke enforcement program and/or inspection/maintenance program aimed
at smoke opacity would be desirable to limit the amount of tampering
as well.
4.2 First-level Emissions Control
This first level of emissions control comprises only the most basic
emission reduction strategies. The technology required is comparable
to that required by pre-1988 U.S. Federal emissions standards, or the
present requirements of ECE R.49, with the exception that particulate
emissions are regulated as well. Using this basic technology,
emissions can be reduced significantly from uncontrolled levels, at
relatively low costs. The cost-effectiveness of this level of control
is excellent, therefore. Table 5 summarizes the estimated emissions
effects and costs of this level of control.
-49-
-------
Table 5
Escimated Emissions and Cost Impacts of
Emissions Control Level Two: First Level Controls
Emissions (g/BHP-hr)
Smoke
Emissions Standards1
Est. Transient Emissions2
Red, from previous level
g/BHP-hr2
percent
Red, from uncontrolled level
g/BHP-hr
percent
Economic Effects
Change in fuel consumption (%)
Change in cost per vehicle (S)
Initial cost
Maintenance cost/year
m
CO
NOx
SI
A
B
C
1.0
4.0
8.0
0.50
15%
7%
25%
0.8
2.5
7.5
0.65
0.2
20%
0.5
17%
3.5
32%
0.15
19%
0.2
20%
1.5
38%
Light-
Heavy
2%
0-50
0
3.5 0.55
32% 46%
Medium-
Heaw
1%
0-100
0
Additional cost per engine family (lQQQs of 5)
Development/testing 0-400 0-500
Certification 50 50
1 13-mode steady-state test.
2 U.S. Heavy-duty transient test or similar.
Heavy-
Heavy
0%
0-200
0
0-600
50
Technology
The technology needed for first-level emissions control is
approximately the lowest common level of Diesel engine technology
worldwide. The basic approach to compliance with these emission
levels involves retarding injection timing somewhat to reduce NOx, use
of low-sac injection nozzles to reduce HC and PM, and smoke controls
as discussed in the last section to limit both PM and visible smoke.
Some minimal optimization of engine breathing and the combustion
chamber may also be required, but is likely to have been performed in
any case in the search for better fuel economy. Due to the moderately
retarded injection timing, mechanical variable timing devices may be
beneficial in improving cold-start performance and light-load HC
emissions.
-50-
-------
Emissions Standards and Test Procedures
As this level of emissions control is intended to require only simple
and inexpensive technology, the use of expensive and sophisticated
test methods such as the U.S. Transient Test would be inappropriate.
So long as only simple mechanical emission controls are in use,
gaseous and steady-state particulate emissions can be controlled with
sufficient accuracy by a steady-state test procedure such as the 13-
mode test. Transient particulate emissions can then be kept within
reasonable limits by the application of stringent standards for
visible smoke. Where electronic controls are used, some additional
testing and regulations would be required to ensure that emissions
calibrations are applied across the entire range of engine operation,
not just in the vicinity of the steady-state test points.
The emissions standards recommended for this level of control
technology are shown in Table 5. The gaseous and particulate
emissions standards shown in the table are based on a steady-state
test procedure similar to the U.S. or ECG 13-mode tests. The
recommended NOx emissions standard of 8.0 g/BHP-hr represents a
significant reduction from uncontrolled levels, but not so much as to
have a significant effect upon fuel economy or PM emissions. The CO
and HC emissions standards are rather lenient, considering the
inherently low emissions of these pollutants from Diesel engines. The
steady-state PM standard of 0.6 g/BHP-hr is also considered readily
attainable using conventional technology. The recommended smoke
emissions standards of 15% A smoke, 7% B smoke, and 25% C smoke are
relatively stringent, but well within the present state of the art as
demonstrated in U.S. certification data.
Costs
Imposition of emissions controls at this level will probably result in
some upgrading of marginal fuel injection equipment to provide higher
injection pressures and variable injection timing, especially in the
heavier end of the Diesel market. Host engines would not require such
upgrades, however, since they are already sold in markets requiring at
least this level of control. The costs of these incremental changes
are estimated at up to $200 for heavy-heavy engines, but considerably
less for medium-heavy and light-heavy engines.
For most engine models, no development and only minimal testing would
be required to assure compliance with these standards. Engine models
not already being sold in jurisdictions where they are subject to
emission controls would require some development and testing effort,
however. Since only steady-state emissions and smoke tests are
required, the costs of this effort would be moderate--in the range of
$400,000 to $600,000, much of which would be for durability testing.
Certification, since it could be carried out using steady-state test
data, would not be very costly.
-51-
-------
Other Effects
The retarded fuel injection timing necessary to limit NOx emissions is
likely to have an adverse effect on fuel economy, especially in
naturally aspirated engines. For the relatively lenient 8 g/BHP-hr
standard shown in the table, however, this effect is unlikely to be
more than two or three percent. In heavier engines, increased fuel
injection pressure is likely to offset any adverse effects of NOx
control at this level. Retarded injection timing should help to
reduce combustion noise somewhat, while the use of low-sac nozzles
will help cut odor and aldehyde emissions as well as HC.
The strict smoke emissions standards recommended for this control
level would require manufacturers to calibrate their naturally
aspirated engines for reduced power output, and turbocharged engines
for reduced acceleration performance, compared to the more lenient
smoke standards recommended for the previous level. Engine
modifications to reduce FM emissions should also Improve smoke,
however, so the net effect is likely to be small.
In-Use Maintenance and Enforcement Requirements
Achieving these emission levels will require no new components beyond
those required for compliance with smoke limitations alone. The
detailed changes in calibration, injection nozzle design, combustion
chamber, etc. are not expected to result in maintenance requirements
different from those for smoke-controlled engines.
The degree of injection timing retardation required at this emissions
control level is not enough to provide significant incentive to tamper
with injection timing. The strict smoke controls will result in some
temptation to tamper with maximum fuel and acceleration smoke limiter
settings, however. Requiring tamper-resistant designs, and tying the
warranty coverage to maintaining intact seals on the adjustments
should reduce this tampering to some degree. As an additional
deterrent, however, an in-use smoke enforcement program and/or
inspection/maintenance program would be desirable.
4.3 Existing Technology
This level of emissions control is intended to reflect current (1988-
1990) Diesel engine technology in the U.S. While this level of
technology is required in the U.S., it is not unique to that country--
many premium engines in Europe and Japan are built to similar
standards, primarily for reasons of power output and fuel economy.
Table 6 summarizes the emissions levels achievable and the cost
estimated for this technology.
Technology
To attain this level of emissions control in the U.S., most
manufacturers have chosen to reduce NOx and PM via turbocharging with
-52-
-------
Table 6
Estimated Emissions and Cost Impacts of
Emissions Control Level Three: Existing Control Technology
Emissions (g/BHP-hr)
ac co is i
Emissions Standards1
Est. Transient Emissions1
Red, from previous level
g/BHP-hr1
percent
Red, from uncontrolled level
g/BHP-hr1
percent
Economic Effects
1.0
0.5
0.3
38%
0.5
50%
4.0
2.5
0.0
0%
1.5
38%
Light-
Heavy
6.0 0.50
5.5 0.4
2.0
27%
0.3
38%
5.5
50%
Change in fuel consumption -5 TO 3%
Change in cost per vehicle ($)
Initial cost 400-1500
Maintenance cost/year 50-200
Additional cost per engine family (1000s of S)
Development/testing 0-2000 0-2000
Certification 270 300
A.
15%
Smoke
B
7%
0.8
67%
Medium-
Heaw
-5 TO 3%
0-1600
0-200
C
25%
Heavy-
Heavy
0 TO 3%
0-1500
0-300
0-2000
350
1 U.S. Heavy-duty transient test or equivalent.
air-to-air or low-temperature air-water aftercooling. Further NOx
control to meet the standards is provided with moderately retarded
fuel injection timing. Increased fuel injection pressure and variable
injection timing are used to minimize the effect of the retarded
timing on fuel consumption and PM emissions. Further control of PM
emissions comes from detailed optimization of air swirl patterns and
combustion chamber geometry, use of low-sac or VCO injection nozzles,
limited control of lube oil consumption, and changes in engine
calibration to reduce transient smoke.
For the most part, this has involved only incremental changes to the
existing technology, rather than major redesign. The same
technological changes (except for the retarded injection timing) have
also been employed in many premium European and Japanese engines, for
reasons of power and fuel economy rather than emissions. Thus, the
effect of regulations in the premium segment of the engine market has
probably been fairly small. However, these regulations have resulted
in the near-disappearance of low-cost, naturally-aspirated, low-BMEP
-53-
-------
engines from the marketplace, to be replaced with more expensive
(albeit more efficient and powerful) turbocharged engines.
Emissions Standards and Test Procedures
At this level of emissions control, compliance requires fairly
sophisticated and expensive technology. The use of similarly
sophisticated and expensive technology for verifying compliance is
therefore not unreasonable. The emissions test procedure recommended
for this level of emissions control is the U.S. Heavy-Duty Transient
Test, or some other, similar test procedure involving real-time
simulation of transient operation. Emissions durability requirements
should be set to cover 70-80% of the engine's full "useful life"
before overhaul (recognizing that the increase in PM and HC emissions
due to lube-oil consumption near the end of an engine's life is beyond
the manufacturer's control).
Achievable emissions standards for this level of control are shown in
Table 6. The NOx standard of 6.0 g/BHP-hr is identical to the current
California NOx standard, and to the U.S. Federal standard for 1990.
This level was originally established for the 1988 model year, but a
lawsuit on procedural grounds resulted in a two-year delay at the
Federal level.
The achievable PM emissions standard shown in Table 6 is 0.50 g/BHP-hr
on the U.S. Transient Test. This is somewhat less than the current
U.S./California standard of 0.60 g/BHP-hr. Since that standard was
set in 1984, progress in PM emissions control has been unexpectedly
rapid, and the lower standard is now clearly feasible. Indeed, most
Diesel engines certified to the U.S. standards have demonstrated PM
emissions below 0.50 g/BHP-hr--considerably below, in some cases.
The HC and CO emissions standards shown in Table 6 are capping rather
than technology-forcing standards, as Diesel CO emissions are already
quite low, while HC emissions have been reduced considerably by the
changes made to reduce PM emissions. These standards are considerably
below the present U.S. standards, however, as those were designed with
gasoline engines in mind. The recommended standards are the same,
numerically, as the steady-state HC and CO standards recommended for
first-level emissions control, but the fact that they are measured on
a transient test cycle makes them (especially the HC standard)
considerably more stringent.
The smoke emissions standards shown in Table 6 are intended primarily
to maintain visual esthetics, rather than particulate emissions
control (although some additional PM control would likely result). At
this level of PM emissions, significant short-term acceleration smoke
could still occur. To ensure public support for an emissions control
program (as well as to limit high short-term exposures to Diesel PM
concentrations) it would be desirable to maintain at least the
preceding level of smoke emissions control.
-54-
-------
Costs
Costs of meeting these emission standards will vary, depending on the
level of technology already in use. For naturally aspirated engines
(found on most light-heavy and many medium-heavy vehicles) the costs
could include a turbocharger with air-air charge air cooling, and an
upgraded fuel injection system. In this case, the loss in fuel
economy would be nil--fuel economy could even improve. Selection of
less-expensive technologies would result in reduced power output and
some loss in fuel-economy. For premium engines already equipped with
air-air aftercooling and high-pressure injection, the additional
hardware costs would be small, but a small fuel economy loss would be
incurred.
The costs of developing engines to meet these emissions levels would
vary, depending on the degree of development that had been done
already. For engines sold in the U.S., the incremental development
costs would be small or zero, as these engines are already meeting
similar standards. For engines not sold in the U.S.,.however, these
costs could range to two million dollars or more, depending on the
degree of development required.
Depending on the specific test procedure chosen, certification costs
could also be small, as U.S. certification results might be
applicable. Otherwise, the costs of durability testing to full life,
with periodic emissions testing, would probably run about $300,000 per
engine family.
Other Effects
Depending on the emissions control strategy used, the effects of this
level of emissions control on fuel economy could range from a small
penalty (especially in heavy-heavy engines, which are typically
turbocharged and intercooled anyway) to a small benefit (where a
naturally aspirated engine is replaced with a more efficient
turbocharged/intercooled one). At any given technology level,
however, advancing injection timing is likely to improve fuel economy
several percent, thus introducing a motive for tampering.
Extensive use of turbocharging and charge air cooling will increase
maximum power levels, providing better vehicle performance, or making
it possible to use a smaller engine.
In-Use Maintenance and Enforcement Requirements
Incremental maintenance costs at this level would range from zero to a
few hundred dollars per year, depending on the approach taken.
Additional maintenance costs would be due to turbocharger and
aftercooler maintenance, and more expensive and precise fuel injection
systems.
-55-
-------
4.4 Near-term Technology (U.S. 1991 Standards1
This level of emissions control reflects essentially the current
state-of-the-art in Diesel engine technologies--those being developed
for compliance with the U.S. 1991 emissions standards. These
standards represent roughly the limits of feasibility for in-cylinder
emissions control alone (using currently available technology).
Table 7 outlines the achievable emissions levels and estimated costs
for this level of emissions control.
Technology
Recent progress in in-cylinder emissions control has been made
possible, in large part, by improved understanding of the Diesel
combustion process, and of the factors affecting pollutant formation
and destruction. Pollutant formation and destruction in the cylinder
are determined by the specific course of the Diesel combustion
Table 7
Estimated Emissions and Cost Impacts of
Emissions Control Level Four: Best In-Cylinder Control Technology
Emissions (g/BHP-hr) Smoke
m
CQ
NOx
EH _A_
B
C
0.50
4.0
5.0
0.25
0.35
2.5
4.7
0.20
0.15
0.0
0.8
0.20
30%
0%
15%
50%
Emissions Standards1
Est. Transient Emissions1
Red, from previous level
g/BHP-hr1
percent
Red, from uncontrolled level
g/BHP-hr1 0.65 1.5 6.3 1.00
percent 65% 38% 57% 83%
Economic Effects Light- Medium- Heavy-
Heavy Heavy Heavy
Change in fuel consumption (%) 0% 0% 0%
Change in cost per vehicle (S)
Initial cost 400-1000 600-1400 800-2000
Maintenance cost/year 50-200 100-300 150-400
Additional cost per engine family (1000s of S)
Development/testing 5000-10000 5000-10000 5000-10000
Certification 270 300 350
1 U.S. Heavy-duty transient test or equivalent.
-56-
-------
process. Modifying this process to minimize pollution involves a
complex multi-dimensional tradeoff between NOx, HC, and PM emissions,
fuel economy, power output, smoke, cold startability, cost, and many
other considerations.
Most engine manufacturers have followed a broadly similar approach to
in-cylinder control, although the specific techniques used differ
considerably from one manufacturer to the next. This typical approach
includes the following major elements:
O Minimize parasitic HC and PM emissions (those not directly
related to the combustion process) by minimizing nozzle sac
volume and reducing oil consumption to the extent possible.
O Reduce PM emissions at constant NOx by refining the turbo-
charger/engine match and improving engine "breathing"
characteristics. Many manufacturers are also experimenting
with variable-geometry turbochargers to improve the
turbocharger match over a wider speed range.
O Reduce PM and NOx (with some penalty in HC) by cooling the
compressed charge air as much as possible, via air-air or low-
temperature air-water aftercoolers.
O Further reduce NOx to meet regulatory targets by severely
retarding fuel injection timing over most of the speed/load
range. Minimize the adverse effects of retarded timing on
smoke, starting, and light-load HC emissions via a flexible
timing system to advance the timing under these conditions.
O Recover the PM increase due to retarded timing by increasing
the fuel injection pressure and injection rate.
O Improve air utilization (and reduce HC and PM emissions) by
minimizing parasitic volumes such as piston/cylinder head
clearance and piston top land volume.
O Optimize in-cylinder air motion through changes in combustion
chamber geometry and intake air swirl to provide adequate
mixing at low speeds (to minimize smoke and PM) without over-
rapid mixing at high speeds (which would increase HC, NOx, and
fuel consumption).
O Control smoke and particulate emissions in full-power
operation and transient accelerations through improved
governor curve shaping and transient smoke limiting (generally
through electronic governor controls).
Taken together, these changes amount to a complete redesign of large
portions of the engine and combustion system, and the costs are
correspondingly high.
In addition to these generally used approaches, a number of other
promising in-cylinder control techniques are under development by
-57-
-------
various manufacturers. These include variable air swirl devices for
improved control of in-cylinder air motion over a range of speeds;
fuel injection pumps with electronic control of the fuel injection
rate; proprietary technology to minimize the initial fuel injection
rate, thus reducing premixed burning and therefore NOx emissions; and
innovative supercharging technologies to minimize or eliminate
turbocharger lag. These technologies will be applied to the extent
that individual manufacturers consider them desirable or cost-
effective .
Emissions Standards and Test Procedures
The recommended test procedure for this level of emissions control is
the U.S. Heavy-Duty Transient Test or a similar test which more
accurately reflects actual operating patterns present Diesel vehicles
in the applicable jurisdiction. As discussed in Section 2.3, the
representativeness of the test cycle becomes an important concern at
these advanced levels of emissions control, due to the potential for
manufacturers to calibrate their control systems to exploit any
differences between testing and actual operation.
Emissions standards recommended for this emissions control level are
shown in Table 7. The recommended NOx and PM emissions standards are
identical to the U.S. 1991 levels. The present consensus among U.S.
manufacturers appears to be that these levels are difficult, but
achievable given low-sulfur fuel in 1991. The CO standard recommended
is the same as the earlier standards, and is intended only as a cap.
The HC standard of 0.5 g/BHP-hr also serves essentially a capping
function: to reduce particulate SOF to the required levels, most
manufacturers will have to control HC to well below 0.5 g/BHP-hr. At
this low level of PM emissions control, no separate smoke standard is
required.
Costs
The incremental costs of this level of emissions control over the
previous one are difficult to assess, due to lack of experience with
the technologies involved. The costs of electronic fuel injection
control--the major technological change--are estimated at around
$200-800 per unit, depending on what functions and sensors are
incorporated. Improved piston and liner materials, required to help
control oil consumption, will also increase costs significantly.
Variable-geometry turbocharging could add $500-800 to the cost of an
engine. Other technological changes would have their costs as well.
Altogether, likely ranges for the incremental costs of this emissions
control level are from around $400 to $1,000 in light-heavy engines to
around $800 to $2,000 in heavy-heavy engines.
Testing and development costs to meet this level of emissions control
will also be high. As noted above, what is required is essentially
the complete redesign of the engine and combustion system. Many new
technologies must be incorporated, and existing technologies extended
-58-
-------
beyond known limits. To assure reasonable durability and reliability
in use, engine manufacturers will need to test extensively, both in
the laboratory and in the field. The costs of this intensive testing
and development program are estimated at roughly five million to ten
million dollars per engine family.
For engines being developed for the U.S. 1991 emissions standards, a
significant portion of these costs has already been incurred, or would
be incurred in any event, in order to market the engines in the U.S.
Thus, for these engines, incremental development and testing costs
will be small. For engines not marketed in the U.S., however, the
full costs of development and testing would still apply.
Other Effects
The universal use of electronic engine control systems should add
significantly to engine performance and driveability, especially in
combination with variable-geometry turbochargers. Other desirable
functions such as road-speed governing, cruise control, driver
performance monitoring, etc. could easily be added as well. The more
sophisticated control strategies possible with these technologies
would probably have a net beneficial effect on vehicle fuel
consumption, even given the lower NOx limit. Vehicle performance and
driver satisfaction should be enhanced as well. Reliability
(especially of the electronic control systems) will be a significant
concern, however, until sufficient experience is accumulated with
onboard electronics in heavy-duty truck applications to provide some
assurance of durability.
In-Use Maintenance and Enforcement Requirements
The widespread use of computerized electronic engine controls in
light-duty vehicles has led to somewhat of a crisis in vehicle
maintenance in the U.S. Many mechanics, especially those working in
non-dealer shops, are simply unable to diagnose or repair these
systems. Similar problems can be expected with widespread use of
electronics in heavy-duty engines. Due to the greater reliance on
manufacturer-authorized service centers and generally superior
mechanic training in the heavy-duty field, this problem will probably
not be as severe, but it will certainly occur. Presently, all U.S.
heavy-duty engine manufacturers are engaged in intensive mechanic
training programs to combat this problem.
Due to the great number of new systems and technologies, overall
engine maintenance costs will probably increase considerably at this
level of control. Durability and reliability of the newly introduced
systems may not be as great as those of the older mechanical systems,
especially at first. In addition, engine durability with the low oil-
consumption levels required for PM compliance has not been fully
demonstrated. On the other hand, use of low-sulfur fuel and low-soot
combustion systems should reduce engine oil contamination and could
thus improve durability.
-59-
-------
The implementation of such sophisticated emission control techniques
may lead to correspondingly sophisticated attempts to tamper with
them. One problem that has already occurred in light-duty vehicles is
the marketing of replacement computer PROMs containing engine maps
optimized for optimum performance and/or fuel economy, rather than for
emissions compliance. Tamper-resistant designs, tie-ins with warranty
coverage, and extended warranty periods may help to reduce this
problem. An effective anti-tampering inspection program, or an I/M
program including gaseous emissions measurements would be helpful in
deterring such tampering as well.
4.5 Most Stringent Non-Trap Technology
This level of emissions control consists essentially of a Diesel
catalytic converter added on to the in-cylinder emissions control
technologies discussed in Section 4.4. This represents approximately
the maximum level of Diesel emissions control possible without the use
of a trap-oxidizer system. Table 8 summarizes the emissions effects
and estimated costs of this level of emissions control.
Technology
Technology for Diesel catalytic converters was summarized in Section
3.7. The use of precious-metal catalyzed particulate traps and
catalytic converters has been demonstrated to reduce particulate SOF
by 60-80%, HC by 40-80%, and CO by 50-90%. With the use of low-sulfur
fuel and low-soot engines to reduce fouling of the
converter,efficiencies near the upper ends of this range can be
expected. Low-sulfur fuel will also be required to limit sulfate
emissions, due to the conversion of gaseous S02 to particulate
sulfates.
Emissions Standards and Test Procedures
Test procedures for this emissions level would be the same as for the
best engine-out technology. Due to the presence of the catalyst,
however, special attention would have to be paid to the operating
conditions during durability testing, to ensure that they
realistically reflect the conditions the catalyst is likely to see in
use.
Feasible emissions standards for this technology are listed in
Table 8. HC and CO emissions standards have been set at levels which
will require an oxidation catalyst to attain. The PM limit reflects a
40% reduction from the limit achieavable with in-cylinder controls
alone. The NOx standard, on the other hand, has not changed from the
previous emissions control level, as NOx would not be affected by an
oxidation catalyst.
-60-
-------
Table 8
Estimated Emissions and Cost Impacts of
Emissions Control Level Five: Best Non-Trap Control Technology
Emissions (g/BHP-hr) Smoke
HC
CO
NOx
A B C
Emissions Standards1 0.20
1.0
5.0
0.15
Est. Transient Emissions1 0.12
0.5
4.7
0.12
Red. from Drevious level
g/BHP-hr1 0.23
percent 66%
2.0
80%
0.0
0%
0.08
40%
Red. from uncontrolled level
g/BHP-hr1 0.88
percent 88%
3.5
88%
6.3
57%
1.08
90%
Economic Effects
Light-
He aw
Medium-
Heaw
Heavy-
He aw
Chanee in fuel consumption (%)
0.5%
0.5%
0.5%
Chanee in cost oer vehicle fS")
Initial cost
Maintenance cost/year
150-300
20-40
300-500
40-80
500-800
40-80
Additional cost oer eneine familv
aooos
of $)
Development/testing 500-1000 500-1000 500-1000
Certification 270 300 350
1 U.S. Heavy-duty transient test or equivalent.
Costs
The costs of a Diesel catalytic converter system would be moderate.
Weaver and Klausmeier (1987) estimated the increase in new-vehicle
costs for this technology at around $186 for a light-heavy, $323 for a
medium-heavy, and $589 for a heavy-heavy-duty vehicle. Actual costs
would depend on the results of further development, but are likely to
fall within a range surrounding these values. The costs of research
and development on catalytic converter systems should also be fairly
moderate, in the range of $500,000 to one million dollars per engine
family. Certification costs would the same as for the in-cylinder
controls alone, or about $300,000 per family.
Other Effects
Added backpressure from the catalytic converter would tend to increase
fuel consumption slightly, but this would be offset to some extent by
-61-
-------
its muffling effect. Overall, an increase of about 0.1-0.5% in fuel
consumption appears likely. Beneficial effects of the catalytic
converter would include reductions in aldehydes, odor, and mutagenic
compounds, as well as regulated emissions of HC, CO, and PM.
In-Use Maintenance and Enforcement Requirements
Since catalytic converters would have little adverse impact on vehicle
performance, there would be little incentive to tamper with them.
Periodic replacement would be necessary, however, in intensively used
vehicles such as long-distance trucks and buses. To ensure that
replacements are performed when necessary, it would be desirable to
add some measurement of catalytic converter efficiency to the
inspection/maintenance program required for in-cylinder emission
controls. Free catalyst replacements (in effect, incorporating the
cost of the replacement into the initial price) and incentives for
having it performed could also help to ensure replacement when needed.
4.6 Maximum Emissions Control: Catalytic Trap-oxidizers
This emissions control level would consist of a catalytic trap-
oxidizer system, added onto the advanced in-cylinder emissions
controls discussed in Section 4.4. As such, it represents the maximum
level of emissions control achievable with a Diesel engine, using
reasonably well-demonstrated technology. Table 9 summarizes the
emission levels achievable and the estimated costs for this level of
emissions control.
Technology
The current state of trap-oxidizer technology was discussed in
Section 3.7. For maximum effectiveness in controlling particulate SOF
(as well as gaseous emissions) it would be desirable to have a
precious-metal catalyst coating on the trap or a separate catalytic
converter. Regeneration, using one of the active or semi-passive
techniques described in Section 3.7, would be under the control of the
engine's electronic control system.
Emissions Standards and Test Procedures
The emissions test procedure for this control level would be
essentially the same as for in-cylinder controls alone--the U.S.
Transient Test or equivalent. Modifications to the test procedure
would be needed to ensure the inclusion of regeneration emissions on a
pro-rata basis. Special attention to the durability test cycle would
also be required, to ensure that the conditions experienced by the
trap in durability testing are reasonably representative of those in
actual service.
-62-
-------
Achievable emissions standards for this level of control are listed in
Table 9. The HC and CO standards of 0.2 and 1.0 g/BHP-hr,
respectively, are the same as those for the catalytic converter
technology, and are intended to force the use of an oxidation
catalyst. The PM standard of 0.08 g/BHP-hr is somewhat lower than the
U.S. 1994 standard level, but should be readily achievable by a
catalytic trap-oxidizer on low-sulfur fuel. The NOx standard of 4.5
g/BHP-hr is somewhat lower than for the in-cylinder control case,
reflecting the change in emissions tradeoffs with the trap-oxidizer
present. Although retarding injection timing to reduce NOx tends to
increase soot and PM emissions, the trap is extremely efficient in
collecting soot. Thus, a lower NOx emissions level is achievable
without significantly increasing PM emissions.
Table 9
Estimated Emissions and Cost Impacts of
Emissions Control Level Six: Maximum Emissions Control
Emissions (g/BHP-hr)
Smoke
HC
CO
NOx ŁJ4
A B C
Emissions Standards1
0.20
1.0
4.5 0.08
Est. Transient Emissions1
0.07
0.5
4.2 0.04
Red. from Drevious level
g/BHP-hr1
percent
0.05
42%
0.0
0%
0.5 0.08
11% 67%
Red. from uncontrolled level
g/BHP-hr1 0.93
percent 93%
3.5
88%
6.8 1.16
62% 97%
Economic Effects
Light-
Heaw
Medium-
Heaw
Heavy-
He aw
Chanee in fuel consumution
(%)
3%
2%
1.5%
Chanee in cost Der vehicle
($*)
Initial cost
Maintenance cost/year
400-1000 800-1400
40-80 100-200
1200-2000
200-300
Additional cost oer engine
familv
(1000s
of
Development/testing
Certification
3000-10000 3000-10000 3000-10000
300 340 400
1 U.S. Heavy-duty transient test or equivalent.
-63-
-------
Costs
The capital costs of different trap-oxidizer system designs were
estimated by Weaver and Klausmeier (1987) at around $400 to $600 for
light-heavy vehicles; $800 to $1,000 for medium-heavy vehicles; and
$1,200 to $1,500 for heavy-heavy vehicles. Other cost estimates by
engine manufacturers have ranged as high as $5,000 for a
heavy-heavy-duty vehicle, but these must be viewed with some
skepticism. Our cost estimates in Table 9 basically follow Weaver and
Klausmeier, but include a higher upper range to reflect the
possibility of higher costs in some specialized applications.
System development and testing costs for trap-oxidizer systems would
be very high, due to the novel nature of the technology and the need
to demonstrate great reliability under a wide variety of operating
patterns. Development and testing costs are roughly estimated at
around three to ten million dollars per engine family, depending on
the complexity of the application and the availability of "packaged"
trap-oxidizer system designs from outside vendors. Certification
costs would be similar to but somewhat higher than those for in-
cylinder controls alone, due to the higher cost of multiple tests to
account for regeneration emissions.
Other Effects
Depending on their design, trap-oxidizer systems would increase fuel
consumption by between 0.5 and 5%. Their size and potential fire
hazard would impose additional problems for vehicle designers. Power
output might also be reduced somewhat by backpressure from the trap.
The trap's effectiveness in collecting soot emissions might make it
possible to relax somewhat on air-fuel ratio control, however,
allowing for increased power output and better acceleration.
In-Use Maintenance and Enforcement Requirements
Maintenance costs for trap-oxidizers are likely to be significant.
Given the complexity of the system, the need for periodic inspection
and checking, replacement of trapping elements in intensively-used
vehicles, and the inevitable system failures could add up to several
hundred dollars per year. This would depend on the vehicle size and
intensity of usage--some rough cost estimates are shown in Table 9.
Trap-oxidizers are likely to encounter resistance from the user
community, due to their (perceived or actual) disadvantages for fuel
economy, power output, and safety. As the trap would also be easy to
remove (or bypass, in the case of active systems using bypass
regeneration) tampering with these systems could be expected to be
widespread in the absence of an effective inspection/maintenance
program. For bypassable traps, even a scheduled I/M program might be
inadequate, since it would be easy to reverse tampering with the
bypass valve immediately before the inspection, restoring it to its
-64-
-------
tampered condition afterward. Thus, some form of in-use anti-
tampering and/or smoke inspection would be required to supplement a
scheduled I/M program.
4.7 Advanced Technologies
A number of advanced engine technologies hold promise for additional
emissions control beyond the levels discussed above. These include
low-heat rejection engines, turbocompounding, closed-loop fuel
injection controls, advanced combustion technologies, exhaust gas
recirculation, and advanced aftertreatment technologies.
Low heat rejection engines
Considerable effort is being devoted to the development of low heat
rejection Diesel engines. The major benefit of such engines would be
the elimination of the engine cooling system, with its attendant power
losses and reliability problems. Another likely benefit would be the
elimination of lubricating oil for the piston/cylinder contact, and
thus a major source of the oil contribution to particulate SOF.
Higher in-cylinder temperatures would help to reduce HC emissions, but
NOx would probably increase some (this might be offset by a reduction
in ignition delay, however). Some test data suggest that soot
emissions might increase, however, due to the higher charge
temperature. Overall, it appears that the technology of low heat
rejection engines is too immature for any final judgement to be made
concerning its emission effects.
Turbocompounding
Several manufacturers are developing turbocompound engines for use in
line-haul trucks (Holtman, 1987). The primary advantage of these
engines is their increased fuel efficiency, due to their ability to
extract work from some of the waste exhaust heat. Other things being
equal, this should result in a small corresponding decrease in
emissions. Another possible emissions-related advantage would be the
potential to bypass the power recovery turbine at low engine speeds,
thus increasing the pressure drop across the compressor turbine and
thus providing more power to the compressor. This should markedly
improve low-speed smoke emissions and transient response.
Fuel injection rate shaping
Technology for reducing the fuel that bums in the premixed burning
phase is not yet completely demonstrated. Some research results,
however, suggest that NOx emissions might be reduced by 30-50% using
this technique, with little or no adverse impact on PM emissions or
fuel consumption. This can be expected to be an area of very active
development in the next few years. If successful, it could make
possible the attainment of PM and HC emissions comparable to those of
-65-
-------
levels four through six above (Tables 7 to 9), in combination with NOx
levels of three g/BHP-hr or less.
Exhaust gas recirculation
The use of EGR could also have a major impact on NOx emissions, with
minimal degradation in fuel economy and PM. Researchers at Ford were
able to reduce NOx emissions from a light-duty DI engine by more than
50 percent (to about two g/BHP-hr in light-duty test cycle) using an
optimized EGR schedule, while maintaining PM emissions comparable to
light-duty IDI engines without EGR. EGR would be less effective in a
heavy-duty engine, due to their more heavily loaded test cycles. The
available data suggest that NOx emissions in the 3-4 g/BHP-hr range
might be achievable without greatly degrading fuel economy or PM
emissions, however.
The major drawback to EGR is its detrimental effect on engine
durability and maintenance costs. The reasons for this effect are not
well understood, however. Some data (and some manufacturers) suggest
that the sulfur dioxide in the exhaust is the major culprit, while
others focus on the role of recycled carbon particles. If sulfur is
the major problem, its effects could be greatly reduced by the use of
low-sulfur fuel--a measure which will probably be required for the
1994 PM standard in any case. If recycled carbon is the problem, its
effects should be virtually eliminated by taking the recycle stream
after the trap-oxidizer. Thus, EGR's effects on engine wear might be
greatly alleviated by emission control measures undertaken for other
reasons. More research is needed to establish the real causes and
potential cures for excessive engine wear due to EGR.
Closed-loop control
Technology for feedback control of heavy-duty Diesel injection timing,
fuel quantities, etc. is still in its infancy. By actually measuring
outputs such as injector needle lift and rates of change in engine
angular velocity, feedback control systems could help to balance fuel
injection quantities between cylinders, compensate for wear or
improper setting of the injection linkage, etc. etc. The major
benefit of such technologies is likely to come from a reduction in in-
use emissions deterioration, as the control system will actually be
able to compensate for wear. Some reduction in the "slippage" between
the performance of certification engines and those actually produced
on the assembly line can also be expected.
Advanced combustion technologies
Diesel NOx and soot formation are inherent in the heterogeneous
combustion process. Some research (Wood, 1988) is now focussed on
ways to make Diesel combustion more homogeneous (i.e. more like that
in an Otto-cycle engine) without giving up the fuel economy and other
advantages of the Diesel. As discussed in Section 6, lean homogeneous
-66-
-------
combustion can achieve very low emissions of both NOx and PM.
Techniques for approaching this condition using Diesel or a similar
fuel are still research curiosities rather than the subject of serious
development efforts, however.
Advanced aftertreatment
Beyond trap-oxidizer systems, potential aftertreatment controls
include electrostatic collection techniques and various approaches to
reducing NOx emissions to harmless nitrogen and oxygen. Currently,
all known NOx reduction techniques require that a chemical reductant
such as ammonia or cyanuric acid be supplied separately. This makes
them infeasible for use in motor vehicles, except under special
circumstances. A system to use Diesel fuel directly as a reductant,
or the vise of direct electrochemical reduction, could make
aftertreatment NOx controls feasible.
-67-
-------
5. EMISSION CONTROL STRATEGIES FOR DIESELS ALREADY IN USE
Programs to control ln-use emissions form an important complement to
new-vehicle emissions standards. Programs such as smoke enforcement
and inspection/maintenance may be required to ensure that the
anticipated emissions benefits of new-vehicle control technologies are
not lost through the effects of poor maintenance and/or tampering with
emission controls. Proper maintenance, encouraged by an effective
inspection/maintenance program, can go far to limit emissions even
from uncontrolled vehicles. Retrofit programs to install emission
controls on existing vehicles can help reduce emissions over the short
term, contributing to more rapid improvement in the environment.
Finally, transportation control measures may also contribute to
improved air quality in specific areas.
5.1 Maintenance
Even in the absence of specific emissions controls, well-maintained
and properly-adjusted Diesel engines usually emit only relatively
small quantities of unburned hydrocarbons and CO, and moderate amounts
of particulate matter. Incorrect adjustment, poor maintenance, or
excessive wear can increase emissions of these pollutants many-fold,
however. NOx emissions, in contrast, are not very sensitive to
maintenance conditions.
In extreme cases, PM and HC emissions in malfunctioning Diesel engines
may be increased by a factor of 10 or 15 over the levels experienced
with proper maintenance. Emission-controlled engines, since they have
lower emissions to start with, may experience an even greater
percentage increase. While the most severe emissions problems tend to
degrade power output and fuel consumption, and are thus likely to be
fixed eventually, this is by no means certain. Increases of up to a
factor of two in HC, CO, and PM have be measured even in the absence
of a significant effect on power or fuel consumption (Ullman et al.,
1984).
The engine systems most likely to contribute to high HC, CO, and PM
emissions are the air induction system (including the turbocharger),
and the fuel injection system. Problems in these areas are generally
due to neglect of routine maintenance, and can be corrected fairly
cheaply. Tampering and improper adjustment of engine controls such as
governor settings, maximum fuel limits, and acceleration smoke
limiters may also result in substantially increased emissions. Major
engine mechanical problems and/or excessive wear may also result in
very high emissions (especially lube-oil emissions).
-68-
-------
The effects of poor maintenance practices on Diesel emissions have
received little quantitative study. One study performed for the
California Air Resources Board (Weaver and Klausmeier, 1988) attempted
to estimate the impact of common emissions-related defects by
estimating the frequency of occurrence of each defect in the truck
fleet and the emissions impact of vehicles having that defect. Defect
frequencies and emissions impacts were different for different classes
of trucks, and for different levels of emission control. Table 10
shows the estimates developed for medium-heavy duty trucks built to
U.S. Federal pre-1988 emissions standards. This emission control
technology is also representative of most European truck engines built
to comply with ECE regulation R.49.
Fuel injection problems
As Table 10 indicates, the most important causes of excess PM and HC
emissions are fuel injection system problems such as leaky, worn, or
clogged injector nozzles. Injector problems range in severity from
minor to very severe. Minor injector problems are likely to go
unrepaired until the next routine maintenance interval, even in well-
maintained fleets. These are not severe enough to impair engine
operation noticeably, even though emissions are increased noticeably.
Moderate injector problems could include clogging with deposits or
significant wear. These problems are severe enough to degrade engine
performance and fuel economy noticeably, and are typically accompanied
by considerable excess smoke. These problems would typically be
repaired at the first convenient opportunity in a well-maintained
fleet. Where
maintenance is less scrupulous, however (that is, in most vehicle
fleets worldwide), they are likely to go unrepaired for some time.
Severe injector problems often involve mechanical damage to the
injector, or else extreme wear or deposit fouling. These problems are
serious enough to degrade engine performance and fuel economy
significantly, and are generally characterized by very heavy smoke.
Problems of this magnitude should cause all but the most marginal
operators to pull the vehicle out of service for repairs. Based on
visual observations, however, such vehicles may constitute as many as
2 to 3 percent of trucks and buses on the road.
Particulate emissions in excess of 50 g/mile have been measured in
city buses with severe injector problems (unpublished New York City
Dept. of Environmental Protection data, summarized in Weaver and
Klausmeier, 1988). The buses were nonetheless driveable, and capable
of performing in service.
Improper fuel injection timing may also result in significant excess
emissions in some circumstances. In fuel injection systems with
mechanical controls, the angular relationship between the fuel pump
and the pump driveshaft is very important. An error of a few degrees
in either direction can significantly increase NOx emissions (if
timing is advanced) or PM emissions (if it is retarded). Where
-69-
-------
Table 10
Estimated Effects of Tampering and Malfunctions on Heavy-Duty Diesel Emissions
Oxides of
Nitrogen
Unbumed
Hydrocarbons
Particulate
Matter
Fuel
Consumption
Type of
Pet.
Ind.
Fleet
Ind.
Fleet
Ind.
Fleet
Ind.
Fleet
Defect
Aff.
Veh.
Avg.
Veh.
Avg.
Veh.
Avg.
Veh.
Avg.
Timing Advanced
10%
50%
5.0%
20%
2.0%
10%
1.0%
0%
0.0%
Timing Retarded
6%
-20%
-1.2%
-10%
-0.6%
30%
1.8%
7%
0.4%
Minor Injector Problems
20%
0%
0.0%
10%
2.0%
35%
7.0%
2%
0.4%
Moderate Injector Problems
15%
-5%
-0.8%
150%
22.5%
200%
30.0%
5%
0.8%
Severe Injector Problems
5%
-10%
-0.5%
500%
25.0%
500%
25.0%
10%
0.5%
Smoke Limiter Misset
18%
0%
0.0%
0%
0.0%
20%
3.6%
1%
0.2%
Smoke Limiter Disabled
15%
0%
0.0%
-20%
-3.0%
50%
7.5%
2%
0.3%
Maximum Fuel High
14%
10%
1.4%
0%
0.0%
30%
4.2%
2%
0.3%
Clogged Air Filter
23%
0%
0.0%
0%
0.0%
50%
11.5%
2%
0.5%
Wrong/Worn Turbo
10%
0%
0.0%
0%
0.0%
40%
4.0%
1%
0.1%
Interoooler clogged
1%
10%
0.1%
-20%
-0.2%
40%
0.4%
2%
0.0%
Other Air Problems
14%
0%
0.0%
0%
0.0%
40%
5.6%
1%
0.1%
Engine Mechanical Failure
2%
-10%
-0.2%
200%
4.0%
150%
3.0%
7%
0.1%
Excess Oil Consumption
8%
0%
0.0%
300%
24.0%
150%
12.0%
0%
0.0%
% All Defects Combined
3.9%
75.2%
140.4%
3.7%
-------
stringent NOx controls are in place, injection timing is likely to be
significantly more retarded than would be optimal for performance and
fuel economy. In this case, vehicle owners and drivers will be
tempted to tamper with the injection timing to improve performance.
Air-fuel ratio problems
Operating with the air-fuel ratio below the smoke limit is another
common cause of excessive PM emissions. This may occur as the result
either of tampering or of neglect of routine maintenance. In
naturally aspirated engines, the maximum power level Is generally
smoke-limited. To increase the power output from the engine, it is
very common for owners of such engines to "turn up" the maximum fuel
stop to increase power. This results in a very large increase in
black smoke and particulate emissions at full load.
A similar problem occurs with acceleration smoke limiters used on
turbocharged engines. By limiting maximum fuel flow during engine
transients, they act to limit smoke emissions. But this also reduces
the engine power output during the transient, and thus impairs the
vehicle's acceleration performance. Smoke limiter settings are
commonly adjustable, and it is very common for vehicle owners or
drivers to adjust them to provide better acceleration, at the cost of
a substantial increase in smoke and PM emissions.
Other causes of air-fuel ratios below the smoke limit include dirty
air filters, which restrict the airflow to the engine; worn-out or
incorrect turbochargers; intercoolers clogged with deposits; and
miscellaneous air intake problems such as air and exhaust leaks in
turbocharged systems, collapsed air intake hoses, crimped exhaust
pipes, etc.
Engine mechanical problems
Vehicles suffering from excessive emissions due to engine mechanical .
failures or excessive oil consumption make up a fairly small fraction
of the vehicle fleet. The increase in emissions from these vehicles
is so large, however, that they represent a significant fraction of
total Diesel emissions. Unlike the other problems leading to high
emissions, these defects are very expensive to correct--typically
requiring that the engine be rebuilt, at a cost of $3,000 to $10,000.
Total impact
The emissions impacts of different types of defects are not
necessarily additive--some defects can interact with others to create
an even greater increase in emissions. For this reason, Weaver and
Klausmeier (1988) estimated the combined impact of all defects by
multiplying the fleet-average impacts of defects in each group. The
results, shown in Table 10, indicate that average HC emissions in this
vehicle class are approximately 75% higher and PM emissions 140%
-71-
-------
higher than they would be if all vehicles were well-maintained.
Similar estimates for heavy-heavy and for light-heavy-duty trucks
showed similar effects. For advanced-technology emission-controlled
vehicles, the effects of tampering and poor maintenance were estimated
to be even larger in percentage terms, although the actual increase in
g/BHP-hr was projected to be smaller.
Figure 7, reproduced from Weaver and Klausmeier (1988), shows the
estimated impact of poor maintenance and tampering with emissions
controls on total Diesel emissions in California through the year 2000
(in the absence of an inspection/maintenance program). As these
figures show, baseline PM and HC emissions (those which would be
experienced if all vehicles were well maintained) are projected to
decrease markedly in the 1990s, as engines built to the stringent
Califomia/U.S. emissions standards replace older models. Excess
emissions of these pollutants are also expected to decline, but by a
much smaller percentage.
Due to the scarcity of data, the estimates of emissions impact and
total emissions developed by Weaver and Klausmeier (1988) include
considerable uncertainty (estimated as -30 to +70 percent for the
excess emissions). Despite this range of uncertainty, it is clear
from the foregoing that poor maintenance has a very large impact in
increasing Diesel HC and PM emissions in California. Since there is
no reason to believe that trucks in other areas receive significantly
better maintenance than those in California, the same conclusion is
doubtless valid elsewhere as well.
5.2 Smoke Enforcement
One approach to reducing the incidence of the most severe emissions
problems in the heavy-duty vehicle fleet is a smoke enforcement
program, in which traffic police or other designated personnel
identify and cite heavily smoking trucks on the highways. Due to the
public offense from excessive Diesel smoke, many jurisdictions have
established such programs.
Most of the programs implemented to date have relied on visual
estimates of smoke opacity, using the Ringelmann scale or some
analogous measure. These measures have generally proven very
difficulty to enforce, due to the need for properly trained observers
and good observing conditions. Smoke opacity estimates based on
Ringelmann Charts have been shown to vary considerably from chart to
chart and observer to observer (Engine Manufacturer's Association,
1983). In some cases, lack of credence in such a "naked-eye"
measurement has led judges to dismiss such citations.
Another problem contributing to the ineffectiveness of most smoke
enforcement programs has been a lack of vigorous enforcement. Smoke,
often, is viewed as less significant than speeding or other traffic
violations, and receives lower priority. Lack of suitable test
procedures, failure to recognize the differences in Diesel smoke
emissions in different operating modes, and excessively lax failure
-72-
-------
Figure 7
Estimated Impact of Poor Maintenance and Tampering
with Emission Controls on Heavy-Duty Diesel Emissions
in California 1985-2000
(Source: Weaver and Klausmeier, 1988)
ExCISS
9U 1 SAC 1907 19M '981 1990 1991
Oxides of
Nitrogen
Unburned
Hydrocarbons
Particulate
Matter
<993 1988 19A7 <9U 19S9 1990 1991 1992 1993 '99* 1993 1999 >997 199A 1999 2000
-73-
-------
criteria have also led to frustration on the part of those charged
with enforcing the programs, and thus to their failure.
Smoke enforcement programs could potentially be an effective spur to
improved maintenance, and thus to reduced Diesel emissions. Proper
program design is essential, however. Recommended elements for an
effective smoke enforcement program are the following.
1. Dedicated enforcement personnel. Especially when a program is
just starting, it is important that the personnel assigned to
smoke enforcement have that as their sole (or at least
primary) responsibility. This will allow for concentration of
effort, for effective training in smoke observation, and for
intensive practice to maintain the skills developed.
2. Suitable test procedures. Although vehicles to be subjected
to enforcement action can be picked out by eye on the road,
the variety of possible operating conditions makes it
impractical for a single standard to apply to all cases. If
the vehicle is made to undergo a standard test, a single
standard becomes feasible. Weaver and Klausmeier (1988)
evaluated a number of short roadside opacity tests, ultimately
recommending one based on full-power acceleration from a stop.
3. Instrumental backup. Much experience has shown that observers
can be trained to judge smoke opacity accurately and
repeatably. Nonetheless, such judgements often lack
credibility in court. Such skills also atrophy quickly if not
used. The use of an inexpensive end-of-stack opacity meter to
confirm an officer's judgement in doubtful or disputed cases
can significantly improve the credibility of the program.
4. Stringent standards. Standards should be set at the lowest
level that will ensure that properly maintained and adjusted
vehicles will pass. A truck need not be continuously belching
opaque black smoke to be emitting far more particulate matter
than it should. The actual standard levels feasible will vary
depending on the level of particulate emissions control in
place. As a general rule, however, no Diesel engine should
emit visible smoke in straight and level cruise at less than
full power; and Diesels subject to level 4 or higher controls
(U.S. 1991 standards or equivalent) should emit no more than a
trace of visible smoke under anv condition.
5. Effective public relations. The public, and especially the
trucking community, should understand the operation of the
program and the reasons for adopting it, and should be
convinced of its fairness. Truck owners should be provided
with adequate time and guidance to bring their vehicles into
compliance before the program goes into effect. Initial
enforcement efforts should concentrate on ticketing the worst
offenders, letting marginal cases go for the time being.
-74-
-------
5.3 Inspection/Maintenance Programs
Inspection and maintenance (I/M) represents the next step beyond an
on-road smoke enforcement program for reducing in-use emissions from
heavy-duty vehicles. As discussed in Section 4, some form of I/M
program is likely to be essential to reap the full benefits of
emission controls such as smoke limiters and trap-oxidizers.
Design of an inspection and maintenance program for heavy-duty Diesel
vehicles presents many difficult issues. Because of the differences
in technology, ownership, and operating patterns, existing I/M
programs for light-duty vehicles may not be a good model for heavy-
duty Diesel I/M. Other existing enforcement programs aimed at heavy-
duty trucks (such as truck weight and safety enforcement programs)
should be considered as well.
To reduce emissions, while minimizing the burden on vehicle owners,
the primary goals of a heavy-duty Diesel I/M program should be the
following:
O deter tampering with emission controls;
0 detect tampering which is not deterred, and require that it be
corrected;
O identify gross-emitting vehicles, and require that they be
repaired; and
O encourage proper maintenance and awareness of the importance
of emission controls in the bulk of the heavy-duty Diesel
fleet.
Program Designs
One approach to designing a heavy-duty Diesel I/M program is to model
it after existing light-duty I/M programs. In these programs,
vehicles owners are required to present them for periodic inspections
and emissions measurements, generally on an annual basis. Inspections
may be performed either at a centralized, government-operated
facility, or at specially licensed garages.
For heavy-duty Diesel vehicles, this type of program has several
drawbacks. The costs of truck and driver time lost in bringing the
vehicle in for inspection are likely to be significant (unless the
emissions test can be combined with a required safety inspection). In
addition, heavy-duty vehicles travel much greater distances annually
than passenger cars, and require correspondingly more frequent
maintenance. An annual inspection will thus have less impact on the
incidence of emissions-related defects in the vehicle fleet.
Another important drawback to periodic inspections is that the
scheduled inspection will be less effective in deterring tampering
than one which is not predictable in advance. Many forms of tampering
-75-
-------
with heavy-duty emissions controls (e.g. maximum fuel stops,
acceleration smoke limiters, injection timing, and trap-oxidizer
bypasses) are reversible. If the inspection time is known, the
tampering can be restored to its original condition immediately before
the inspection, then returned to its tampered condition immediately
afterward.
An alternative to periodic inspections is a program of in-use
inspection for heavy-duty vehicles. Many jurisdictions already
perform truck weight checks and safety inspections on this basis, and
the addition of an emissions check and/or anti-tampering inspection to
these programs would be straightforward. Random anti-tampering
inspections (possibly in conjunction with smoke enforcement) could
also serve as a useful adjunct to a periodic inspection/maintenance
program.
Existing Programs
Inspection/maintenance programs for Diesel vehicles are already
operating in a number of jurisdictions. In West Germany and several
U.S. I/M programs, Diesel trucks are subject to the same idle tailpipe
emission tests used for passenger cars, or to an idle smoke opacity
test. These tests are generally justified on grounds of "fairness"--
since gasoline vehicles are required to be tested, Diesels should too.
Such measurements provide no meaningful test of actual Diesel
emissions performance, however, and are essentially a waste of time
and money. Failure rates in these programs are extremely low, as only
a grossly malfunctioning engine could possibly fail.
A few jurisdictions have established meaningful Diesel I/M programs.
One, established in the U.S. cities of Tucson and Phoenix, Arizona,
measures Diesel smoke opacity under peak torque conditions on a
chassis dynamometer. The maximum opacity level is 20%, or 5% over the
U.S. smoke emissions standard. The failure rate in this program is
about 7.5%.
Another successful program, in Santiago, Chile, is combined with a
periodic safety inspection. Trucks are required to be tested every
seven months, and buses every three months. Emissions testing in this
program is performed by placing the drive wheels on a set of free
rollers, placing the engine in gear, and then loading the engine down
to the peak torque point using the vehicle's service brake. Smoke is
measured using a Bosch smokemeter. To minimize potential problems
with reversible tampering, smoke inspection facilities are required to
seal the maximum fuel adjustment after the vehicle passes.
The Chilean program is a good example of the potential benefits of an
effective I/M program, as well as some of the potential problems.
Buses in Santiago are privately owned and operated, and maintenance
practices are generally poor. However, Santiago's buses--unlike those
in nearly every other third-world city--do not emit excessive
quantities of black smoke. While about 25% of the buses emit visible
smoke, the smoke is bluish or grayish--characteristic of unburned
-76-
-------
lubricating oil, rather than soot. The Bosch smoke test method used
in the Chilean program is effective in measuring soot emissions, but
is insensitive to oil smoke. It thus appears that the program is
successfully controlling only those smoke emissions measured in the
test. Chile has begun a the transition to a light-transmission
opacity measuring technique, which will allow oil smoke to be detected
as well. This will presumably result in decreased oil smoke
emissions.
Test Procedures
The Diesel emissions of greatest concern from an I/M standpoint are
particulate matter and hydrocarbons. Where (as in the U.S.) NOx
emissions controls are stringent enough to affect fuel economy
significantly, measurement of NOx emissions as well may be needed to
deter tampering. While direct measurement of Diesel particulate
matter under I/M conditions is impractical, smoke opacity provides an
excellent proxy measurement. Smoke opacity can be measured using
Bosch, Hartridge, or similar test apparatus, or by an inexpensive end-
of-stack light-transmission opacity meter. The latter has the
advantage that it registers light-colored blue or gray smoke, as well
as the black smoke resulting from soot emissions.
Diesel HC and NOx concentrations can be measured directly by sampling
the exhaust gas, using readily available detectors. Because of the
higher molecular weight of Diesel HC emissions, they cannot be
measured using the same unheated infrared-type analyzers used for
gasoline engine HC measurements, however. Instead, a flame-ionization
type detector (with heated sample lines to prevent HC condensation in
the line) is required. NOx can be measured by chemilumenescent or IR
techniques.
Meaningful Diesel emissions measurements can only be performed with
the engine under load. Alternatives for supplying this load include
use of a chassis dynamometer, the vehicle's own inertia and/or that of
its drivetrain, the hydraulic torque converter on a vehicle so
equipped, or the vehicle's service brake. The specific speed and load
conditions during the test have an important effect on the emission
levels. The choice of transient versus steady-state test conditions
is also significant--many turbocharged Diesels have low emissions in
steady-state operation, but smoke heavily during transients.
In work for the California Air Resources Board, Weaver and Klausmeier
(1988) analyzed the results of a number of different test modes, using
both chassis dynamometer and in-motion tests with a portable opacity
meter. NOx, HC, and PM emissions were also measured during the
chassis dynamometer tests. Weaver and Klausmeier also analyzed a
database compiled by the New York City Department of Environmental
Protection, consisting of smoke and chassis transient emissions
measurements for several hundred trucks and buses.
Based on these analyses, Weaver and Klausmeier recommended measuring
smoke opacity under full-power acceleration from stop as the most
-77-
-------
effective and practical roadside test for high emissions. For
vehicles with hydraulic torque converters, an alternative test--
accelerating the engine from idle to full load against the stalled
torque converter (with the vehicle in gear, but held stationary by the
brakes) gave similar results, and was also recommended. The
recommended failure level for both test modes was 35% opacity (for
engines built to the pre-1988 U.S. standards). Initial smoke opacity
screening can be done by eye, with doubtful or disputed cases resolved
by a portable end-of-stack opacity meter. California now plans to
implement an in-use I/M program based on these recommendations.
For chassis dynamometer testing, Weaver and Klausmeier recommended
smoke opacity measurements in transient acceleration, steady-state
operation at peak torque, and at rated engine speed and three-quarters
power. The latter condition (corresponding to highway cruise) is a
very favorable one for emissions, and the recommended failure
-criterion is correspondingly stringent--4% opacity. Recommended
failure criteria for the acceleration and peak torque test modes were
35% and 15% opacity, respectively.
Weaver and Klausmeier also found good correspondence between HC and
NOx emissions in specific operating modes and overall emissions. For
NOx, the best correspondence was found at 50% and 75% power at rated
speed (the latter mode being the same as for highway cruise opacity
measurement). For HC, a weighted combination of emissions at 100%
power/rated speed and no-load/governor speed gave the best correlation
with overal emissions. Further research (with reference to the
specific emissions standards involved) would be required to establish
failure criteria for NOx and HC concentrations, however.
Chassis dynamometer testing, as investigated by Weaver and Klausmeier
(1988), is quite expensive, due to the cost of the dynamometer itself
(generally $100,000 to $200,000 installed), and the time required to
mount and secure the truck safely. Where these costs are prohibitive,
the free-roller technique as used in Santiago may be an attractive
alternative. The rollers are much less costly than dynamometer units,
and the time requirements are less. Since no horizontal force is
transmitted to the rollers, it is unnecessary to secure the truck in
place before testing, and it can simply be driven on and off by
locking the rollers. This has the disadvantage that it is difficult
to measure transient smoke emissions, however.
Cost-Effectiveness
The cost-effectiveness of a heavy-duty Diesel I/M program will depend
heavily on the degree to which it can be combined with existing
activities, and on the effectiveness of the emissions tests and anti-
tampering inspections in identifying high-emitting vehicles. In their
study for the California Air Resources Board, Weaver and Klausmeier
(1988) estimated the costs of a periodic I/M program using chassis
dynamometers at around $5,500 per ton of PM eliminated. This estimate
included a large credit for reducing NOx and HC emissions, however--
without these credits, the cost would have been more than $12,000 per
-78-
-------
ton. More than a third of this cost was due to truck and driver time
lost in going to and from the inspection.
For an in-use inspection program, using roadside smoke opacity tests
and anti-tampering inspections combined with California's existing
safety and truck weight inspections, the cost-effectiveness was
estimated at only about $1,500 per ton of PM eliminated. Without the
NOx and HC credits, this would have been $4,000 per ton instead. This
greater cost-effectiveness is due to a reduction in time wasted in the
inspection process, and to increased program effectiveness due to the
more effective inspection frequency for the roadside program.
5.4 Fuel Modification
Modifications to Diesel fuel composition have drawn considerable
attention in the United States as a quick and cost-effective means of
reducing emissions from existing vehicles. The two modifications
which show the most promise are a reduction in sulfur content, and in
the fraction of aromatic hydrocarbons in the fuel. Of the two, the
sulfur reduction is by far the cheaper and more cost-effective.
Current proposals being studied by the U.S. EPA would reduce sulfur
content in Diesel fuel to a maximum of 0.05% by weight. Possible
limits on the cetane index or aromatic content of the fuel are also
under consideration. The State of California has already adopted
regulations mandating 0.05% sulfur in the Los Angeles area, and is
presently considering limits of 0.05% sulfur and 10% aromatic
hydrocarbons, s tatewide.
Sulfur Content
The effects of Diesel fuel sulfur and aromatic hydrocarbon content on
emissions were addressed in a previous report by one of the authors
and others (Weaver et al., 1986). Based on the limited data then
available, this report concluded that reducing the fuel aromatic
content would significantly reduce Diesel NOx, HC, and PM emissions; .
while reducing the sulfur content would reduce PM emissions and
corrosive wear, thus increasing engine life. The savings due to
extended engine life were projected to more than compensate for the
increased refining cost to remove the sulfur, resulting in net
economic benefits to society, as well as substantial environmental
benefits. These conclusions proved highly controversial.
Since the publication of the 1986 report, many new data have come to
light on the relationship between fuel sulfur and engine wear. In
submissions to EPA, virtually every heavy-duty engine manufacturer has
stated that reducing fuel sulfur will beneficially affect engine life
(although the magnitude of these benefits is uncertain, and may have
been overestimated by Weaver et al.). Preliminary results from a
study of oil analyses in Southern California RTD buses show roughly a
30 percent reduction in wear metals in switching from the previous
fuel to fuel containing 0.05 percent sulfur. Tests by a major engine
manufacturer using a low-emission engine with air-to-air intercooling
-79-
-------
showed more than a two-thirds reduction in piston ring wear rate in
going from 0.27 percent to 0.05 percent sulfur in the fuel. Low-
emission engines would be expected to be more susceptible to corrosive
wear, since their low exhaust temperatures and high boost pressures
would make it easier for sulfuric acid to condense on the cylinder
walls. It appears clear from the available data, therefore, that a
reduction in fuel sulfur content is likely to result in measurable
benefits in engine life.
Fuel sulfur may also be a major reason for the increase in Diesel
engine wear due to EGR, as discussed in Section 3.6. If so, low
sulfur fuel would be required in order for EGR to become a practical
emission control technique for heavy-duty Diesel engines.
In addition to a direct reduction in emissions of S03 and sulfate
particles, reducing the sulfur content of Diesel fuel would reduce the
indirect formation of sulfate particles from S03 in the atmosphere.
In Los Angeles, it is estimated that each pound of S02 emitted results
in roughly one pound of fine particulate matter in the atmosphere. In
this case, therefore, the indirect FM emissions due to S02 from Diesel
vehicles are roughly as great as their direct particulate emissions.
S02 conversion to particulate matter is highly dependent on local
meteorological conditions, however, so the effects could be greater or
less in other cities.
The sulfur content of Diesel fuel can be reduced to 0.05% or less by
weight through hydrotreatment under moderate pressures in the presence
of a catalyst. This process is widely used in the U.S. Cost
estimates for adding this capability to a refinery without it range
from around $0.01 to $0.05 per gallon of Diesel fuel treated,
depending on the specific refinery and the availability of surplus
hydrogen for treatment. Even taking the upper end of this range, and
neglecting maintenance and engine life benefits, the cost-
effectiveness of reducing sulfur from 0.3 to 0.05% would be about
$6,000 per ton of PM eliminated, assuming the same S03 to particulate
conversion rate as Los Angeles. This is competitive with trap-
oxidizers. Factoring in maintenance benefits would reduce the net
cost considerably, as would a lower desulfurization cost.
Regulations limiting Diesel fuel sulfur content to 0.05% in the Los
Angeles area of California have been in force for several years.
Spot-market prices for such low-sulfur fuel (CIF Los Angeles Harbor)
are typically $0.02 per gallon higher than for standard distillate
(max 0.5% sulfur by weight). This suggests that the actual costs of
producing desulfurized fuel in a modern refinery are probably much
closer to $0.02 per gallon or lower, rather than $0.05.
Aromatic Hydrocarbons
A reduction in the aromatic hydrocarbon content of Diesel fuel may
also help to reduce emissions, especially where fuel aromatic levels
are high (as they are in the U.S. and Canada). For existing (pre-
1988) Diesel engines, a reduction in aromatics from 35% to 20% by
-80-
-------
volume would be expected to reduce transient particulate emissions by
10 to 15% and NOx emissions by 5 to 10%. HC emissions, and possibly
the mutagenic activity of the particulate SOF, would also be reduced.
Modelling studies of the refining industry have shown that aromatic
reductions of this magnitude can often be obtained through alterations
in Diesel fuel production and blending strategy, without a need for
major new investments in additional processing capacity.
The costs of reducing fuel aromatic content to 20% in the U.S. have
been estimated in EPA studies at $0,018 to $.033 per gallon, assuming
sulfur controls are in place. Studies conducted for industry have
estimated costs in the neighborhood of $0.08 per gallon. If the cost
were $0.03 per gallon, and PM emissions were reduced 10%, the cost-
effectiveness of this measure (for vehicles using pre-1991 technology)
would be about $20,000 per ton of PM eliminated, which is more than
the cost of other PM controls that have been adopted in the U.S.
Vehicles with advanced emissions controls would show a smaller
benefit, so that the costs per ton would be even higher. If half of
the cost were assigned to the NOx, HC, and mutagenicity benefits,
however, the resulting cost-effectiveness for pre-1991 vehicles would
be comparable to that for other emission control measures.
Reduced Diesel fuel aromatic content would have other environmental
and economic benefits which, if factored in, might help to improve its
cost-effectiveness. The reduced aromatic content would improve the
fuel's ignition quality, improving cold-starting and idling
performance and reducing engine noise. The reduction in the use of
catalytically cracked blending stocks should also have a beneficial
effect on deposit-forming tendencies in the fuel injectors, reducing
maintenance costs. On the negative side, however, the reduced
aromatics might result in some impairment of cold-flow properties, due
to the increased paraffin content of the fuel.
Fuel Additives
A number of well-controlled studies have demonstrated the ability of .
detergent additives in Diesel fuel to prevent and remove injector tip
deposits, thus reducing smoke levels. The reduced smoke probably
results in reduced PM emissions as well, but this has not been
demonstrated as clearly, due to the great expense of PM emissions
tests on in-use vehicles. Cetane-improving additives are also likely
to result in some reduction in HC and PM emissions in marginal fuels.
Claims for emissions benefits with specific fuel additives abound, and
a moderate (perhaps 10 to 20 percent) reduction in PM emissions with a
well-formulated cetane and detergency improving additive appears
credible. Hard experimental data in heavy-duty engines to support
these claims have been difficult to find, however. Therefore, while
fuel additives may offer a modest but cost-effective reduction in
Diesel emissions, such claims for any particular additive formulation
should be supported by actual test data, preferably long-term
comparative tests in heavy-duty DI engines.
-81-
-------
5.5 Retrofitting Emission Controls
Attempts to retrofit emissions controls to light-duty vehicles in
consumer service have been notably unsuccessful. Heavy-duty vehicles,
due to their more intensive usage, represent a more attractive target
for retrofits. A successful retrofit requires considerable care and
engineering development work, however--it is not enough simply to
"bolt on" some emissions control technology in the expectation that it
will be effective. Proper design, prototype testing (including
emissions testing), and manufacturing are required. Design and
development will require at least the services of an experienced
engine emissions laboratory, and preferably the assistance of the
original equipment manufacturer. In most cases, the systems developed
will be unique to a single engine or vehicle type.
Due to the expense involved in development, retrofits will generally
be cost-effective only where a large number of vehicles of similar
type and design are available for retrofit. Examples include transit
bus fleets, garbage collection fleets, urban delivery fleets, etc.
The highest priority for retrofit programs should probably go to
transit buses and to other vehicles operating in congested urban
areas, especially those with high-emission stop-and-go driving cycles.
Such programs could best be undertaken, at least initially, on a
voluntary or quasi-voluntary basis. Because of this, government-owned
vehicle fleets are especially suitable. Given the present state of
the technology, enforcement of a mandatory retrofit program for
privately owned vehicles would be very difficult.
Engine Technologies
Possible engine retrofits to reduce emissions range from simple
changes in settings to addition of new turbochargers and replacement
of fuel injection systems.
Maximum fuel rate--Particulate emissions from naturally aspirated
Diesel engines can often be reduced by simply derating the engine.
The extent to which PM emissions are reduced depends on the initial
smoke level and the extent of the derating. Where engines are not
subject to smoke or PM emissions limits, manufacturers may establish
maximum power ratings which require operation beyond the smoke limit.
Smoke levels at high altitudes may be excessive even where sea-level
smoke is low. In most cases, derating the engine by as little as 10%
is enough to produce a drastic reduction in smoke and PH. The impact
of this derating on the vehicle's intended use may be significant,
however, and must be assessed on a case-by-case basis.
Turbocharger--An alternative to derating a naturally aspirated engine
to reduce its smoke levels is to equip it with a turbocharger to
provide extra air for combustion. This requires some care, however.
Although many manufacturers produce both turbocharged and naturally
aspirated versions of the same engine, the turbocharged engines often
differ in compression ratio, injection timing, and piston design from
their naturally aspirated brethren. Simply bolting a turbocharger
-82-
-------
onto the naturally aspirated version may unduly increase maximum
cylinder pressures and possibly damage the engine. In most cases, it
will be necessary to rebuild the engine with pistons, etc. intended
for the turbocharged model. If the engine is in need of rebuilding
anyway, the incremental cost to rebuild it in turbocharged form will
generally be only about $500 to $1,500. Otherwise, the costs will be
much higher.
One situation where it will often be feasible to "bolt on" a
turbocharger without other modifications is where naturally aspirated
engines are used exclusively at high altitude. In this case, the
effect of the turbocharger is simply to restore the intake air
pressure for which the engine the engine was designed for.
Fuel/air ratio controls--Fuel-air ratio controls are of two types:
limits on the maximum fuel quantity injected, and acceleration smoke
limiters in turbocharged engines. Maximum fuel quantity levels are
adjustable on most types of fuel injection equipment, and could
readily be reduced to limit excess smoke and PM emissions at full
load. The cost of this adjustment would be very low, but it would
result in some loss of power--possibly impairing the vehicle's ability
to perform its intended functions.
Many turbocharged engines are already equipped with acceleration smoke
limiters for reasons of consumer acceptance, even where these are not
required by regulations. Where this is the case, more stringent smoke
standards could be achieved simply by adjusting the control settings,
at minimal cost. Even where acceleration smoke limiters have not been
installed as original equipment, most fuel injection equipment
manufacturers offer them as a standard option on their pumps. In many
cases, therefore, it would be possible to retrofit smoke limiters to
engines already in use. The cost of the additional parts required
would be of the order of up to $150. Since it would be necessary to
remove, disassemble, reassemble, and recalibrate the pump, labor costs
would be fairly high, however, unless the pump needed to be removed
and serviced anyhow.
Injection timing--Except where they are subject to stringent NOx
controls (as in California), Diesel engine fuel injection timing is
generally set to provide optimum fuel economy. Base injection timing
on most Diesel engines is adjustable to some degree. By retarding the
injection timing a few degrees, NOx may often be reduced 30 to 40%, at
some cost in fuel economy and particulate emissions. The relationship
between NOx and fuel economy/particulates depends on the specific
engine model and calibration. Typically, however, the first few
increments of NOx reduction have comparatively little impact, with the
effects growing rapidly more severe after some point. By resetting
the injection timing to optimize between NOx and PM emissions, it may
be possible to obtain significant NOx reductions at relatively little
cost.
Fuel injection system--As discussed in previous chapters, fuel
injection technology has advanced considerably in recent years. Fuel
injection pumps and nozzles suffer wear, and must be rebuilt or
-83-
-------
replaced every few years. By replacing the pump with a unit capable
of higher injection pressures, and injection nozzles with improved
designs, it may be possible to retard injection timing to reduce NOx
while still reducing PMand HC emissions and possibly fuel consumption.
Further benefits may be obtainable during engine overhaul by replacing
the pistons with updated models optimized for use with the new
injection system. Some U.S. engine manufacturers now offer standard
"updating" kits to bring their older engine models up to the
technology and performance levels of their current production.
Aftertreatment Technologies
Because they require little or no modification to the engine itself,
aftertreatment technologies may be more cost-effective than in-
cylinder emission controls in retrofit applications. The two
aftertreatment technologies that appear most promising as retrofits
are catalytic converters and particulate trap-oxidizers.
Catalytic converters--The advantages of catalytic converters for
Diesel engines have already been discussed. Although ineffective in
oxidizing Diesel soot, they can sharply reduce emissions of
particulate SOF, HC, CO, aldehydes, and mutagenic compounds such as
PNA and nitro-PNA. To avoid an increase in sulfate emissions, low-
sulfur fuel would be required. As discussed above, however, low-
sulfur fuel appears to be a highly cost-effective emissions control
measure in any case.
Due to the lower temperature and combustible content of Diesel engine
exhaust, catalytic converter temperatures would be much lower than in
conventional gasoline engines, and thermal protection requirements
would be minimal. The major cost in a retrofit program would be the
catalytic converter itself, therefore. Given large scale production,
this would probably be about $400 to $800 for typical heavy-duty
Diesel vehicles.
Trap-oxidizers--A number of programs have been undertaken to retrofit;
trap-oxidizers to existing vehicles, with varying degrees of success.
One successful program has occurred in Athens, Greece, where a number
of buses have successfully been retrofitted with ceramic monolith
traps and a manually controlled throttling regeneration system. Where
(as in buses and other centrally-serviced fleets) periodic manual
control of regeneration is possible, the difficult control and
reliability problems that have plagued trap-oxidizer development can
be significantly alleviated.
Even where manual control is not feasible, it appears likely that some
manufacturers will soon be able to offer more-or-less stand-alone
trap-oxidizer systems--including the trap, regeneration hardware,
sensors, and control unit--in a form which could be suitable for
retrofit applications. Much work is presently being done in this
area, and the fruits of this work should be commercially available
within a few years. The cost of such systems is presently unknown,
-84-
-------
but a range of about $1200 to $3000--depending on size and production
level--appears reasonable.
Alternative Fuels
Diesel engines can be modified to bum Ignition-improved methanol by
replacing the fuel injection pump with one sized for the larger
methanol quantities required, and resistant to methanol's corrosive
effects. With somewhat more effort, spark ignition or glow-plug
ignition may be used to ignite the methanol instead of additives.
Diesel-derived methanol engines produce very little PM, and less NOx
than a standard Diesel, but may produce unacceptable levels of
aldehydes and unburned methanol. Diesel engines can also be modified
fairly readily to utilize natural gas or LP gas, in either spark-
ignition or dual-fuel modes. The combination of spark-ignition, very
lean combustion, and an oxidation catalyst in the exhaust results in
very low emissions of all pollutants. These technologies are
discussed at greater length in Section 6.
5.6 Transportation Control Measures
Measures to reduce vehicle traffic should be considered in any program
to reduce mobile-source emissions. Unfortunately, the scope for such
measures is very limited where heavy-duty vehicles are concerned.
Trucks are an essential component of urban commerce, and there is no
obvious transportation mode available to replace them. The high costs
of truck operation have already prodded most truck operators to
eliminate unnecessary trips, and to use the smallest vehicle that can
do the job effectively. The situation for buses is much the same--
indeed, many programs to control light-duty vehicle traffic are likely
to increase, not decrease, the use of transit buses. Replacement of
buses with electric rail vehicles or trolley buses is seldom cost-
effective. Alternative fuel vehicles using compressed natural gas or
methanol are likely to prove a more economic alternative.
Changes in traffic patterns to improve the flow of traffic, and reduce
time and fuel-consuming stop-and-go operation have some potential for
heavy-duty Diesel emission reductions. For buses, establishment of
special priority bus lanes, limitations on private car access to
congested areas, and other traffic improvements have often proven very
effective in reducing traffic congestion and delays (World Bank,
1988). By reducing stop-and-start operations as well as congestion,
these measures can help considerably in reducing emissions as well.
In the case of heavy-duty trucks, regulations prohibiting operation
during peak traffic hours could possibly be justified. Such
regulations are now under consideration in Los Angeles. A requirement
that truck deliveries to the most congested city areas be made only at
night could also prove effective. A similar requirement has been in
effect in New York for some time.
-85-
-------
Changes in transportation patterns that reduce Diesel fuel consumption
are also likely to reduce emissions. Since one large truck uses less
fuel (and emits less) than two small ones, governments may wish to
review truck size and weight regulations to assure that they are
actually justified. Changing to alternative transport modes (such as
piggyback rail service instead of long-distance trucking) may help to
reduce the regional impacts of Diesel emissions. By concentrating
trucking activity around one central point, however, such measures may
significantly increase the local emissions impact.
-86-
-------
6. ALTERNATIVE FUELS
The possibility of substituting cleaner-burning alternative fuels for
Diesel fuel has drawn increasing attention during the last decade.
Motivations advanced for this substitution include conservation of oil
products and energy security, as well as the reduction or elimination
of certain pollutant emissions such as particulate matter and visible
smoke. Care is necessary in evaluating the air-quality claims for
alternative fuels, however. While many alternative fuel engines do
display greatly reduced particulate and SOx emissions, emissions of
other gaseous pollutants such as unburned hydrocarbons, CO, and in
some cases NOx and aldehydes may be much higher than from Diesels.
The principal alternative fuels presently under consideration are
natural gas and methanol made from natural gas. Other fuels having
present or potential local applications in heavy-duty vehicles include
liquified petroleum gas (LPG) and ethanol from biomass. Table 11
compares some of the key physical and combustion properties of these
fuels with those of gasoline and Diesel fuel.
This section provides an overview of all four alternatives as fuels
for heavy-duty vehicles. The properties and characteristics of the
fuels themselves are discussed, and the key issues surrounding their
utilization are summarized. These issues include fuel cost and
availability, utilization technologies, emissions tradeoffs, timing,
and economics.
6.1 Natural Gas
Natural gas has many desirable qualities as an alternative to Diesel
fuel in heavy-duty vehicles. Clean-burning, cheap, and abundant in
many parts of the world, it already plays a significant vehicular role
in a number of countries. Buses equipped with natural gas engines are
now in series production in Brazil, and numerous heavy-duty vehicles
around the world have been retrofitted to use this fuel. The major
disadvantage of natural gas as a motor fuel is its gaseous form at
normal temperatures.
Fuel Properties
Pipeline-quality natural gas is a mixture of several different gases.
The primary constituent is methane, which typically makes up 90-95% of
the total volume. Other, minor constituents include nitrogen, carbon
dioxide, higher hydrocarbons such as ethane and propane, and traces of
hydrogen sulfide and water. The presence of these minor constituents
has some effect on the actual properties of the gas, but for most
-87-
-------
Table 11
Properties of Alternative and Conventional Fuels
Diesel Gasoline Methanol Ethanol Propane Methane
Energy Content 45.15 43.65 23.86 29.73 50.37 55.53
(HHV MJ/kg)
Liquid Density .843-.848 -.735 .7914 .7843 .5077 .4225
(kg/1)
Energy Density of 38.16 32.1 18.9 23.32 25.6 23.46
Liquid (MJ/1)
Energy Density of
Gas (MJ/1)
@ STP* --- --- --- ... 0.036
@ 200 BAR --- --- --- --- --- 7.47
Normal Boiling 140-360 37-205 65 79 -42.15 -161.6
Point °C
Research Octane No. -25 91-97 112 111 125 130
Cetane No. 45-55 0-5 5 5 -2 0
*
standard temperature and pressure
practical purposes natural gas can be treated and analyzed as pure
methane.
Methane is a nearly ideal fuel for Otto-cycle (spark-ignition)
engines. As a gas under normal conditions, it mixes readily with air
in any proportion, eliminating cold-start problems and the need for
cold-start enrichment. It is flammable over a fairly wide range of
air-fuel ratios. With a research octane number of 130 (the highest of
any commonly used fuel), it can be used with engine compression ratios
as high as 15:1 (compared to 8-9:1 for gasoline), thus giving greater
efficiency and power output. The low lean flammability limit permits
operation with extremely lean air-fuel ratios--having as much as 60%
excess air. On the other hand, its high flame temperature tends to
result in high NOx emissions, unless very lean mixtures are used.
Because of its gaseous form and poor self-ignition qualities, methane
is a poor fuel for Diesel engines. Since Diesels are generally
somewhat more efficient than Otto-cycle engines, natural gas engines
are likely to use somewhat more energy than the Diesels they replace.
-88-
-------
The high compression ratios achievable with natural gas limit this
efficiency penalty to about 10% of the Diesel fuel consumption,
however. In many cases, the increase in energy consumption is more
than offset by the lower cost of gas, resulting in a net savings in
fuel costs.
The fact that methane is a gas under normal conditions creates
significant problems with fuel storage aboard the vehicle. At
present, natural gas is stored either as a gas in high-pressure
cylinders or as a cryogenic liquid in an insulated tank. Both forms
of storage are considerably heavier, more expensive, and bulkier than
storage for an equivalent amount of Diesel fuel. The costs of
compressing or liquifying natural gas in order to store it are also
substantial, in some cases more than offsetting its lower cost.
Engine Technology and Emissions
Although an excellent fuel for Otto-cycle engines, natural gas is not
suitable by itself as a fuel for Diesels. Options for using natural
gas in heavy-duty vehicle engines are thus limited to the following:
- Fumigation, or mixing the gas with the Diesel intake air to be
ignited by Diesel fuel injected in the conventional way;
Conversion of the existing Diesel engine to Otto cycle
operation; or
Replacement of the Diesel engine with a conventional spark-
ignition engine.
These options are discussed below.
Fumigation--The simplest way to use natural gas in a Diesel engine is
simply to mix it with the intake air. Injection and combustion of the
Diesel fuel then ignites and bums the alternate fuel as well. Since
the gas supplies much of the energy for combustion, the Diesel fuel
delivery for a given power level is reduced. This results in reduced
smoke and particulate (PM) emissions at high load, and can increase
the smoke-limited power of the engine. However, incomplete combustion
(especially at light loads) usually increases CO and HC emissions
considerably. The increased HC emission are of less concern with
natural gas than with other hydrocarbon fuels, since the principal
component, methane, is non-toxic and has very low photochemical
reactivity. On the other hand, methane is a very active greenhouse
gas, and the other, minor components of the HC emissions include some
formaldehyde as well as higher hydrocarbons.
Fumigated engines are normally set up to idle on Diesel fuel only,
with the gas being added in increasing amounts at higher loads, and
the amount of Diesel fuel injected per stroke (the "pilot" fuel
quantity) kept constant. The pilot fuel quantity required to ensure
good combustion in a fumigated engine is of the order of 5 to 20% of
-89-
-------
the full-load energy consumption of the engine. Near full load, up to
95% of the total fuel energy can be supplied by the gas, but at lower
loads, much more Diesel energy is required to ensure good combustion.
Because idling and light-load operation account for a large part of
total engine fuel consumption, it is typically possible to substitute
natural gas for only about 40-70% of the total energy consumption.
The major advantage of the fumigation technique is the fact that no
major modifications are required to the engine, making it easy to
install in an existing vehicle. A number of commercial kits for
converting Diesel engines to dual fuel operation are available. The
major disadvantages of the fumigation approach are the fact that only
partial substitution is possible, and the large increases in HC and CO
emissions that may result. The need for two fuel supply systems adds
cost and weight, and creates additional safety hazards. The fact that
a large part of the energy is still provided by Diesel fuel also
limits the emissions benefits. Although fumigation greatly reduces
smoke and particulate emissions at full load, the interaction between
the gaseous fuel and the pilot may actually increase particulate
production at lower loads. Thus, the net effect on particulate
emissions will depend on the duty cycle.
Spark-ignition engines--Natural gas fueled, spark-ignition versions of
heavy-duty truck Diesel engines have been used for stationary
applications for many years, and a number of similar conversions have
been carried out with vehicular engines. The modifications required
to convert a Diesel engine to Otto-cycle operation are machining the
cylinder head to accept a spark plug instead of a fuel injector;
redesign of the pistons to reduce the compression ratio; replacement
of the fuel injection pump with an ignition system and distributor;
replacement of exhaust valves and valve seats with wear resistant
materials; and addition of a carburetor (or fuel injection system) and
throttle assembly.
For new engines, these changes involve only minor changes in the
existing Diesel manufacturing process, and would be economic even on a
very small scale. MAN, Saab-Scania, Daimler-Benz, Caterpillar, and .
Cummins, among others, have produced such engines on a demonstration
or small-volume production basis. This conversion can also be
performed on an existing Diesel engine, at a cost comparable to that
of a major overhaul. A number of such conversions have been performed
in New Zealand, Australia, and Brazil.
Current heavy-duty Otto-cycle engines can be divided into two groups:
"rich burn" and "lean-burn". Conventional stoichiometric or "rich
burn" engines operate with the air-fuel ratio X close to or even
somewhat less than 1.0. The rich-burn approach provides the maximum
power output for a given engine displacement and turbocharging level,
at some cost in fuel economy.
"Lean-burn" engines operate with X in the range of 1.4 to 1.6, or even
higher where stratified-charge techniques are used. Since lean
mixtures knock less readily than stoichiometric ones, these engines
can use higher compression ratios. Throttling losses in the lean-burn
-90-
-------
engine are also lower at a given power level, since a given amount of
fuel occupies a larger volume of mixture. All of these effects help
to increase the thermodynamic efficiency of the lean-bum engine
compared to a rich-burn type. Lean-burn natural gas engines used in
stationary applications commonly attain thermal efficiencies at full
load comparable to those of Diesels.
The choice between lean-burn and rich-burn has an important effect on
emissions. Figure 8 shows the typical relationship between HC, CO,
and NOx and A for Otto-cycle engines. CO emissions are governed by
oxygen availability, while NOx emissions are primarily a function of
flame temperature. For natural gas engines, the peak NOx emissions at
A - 1.1 are in the range of 20 to 30 grams per BHP-hr. Typical NOx
emissions at stoichiometry (A - 1) are about 10 to 15 g/BHP-hr. This
can be reduced to less than 2 g/BHP-hr, however, through the use of a
three-way catalyst and closed-loop mixture controls like those on
light-duty passenger cars. The durability of such catalysts under the
high temperatures experienced in heavy-duty operation is doubtful,
however. The durability of electronic control systems under heavy-
duty conditions is also unproven.
Figure 8
Typical Variation of Emission Levels with Air-Fuel Ratio A
for an Otto-Cycle Engine
(Source: Bergmann and Busenthur, 1986)
misfire limit
NOx
CO
emission
0-
1.5
2
air/fuel ratio A
-91-
-------
An alternative approach to NOx control is to operate very lean. At
A - 1.5, which is about the limit for smooth running with normal
ignition systems, the lower flame temperature results in NOx levels
around 3.5 to 4.5 g/BHP-hr. Through the use of high-energy ignition
systems and careful optimization, homogeneous mixtures with A as lean
as 1.6 can be burned. NOx emissions from such engines are typically
in the 2 g/BHP-hr range. The drivability of such engines under
transient conditions has not yet been demonstrated, however. The lean
combustion limit can be extended even further by using a stratified
charge. Stationary engines using this technique have demonstrated NOx
emissions less than one g/BHP-hr.
To avoid excessive smoke, Diesel engines are designed to run at rather
lean air-fuel ratios, with A typically in the range of 1.5 to 1.8 at
full power. They are well-adapted, therefore, to lean-bum operation
on natural gas, and can typically match or exceed their Diesel power
rating. Stoichiometric operation, with its higher exhaust
temperatures, is likely to lead to valve and turbocharger problems,
however.
Existing Otto-cycle engines, on the other hand, are designed for near-
stoichiometric operation. Lean-bum operation in these engines is
likely to result in a significant power loss. In addition, these
engines typically lack the mechanical strength and other design
features required to take full advantage of the high compression
ratios possible with natural gas. Thus, the fuel efficiency of
gasoline-type engines converted to natural gas is likely to be
considerably less than that of a Diesel-engine-based lean-bum engine.
Fuel Storage and Refueling
Natural gas may be stored on-board a vehicle either as a compressed
gas in high-pressure cylinders or as a cryogenic liquid. Although
both processes are expensive and energy-intensive, compression is much
less so than liquefaction. The current maximum working pressure for
CNG cylinders is 3,000 PSIG (200 bar). The volumetric energy content
of natural gas at this pressure is about one-fifth that of Diesel
fuel, while that of LNG is a little over half that of Diesel.
The high-pressure cylinders needed for CNG weigh more and occupy more
space than the vacuum-insulated tanks used for LNG, but the cost of
the two storage systems is about the same. The lesser weight and
volume required by LNG tanks is a critical advantage in some
applications (e.g., long-distance hauling). For typical urban
vehicles, however, there is sufficient unused space between the frame
rails to accommodate enough CNG cylinders for a full day's operation
with some reserve.
A major drawback of LNG storage is the need to provide for venting of
gas. As the tanks are not perfectly insulated, some heat leaks in.
During normal operation, this heat is removed by drawing off the vapor
above the tank, thus cooling the liquid by evaporation. Upon
prolonged standing, however, pressure builds up in the tank. While
-92-
-------
the Cank is capable of withstanding 20-40 atmospheres, it must
eventually vent the excess vapor. For available tank designs, the
standing time before this occurs is of the order of a week (Deluchi et
al., 1988).
Two types of CNG refueling systems are in common use: slow-fill and
quick-fill. In the slow-fill approach, the fuel tanks of all vehicles
at a given facility are connected to a common fuel manifold. A
compressor pumps high-pressure gas through the manifold into all of
the tanks simultaneously (usually overnight). This system is well
suited to the operating patterns of buses and other fleet vehicles
which return to the same location each night. In the alternative,
quick-fill system, the compressor is connected to one vehicle at a
time, in sequence, in the same way as a Diesel refueling pump. The
refueling rate is then limited by the capacity of the compressor. As
compressor capacity is the major cost in a CNG system, rapid refueling
by this means can be very expensive. Often, quick-fill systems will
include a small amount of high-pressure gas storage, arranged in a
"cascade" system, tfith this approach, vehicles can be refueled
quickly from the cascade, which is then recharged more slowly by the
compressor.
Fuel Supply and Costs
After coal, natural gas is the most abundant fossil fuel. The ratio
of proven gas reserves to annual production is double that of
petroleum, and a larger fraction of world gas reserves than petroleum
reserves are found outside the Middle East. Today, most major urban
centers and many minor ones in industrial countries are served by a
large network of gas pipelines.
Transport and distribution of natural gas by pipeline is a well-
developed and relatively low-cost technology. Use of compressed
natural gas (CNG) from pipelines as a vehicular fuel is also well-
developed in Italy (where it has been used for more than 40 years),
Canada, and New Zealand. Other technologies for natural gas
transportation and distribution include liquefaction and shipment in
liquid form (LNG), and short-distance transport of CNG in large banks
of cylinders. Japan and many countries of Western Europe now import
significant quantities of natural gas in the form of LNG.
Owing to the difficulty of transportation, the costs of natural gas
vary greatly from country to country, and even within countries.
Where gas is available by pipeline from the field, its price is
normally set by competition with residual fuel oil or coal as a burner
fuel. The market-clearing price of gas under these conditions is
typically about $3.00 per million BTU (equivalent to about $0.41 per
gallon of Diesel fuel equivalent). Compression costs for CNG use can
add another $0.50 to $2.00 per million BTU, however, depending on the
size of the facility and the natural gas supply pressure. The higher
the pipeline supply pressure of the gas, the lower the capital and
operating costs of the compression facility.
-93-
-------
The cost of LNG varies considerably, depending on specific contract
terms (there is no effective "spot" market for LNG). The cost of
small-scale liquefaction of natural gas is about $2.00 per million
BTU, making it uneconomic in comparison to CNG in most cases. Where
low-cost remote gas is available, however, LNG production can be quite
economic. Typical 1987 costs for LNG delivered to Japan were stated
as about $3.20 to $3.50 per million BTU (Oil and Gas Journal, 1988),
The costs of terminal receipt and transportation would probably add
another $0.50 or so to this cost at the wholesale level. Compression
costs would be nil, however, since LNG could either be used directly,
or converted directly to CNG by allowing it to vaporize under
pressure. The costs of LNG for vehicular fuel (given that a nation
was importing LNG anyway) would thus be comparable to those of CNG.
Capital and Maintenance Costs
Lean-bum natural gas engines based on Diesel engine designs should
cost no more to manufacture than the parent Diesel, if produced in
moderate quantity. Additional capital costs for new vehicles,
therefore, would be limited essentially to the extra costs of the fuel
storage. For heavy-duty vehicles using CNG, the additional cost of
the cylinders would amount to about $2,000 to $5,000, depending on
vehicle size and range requirements. The costs of LNG tanks would
probably be similar, but such tanks are not commercially available at
present.
To retrofit an existing Diesel vehicle for lean-burn operation using
CNG would be more expensive. Assuming that a large number of vehicles
were to be converted, the costs per vehicle would probably range from
about $5,000 to $10,000, with about half of the cost being due to the
CNG storage system.
Maintenance costs for natural gas engines based on Diesel engine
components should be lower than those for the parent Diesel. Due to
the clean-burning nature of the fuel, oil-change intervals can be
extended, and the engine life is likely to be considerably longer than
with Diesel fuel. The added costs of maintaining spark plugs,
ignition systems, and fuel storage systems would be offset by the
avoided costs of fuel injection system maintenance. Other maintenance
costs should be essentially the same between the two technologies.
Current Activities
The use of natural gas in heavy-duty vehicle engines is the subject of
a fair amount of activity around the world. Commercial kits for
converting Diesels to dual fuel operation are available in the U.S.
and Europe. A number of heavy-duty truck and bus engines have been
converted to natural gas spark-ignition operation in New Zealand,
Australia, and Brazil. In Brazil, Daimler-Benz AG is producing new
natural-gas fueled engines and buses on its production lines for
domestic use. Cummins Engine Co. in the U.S. is developing a lean-
burn, spark-ignition conversion of its L-10 bus engine, with a view
-94-
-------
Coward compliance with the U.S. 1991 transit bus emissions
regulations. Several other heavy-duty Diesel engine manufacturers
have demonstrated spark-ignition conversions of their existing Diesel
engine models, although only Daimler-Benz is currently producing such
engines for vehicular use.
6 . 2 Liquified Petrnlpnni Gas fLPO
Liquified petroleum gas is already widely used as a vehicle fuel in
the U.S., Canada, the Netherlands, and elsewhere. As a fuel for
spark-ignition engines, it has many of the same advantages as natural
gas, with the additional advantage of being easier to carry aboard the
vehicle. Its major disadvantage is the limited supply, which would
rule out any large-scale conversion to LPG fuel.
Fuel Properties
Liquified petroleum gas or LPG is typically a mixture of several gases
in varying proportions. Major constituent gases are propane (C3H8)
and normal butane (C4H10), with minor quantities of propylene and
other hydrocarbon gases. Because of its superior knock-resistance,
propane is preferred to butane or propylene as an automotive fuel.
Propane presents a useful combination of combustion and storage
properties. Like methane, it is a gas at normal temperatures and
pressures, and thus mixes readily with air in any proportion. Cold
starting is not a problem, therefore, and cold-start enrichment is
unnecessary. Propane's research octane rating of 125, while somewhat
lower than that of methane, is still much higher than gasoline, and
permits the use of compression ratios in the range of 11-12:1. The
lean combustion limit of propane-gasoline mixtures is also
considerably leaner than for gasoline, allowing the use of lean-burn
calibrations which increase efficiency and reduce emissions. These
mixtures are also more resistant to knocking, permitting the use of
still higher compression ratios. This is risky, however, as
inadvertent contamination with butane (which is denser and has a lower
octane value) can cause destructive detonation in the engine.
Engine Technology and Emissions
The technologies available for LPG are the same as those available for
natural gas: fumigation, or spark ignition using either stoichiometric
or very lean mixtures. Due to the lower octane value of LPG, the
compression ratio (and thus the thermal efficiency) possible with this
fuel in spark-ignition operation is lower than with natural gas,
although still considerably higher than with gasoline. Aside from
this, the engine technologies involved are very similar. Due to the
lower octane value (and higher photochemical reactivity) of LPG,
however, it is not as good a candidate for use in fumigation as
natural gas.
-95-
-------
Like natural gas, LPG in spark-ignition engines is expected to produce
essentially no particulate emissions (except for a small amount of
lubricating oil), very little CO, and moderate HC emissions. NOx
emissions are a function of the air-fuel ratio, A. LPG does not burn
as well under very lean conditions as natural gas, so the NOx levels
achievable through lean-burn technology are expected to be somewhat
higher--probably in the range of 3 to 5 g/BHP-hr. For stoichiometric
LPG engines, the use of a three-way catalyst and closed-loop air-fuel
mixture control results in very low NOx emissions (Van der Weide et
al., 1988), assuming that such systems can be made sufficiently
durable.
Fuel Storage and Handling
Propane is stored on the vehicle as a liquid under pressure. Propane
tanks, since they must contain an internal pressure of 20-40
atmospheres, are generally cylindrical with rounded ends, and are much
stronger than tanks used for storing gasoline or Diesel fuel, albeit
much less so than those used for CNG. Propane can be pumped from one
tank to another like any liquid, but the need to maintain pressure
requires a gas-tight seal. Except for the need for a standardized,
gas-tight connection, propane used as vehicle fuel can be dispensed in
much the same way as gasoline or Diesel fuel. To ensure that some
vapor space is always available for expansion, propane tanks used in
automotive service must never be filled more than 80% full. Automatic
fill limiters are incorporated in the tanks to ensure that this does
not occur.
Fuel Supply and Costs
LPG is produced in the extraction of heavier liquids from natural gas,
and as a byproduct in petroleum refining. Presently, LPG supply
exceeds the demand in most petroleum-refining countries, so the price
is low compared to other hydrocarbons. Wholesale prices for consumer-
grade propane in the U.S. have ranged between $0.25 and $0.30 for
several years, or about 30% less than the wholesale cost of Diesel on
an energy basis. Depending on the locale, however, the additional
costs of storing and transporting LPG may more than offset this
advantage. Differences in road taxes, sales taxes, etc. between LPG
and Diesel may also have an important effect on the economics as
perceived by private consumers. Such taxation differences do not
affect the social cost, however.
Because the supply of LPG is limited, and small in relation to other
hydrocarbon fuels, any large-scale conversion of heavy-duty vehicles
to LPG use would likely absorb the existing glut, causing prices to
rise. For this reason, LPG probably makes the most sense as a special
fuel for use in vehicles, such as urban buses and delivery trucks,
operating in especially pollution-sensitive areas.
-96-
-------
Capital and Maintenance Costs
The costs of producing lean-burn LPG engines based on Diesel engine
designs should be no more than the costs of the existing Diesels.
While LPG storage tanks are somewhat more expensive than Diesel fuel
tanks, this difference would amount to no more than a few hundred
dollars in a heavy-duty vehicle. For practical purposes, therefore,
the capital costs of Diesel and LPG vehicles can be considered to be
equivalent. To convert an existing Diesel vehicle to LPG operation
would be somewhat more expensive, however. If done on a relatively
large scale, the cost per vehicle would probably range from about
$2,000 to $6,000, depending on vehicle size and configuration.
The maintenance costs of heavy-duty vehicles using LPG fuel should be
essentially the same as those for natural gas vehicles. Extended oil
change intervals and engine life would be major benefits of a switch
to LPG fuel.
Current Activities
Several hundred LPG-fuelled city buses are now in use in Vienna,
Austria. The engines for these buses are spark-ignition conversions
of the MAN Diesel bus engine (this engine has also been adapted to
methanol and natural gas). Research on heavy-duty LPG engines is also
underway at TNO in the Netherlands, in New Zealand, and in Canada.
LPG conversions of light and heavy-duty spark-ignition engines have
been routine for some time, and several hundred thousand such vehicles
are in use around the world.
6.3 Methanol
Widely promoted in the U.S. as a "clean fuel," methanol in fact has
many desirable combustion and emissions characteristics, including
good lean combustion characteristics and relatively high octane (both
of which can lead to high efficiency), plus low flame temperature
(leading to low NOx emissions) and low photochemical reactivity. One
drawback of methanol as a fuel is its cost--it is highly variable, and
will probably be somewhat higher than Diesel fuel. Methanol is
produced in rather small volumes, and there have recently been some
rather large but volatile new demands such as for MTBE production. As
a result, spot prices in the last few years have varied between about
$0.25 and $0.75 per gallon.
As little as one methanol plant coming on-line or off-line can
markedly alter methanol spot prices, making long term planning
difficult. However, a recent EPA report (EPA, 1989) has estimated
that, with large-scale production, fuel methanol could be produced and
landed in the U.S. at a total cost of $0.30 to $0.40 per gallon.
Other estimates are generally higher. Because of the differences in
energy content per gallon between Diesel and methanol fuels, an engine
would consume more than 2.2 times as much methanol as Diesel fuel per
mile. This must be considered when comparing future fuel prices. For
-97-
-------
example, if current methanol prices were $0.30 per gallon, methanol
would be more expensive on an energy basis than Diesel fuel at the
present bulk price of approximately $0.60 per gallon.
Fuel Properties
As a liquid, methanol can either be burned in an Otto-cycle engine or
injected into the cylinder as in a Diesel. With a fairly high
research octane number of 112, and excellent lean combustion
properties, methanol is a good fuel for lean-burn Otto-cycle engines.
Its lean combustion limits are similar to those of natural gas, while
its low energy density results in a low flame temperature compared to
hydrocarbon fuels, and thus lower NOx emissions. Methanol burns with
a sootless flame and contains no heavy hydrocarbons. As a result,
particulate emissions from methanol engines can be very low-
consisting essentially of unburned lubricating oil.
Methanol's high octane number results in a very low cetane number, so
that methanol cannot be used in a Diesel engine without some
supplemental ignition source. Investigations to date have focused on
the use of ignition-improving additives, spark ignition, glow-plug
ignition, or dual injection with Diesel fuel. Converted heavy-duty
Diesel engines using each of these approaches have been developed and
demonstrated.
The low energy density of methanol means that a large amount (more
than 2.2 times the mass of Diesel fuel) is required to achieve for the
same power output. The high heat of vaporization of methanol,
combined with the large amounts required, makes it difficult to ensure
complete vaporization in Otto-cycle engines, and requires special
attention to the design of intake manifolds and cold-start procedures.
Current Otto-cycle engine designs using liquid methanol become nearly
impossible to start below about 5° C without the use of special pilot
fuels or supplemental heating techniques. Engines using Diesel-type
direct liquid injection (and glow plugs) do not have this problem, and
are frequently easier to start than the parent Diesels under cold
conditions.
Most methanol sold in commerce is of chemical grade. For chemical
purposes, methanol is required to be very low in dissolved water, and
essentially free of hydrocarbons. Methanol used as motor fuel does
not require the same high standards of purity, and could thus be
produced and handled more cheaply than chemical-grade methanol. The
admixture of substantial quantities of higher alcohols, one or two
percent of hydrocarbons, and/or as much as 2% dissolved water would
have little adverse effect, other than changes in volumetric energy
content.
Engine Technology and Emissions
Options for methanol utilization in heavy-duty engines include both
Otto-cycle and Diesel-cycle operation. As a liquid, methanol can
-98-
-------
readily be injected directly into the cylinder, in the same way as
Diesel fuel in a Diesel engine. It can also be atomized, vaporized,
or chemically reformed and mixed with the air charge, in the same way
as gasoline in a conventional spark-ignition engine.
Otto-cvcle--When burned in an Otto-cycle engine, methanol performs
similarly in many respects to natural gas or LPG. The major
differences are the lower flame temperature (which reduces NOx
emissions by about half from those experienced with hydrocarbon fuels
at the same air-fuel ratio), and the fact that liquid methanol must be
vaporized and mixed thoroughly with air before it burns. One approach
used by Daimler-Benz is to vaporize the methanol before mixing it with
the air, using a separate vaporization system, and then burn it as a
lean mixture in a Diesel engine converted to spark ignition. This
results in HC and CO emissions very similar to those of similar
engines using natural gas or propane (Bergmann and Busenthur, 1986) .
(Note that "HC" emissions for methanol engines are mostly unburned
methanol, which shows less photochemical reactivity than propane, but
more than natural gas.) Where methanol is mixed with the air as a
liquid, somewhat higher HC and CO emissions would be expected,
especially under cold start conditions. Methanol can also be used in
stoichiometric, closed-loop combustion systems such as those found in
light-duty vehicles. There is little information available on the
performance of these systems under heavy-duty conditions, however.
Diesel iniection--Recent development activities in the area of heavy-
duty methanol engines have focussed on Diesel injection techniques,
due to the greater fuel-efficiency possible with this approach.
Although methanol can be injected into the cylinder as in a Diesel
engine, the Diesel injection pump supplied with the engine would not
be suitable in most cases. Since about 2.2 times as much methanol as
Diesel fuel is required to supply the same amount of energy, a larger
volumetric capacity is required. In addition, Diesel injection pumps
are fuel-lubricated. Since methanol is a poor lubricant, a separate
oil supply to the pump would be required. Other changes to the high-
pressure lines, injector nozzles, and so forth are required to prevent
cavitation and premature wear. All of these changes are
straightforward, however, and injection pumps suitable for use with
methanol have been produced for research and development purposes
Due to its poor cetane number, methanol will not self-ignite reliably
in a Diesel engine; thus, some form of ignition assistance is
required. Ignition approaches that have been demonstrated include
spark plugs, glow plugs, dual injection with "pilot" Diesel fuel, and
the use of ignition-improving additives mixed with the methanol. All
of these approaches can produce thermal efficiencies as high or higher
than those of a conventional Diesel with similar levels of HC and
aldehyde emissions, low NOx and CO, and (except with pilot injection)
virtually no particulate matter.
Of the four ignition techniques, the spark plugs and ignition-
improving additives appear most attractive. Diesel pilot injection
suffers from many of the drawbacks discussed above in the case of
fumigation, particularly the limited potential for fuel substitution
-99-
-------
and emissions control. In addition, the need for two separate fuel
injection and supply systems adds considerable complexity and expense.
Glow plugs consume much more energy than spark plugs and are harder to
control, and present designs have much shorter service lives. They
are also slower to ignite the fuel, reducing engine efficiency
somewhat. Although spark plugs require a high-voltage ignition system
and major changes to the cylinder head, their advantages are
considered to outweigh these drawbacks.
Perhaps the most technically attractive approach to igniting the
methanol is the use of ignition-improving additives such as organic
nitrates. With these additives, methanol can be used in the same way
as Diesel fuel, with no need for external energy sources for ignition.
This approach would also involve minimal modifications to the engine,
as it would be necessary only to replace the fuel injection pump and
related components. The cylinder head, pistons, and other internal
engine components could be left in the same configuration as for the
Diesel engines. This would allow a ready conversion back to Diesel
fuel, if required, and (in an emergency) the engine could even be run
on Diesel fuel in the methanol configuration.
Concerns with the ignition improver approach include the costs of the
additive, and potential effects on emissions. Currently available
additives have demonstrated effectiveness at concentrations of 0.5 to
5% by volume, so that even a fairly expensive additive would have a
limited effect on total cost. One supplier quoted a rough cost for
large quantities of a commercial additive ("Avocet") of about $1,300
per ton. Blended at a rate of 1%, this would add about $.02 per liter
to the cost of the finished product. Tests have also shown no adverse
effects on pollutant emissions due to the use of this additive.
Emissions--Methanol combustion does not produce soot, so particulate
emissions from methanol engines are limited to a small amount of
lubricating oil. Methanol's flame temperature is also lower than that
for hydrocarbon fuels (at the same X ratio), resulting in NOx
emissions which are typically 50% lower. CO emissions are generally
comparable to or somewhat greater than those from a Diesel engine
(except for stoichiometric Otto-cycle engines, for which CO emissions
may be much higher). These emissions can be controlled with a
catalytic converter, however.
The major pollution problems with methanol engines come from emissions
of unbumed fuel and formaldehyde. Methanol (at least in moderate
amounts) is relatively innocuous--it has low photochemical reactivity,
and--while acutely toxic in large doses--displays no significant
chronic toxicity effects. Formaldehyde, the first oxidation product
of methanol, is much less benign, however. A powerful irritant and
suspected carcinogen, it also displays very high photochemical
reactivity. While all combustion engines produce some formaldehyde,
some early-generation methanol engines exhibited greatly increased
emissions compared to Diesels (Ullman and Hare, 1982). Recent
progress, however, has been such that engine-out emissions are only
slightly above Diesel-equivalent levels.
-100-
-------
The potential for large increases in formaldehyde emissions with the
widespread use of methanol vehicles has raised considerable concern
about what would otherwise be a relatively benign fuel from an
environmental standpoint. It remains to be seen if the formaldehyde
increase from Diesel levels becomes in fact a major problem.
Formaldehyde emissions can be reduced through changes in combustion
chamber and injection system design, and are also readily
controllable through the use of catalytic converters, at least under
warmed-up conditions. A significant problem with the catalytic
converter approach (especially for Diesels) is that--at low
temperatures--certain precious metal catalyst formulations can
actually increase emissions of formaldehyde considerably. This occurs
through the catalytic partial oxidation of methanol in the exhaust.
If methanol emissions are high (as they tend to be during cold
operation) and the catalyst temperature is low, the resulting
formaldehyde emissions could present an acute problem. Efforts to
resolve this problem are focussing on special low-formaldehyde
catalysts, and on minimizing unbumed methanol emissions. These
efforts, are promising, and have demonstrated control of formaldehyde
to roughly Diesel equivalent levels.
Fuel Storage and Handling
As a liquid at normal temperatures and pressures, methanol presents
few special storage or handling problems other than those of materials
compatibility. Methanol is corrosive to some aluminum alloys, lead,
and zinc; so these materials must not be used in methanol-handling
equipment. It also attacks many of the elastomeric sealing materials
that are commonly used with hydrocarbons, causing them to swell and
crack. The use of methanol-tolerant elastomers and other materials is
required. Although pure methanol is relatively uncorrosive of carbon
steel, the addition of water (as in fuel-grade methanol) increases the
corrosive effect. The use of anti-corrosive additives in fuel
methanol may be necessary in order to avoid excessive corrosion of
storage and distribution equipment. Otherwise, methanol shares the
advantage of ease of storage and handling with other liquid fuels.
Fuel SuppIv and Costs
Methanol can be produced from natural gas, coal, or biomass. At
current and foreseeable prices, the most economical feedstock for
methanol production is natural gas, especially natural gas found in
remote regions where it has no ready market. The current world market
for methanol is as a commodity chemical, rather than a fuel, and world
methanol production capacity is limited and projected to be tight at
least through the mid-90s. Any large-scale conversion of vehicles to
methanol fuel would require new methanol production capacity to be
built if prices were not to rise significantly.
U.S. methanol prices are commonly quoted FOB the U.S. Gulf Coast. The
price of methanol on the world market has fluctuated dramatically in
-101-
-------
the last decade, from around $0.25/gallon in the early 1980's to
$.60-.70 in 1988, and back to $0.25 at the present time. The lower
prices reflect the effects of a glut; while the higher prices reflect
a temporary shortage.
Economies of scale and the lower quality requirements for fuel
methanol are expected to make its production somewhat less costly than
chemical methanol. A recent EPA report (EPA, 1989), using capital and
transportation cost estimates developed by Bechtel Corporation,
estimates that fuel methanol could be produced for $0.25-0.35 per
gallon. This study assumed that remote natural gas resources, without
another market, would be available at a price of $0.50 to $1.00 per
million BTU. Transportation by tanker to consumption centers in the
U.S. was estimated to add another $0.05 per gallon, for a total landed
cost estimated at $0.30 to $0.40 per gallon. Other estimates based on
the same data place this range at $0.36-$0.41 (California Energy
Commission, 1989). For comparison, the landed price of Diesel fuel
(with 2.2 times as much energy per gallon) is approximately $0.60 in
early 1990.
In addition to new methanol supply capacity, any large-scale use of
methanol for vehicle fuel would require substantial investments in
fuel storage, transportation, and dispensing facilities, which would
further increase the delivered cost of the fuel. EPA estimates that
long-range and local distribution from the point of entry would add
$0.03 per gallon to the cost of the methanol delivered to the service
station. Since 2.2 times as much methanol as Diesel would have to be
delivered, this would result further increase the difference in costs.
Differences in actual retail price are much harder to estimate, as
they would depend on the markup demanded (which in turn would depend
on the competitive situation), as well as any differences in taxation.
Capital and Maintenance Costs
In volume production, the costs of a methanol-fueled vehicle using an
Otto-cycle engine would probably be about the same as those of a
Diesel. For a methanol-Diesel engine, the costs would be several
hundred dollars higher (or about 2 to 5%), due to the added cost of
the more sophisticated fuel injection equipment and the ignition
system. Retrofit costs for either system are estimated at about
$2,000 to $6,000 per vehicle, depending on the vehicle type and
requirements.
Maintenance costs with methanol-fueled engines could be higher than
those for Diesels. Methanol combustion products are fairly corrosive,
and experience with certain engine designs tested to date suggests
that engine life for these designs is likely to be considerably
shorter. Other designs have experienced no significant added wear.
Oil change intervals may also need to be shortened to counteract the
increased corrosiveness of the combustion products. Fuel injection
equipment in most methanol-Diesel engines tested to date has suffered
frequent blockage due to deposits, and may also wear out more quickly
-102-
-------
than in the standard engine, due to the poor lubricating qualities of
methanol.
Current Activities
Much work on heavy-duty methanol engines is underway in the U.S.,
where methanol has been targeted for some time as the alternative fuel
of choice. Nearly 100 methanol buses are in use in various
demonstration fleets in the U.S. Most of these buses are equipped
with Detroit Diesel two-stroke Diesel engines using compression-
ignition (augmented by glow plugs at light loads). Detroit Diesel
Corporation has stated publicly that it plans to offer only these
engines for use in buses subject to the U.S. 1991 bus emissions
standards. A significant number of MAN methanol-Dlesel engines using
spark-ignition are also in operation.
Daimler-Benz has demonstrated the use of methanol with ignition-
improving additives in a number of engines in South Africa and
elsewhere; and Cummins Engine in the U.S. is using the same approach
to develop a methanol-Diesel version of its L-10 bus engine. Two such
engines are currently being demonstrated as part of the MILE (Methanol
in Large Engines) program in Canada. Two Caterpillar methanol-Diesel
engines, using glow-plug ignition, are also being demonstrated in that
program.
6.k Ethanol
Ethanol has attracted considerable attention as a motor fuel due to
the success of the Brazilian Proalcool program. Despite the technical
success of this program, however, the high cost of producing ethanol
(compared to hydrocarbon fuels) means that it continues to require
heavy subsidization.
Fuel Properties
As the next higher of the alcohols in molecular weight, ethanol
resembles methanol in most combustion and physical properties. The
major difference is in the higher volumetric energy content of
ethanol. Fuel grade ethanol, as produced in Brazil, is produced by
distillation, and contains several volume percent of water.
Engine Technology and Emissions
Technology for ethanol utilization in heavy-duty engines is
essentially the same as for methanol. Both Otto-cycle engines and
ethanol-Diesel engines (using spark ignition, glow plugs, or ignition-
improving additives) have been developed. Emissions from these
engines are not well characterized, but are believed to be high in
unbumed ethanol, acetaldehyde, and other aldehydes. These could
presumably be controlled with a catalytic converter (subject to the
-103-
-------
concerns regarding aldehyde production discussed above for methanol).
NOx emissions are somewhat higher than for methanol, but still
considerably lower than for Diesel engines. Particulate emissions, as
for methanol engines, should be effectively nil.
Fuel Supply and Costs
Ethanol is produced primarily by fermentation of starch from grains or
sugar from sugar cane. As a result, the production of ethanol for
fuel is in direct competition with the food production in most
countries. The resulting high price of ethanol (ranging from $1.00 to
$1.60 per gallon in the U.S. in the last few years) has effectively
ruled out its use as a motor fuel for any but very specialized
applications.
The Brazilian Frooalcool program to promote the use of fuel ethanol in
motor vehicles in that country has attracted worldwide attention as
the most successful example of an alternative fuel implementation
program extant. Despite the availability of a large and inexpensive
biomass resource, however, this program still depends on massive
government subsidies for its viability, and cannot be considered a
success in any economic sense.
Capital and Maintenance Costs
Capital costs for ethanol-fueled heavy-duty vehicles should be
essentially the same as those for methanol vehicles. Maintenance
costs should be somewhat lower, however. Ethanol's combustion
products are less corrosive than those of methanol, so the adverse
impact on engine life and oil change frequency should be less. The
lubricating qualities of ethanol are essentially the same, however, so
that increased fuel injection system wear would still be expected in
ethanol-Diesel engines.
Current Activities
The Proalcool program in Brazil has led to some efforts to develop
and market ethanol-fueled heavy-duty vehicles. A number of heavy-duty
vehicles using converted gasoline-type engines are in use in Brazil,
and engine manufacturers have experimented with ethanol-Diesel engines
using fuel additives and spark ignition. The drop in world oil
prices, coupled with Proalcool's success in promoting ethanol use in
light-duty vehicles, led the government to de-emphasize the program,
however, and such development has largely stopped. Present
alternative fuel development for heavy-duty engines in Brazil is
focussed on compressed natural gas.
-104-
-------
7. REFERENCES
Bergmann, H., and B. Busenthur (1986). "Facts Concerning the
Utilization of Gaseous Fuels in Heavy Duty Vehicles" from Gaseous
Fuels For Transportation I. Institute of Gas Technology, Chicago,
IL.
Cadman, W. and J.H. Johnson (1986) "The Study of the Effect of Exhaust
Gas Recirculation on Engine Wear in a Heavy-Duty Diesel Engine
Using Analytical Ferrography", SAE Paper No. 860378, Society of
Automotive Engineers, Warrendale, PA.
California Air Resources Board staff (1984). "Status Report: Diesel
Engine Emissions Reductions through Modification of Motor Vehicle
Diesel Fuel Specifications," CARB, Sacramento, CA.
California Energy Commission (1989) AB 234 Report: Cost and
Availability of Low-Emission Motor Vehicles and Fuels, draft
report, CECm, Sacramento, CA.
Cartellieri, W.P., and W.F. Wachter (1987) "Status Report on a
Preliminary Survey of Strategies to Meet US-1991 HD Diesel Emission
Standards Without Exhaust Gas Aftertreatment", SAE Paper No.
870342, Society of Automotive Engineers, Warrendale, PA.
Cometti, G.M., K. Klein, G.J. Frankle, and H.J. Stein (1988). "US
Transient Cycle Versus ECE R.49 13-Mode Cycle," SAE Paper No.
880715, Society of Automotive Engineers, Warrendale, PA.
DeLuchi, M. A., D. Sperling, and R. A. Johnston (1987). A Comparative
Analysis of Future Transnoratlon Fuels. Research Report UCB-ITS-
RR87-13, Institute of Transportation Studies, Berkeley, CA.
Difiglio, C. and M. F. Lawrence (1987). "Economic and Security Issues
of Methanol Supply" in METHANOL-Promise and Problems. SAE Paper No.
872062, Society of Automotive Engineers, Warrendale, PA.
Draper, W.M., J. Phillips, and H.W. Zeller (1988). "Impact of a
Barium Fuel Additive on the Mutagenicity and Polycyclic Aromatic
Hydrocarbon Content of Diesel Exhaust Particulate Emissions," SAE
Paper No. 881651, Society of Automotive Engineers, Warrendale, PA.
Engine Manufacturer's Association (1985). "Heavy-Duty Diesel Engine
In-Use Emission Testing Meeting," briefing package for a
presentation to the California Air Resources Board Staff, El Monte,
CA.
-105-
-------
Engine Manufacturer's Association (1983). Ringlemann Numbers Vs.
Opacity, photo copy, EMA, Chicago, IL.
Gill, A.P. (1988). "Design Choices for 1990's Low Emission Diesel
Engines," SAE Paper No. 880350, Society of Automotive Engineers,
Varrendale, PA.
Gulati, S. and R.P. Merry (1984). "Design Considerations for Mounting
Material for Ceramic Wall-Flow Diesel Filters", SAE Paper No.
840074 in Diesel Particulate Traps. P-140, Society of Automotive
Engineers, Varrendale, PA.
Hardenberg, H.O. (1987) "Urban Bus Application of a Ceramic Fiber Coil
Particulate Trap", SAE Paper No. 870011 in Diesel Particulates: An
Update. SP-702. Society of Automotive Engineers, Warrendale, PA.
Hardenberg, H.O., H.L. Daudel, and H.J. Erdmannsdorfer (1987a)
"Experiences in the Development of Ceramic Fiber Coil Particulate
Traps", SAE Paper No. 870015 in Diesel Particulates: An Update. SP-
702, Society of Automotive Engineers, Varrendale, PA.
(1987b) "Particulate Trap Regeneration
Induced by Means of Oxidizing Agents Injected Into the Exhaust
Gas", SAE Paper No. 870016 in Diesel Particulates: An Update. SP-
702, Society of Automotive Engineers, Varrendale, PA.
Ishida, A., T. Kanamoto, and S. Kurihara (1986). "Improvements of
Exhaust Gas Emissions and Cold Startability of Heavy Duty Diesel
Engines by New Injection-Rate-Control Pump," SAE Paper No. 861236,
in Combustion. Heat Transfer and Analysis. P-182, Society of
Automotive Engineers, Varrendale, PA.
Kittelson, D.B., and D.Y.H. Pui (1986). "Electrostatic Collection of
Diesel Particles," SAE Paper No. 860009 in Advances in Diesel
Particulate Control. P-172, Society of Automotive Engineers,
Varrendale, PA.
Plee, S.L., T. Ahmad, and J.P. Myers (1983). "Diesel NOx Emissions -
A Simple Correlation Technique for Intake Air Effects," in
Proceedings of Nineteenth International Symposium on Combustion,
pp. 1495-1502, The Combustion Institute, Pittsburgh, PA.
Polach, W. and K-H Hagele (1984). "Erste Ergebnisse Mit Dem
Elektrostatischen RuBabscheiden An Einem PKV-Dieselmotor," FISITA
Paper No. 845080, Proc. XX FISITA Conference, 1984.
Shiga, S., H. Ehara, T. Karasawa, and T. Kurabayashi (1985). "The
Control of Diesel Combustion by Means of Exhaust Gas
Recirculation," SAE Paper No. 851544, Society of Automotive
Engineers, Varrendale, PA.
-106-
-------
Shimada, T., K. Sakai, and S. Kurihara (1986). "Variable Swirl Inlet
System and Its Effect on Diesel Performance and Emissions," SAE
Paper No. 861185, in Combustion. Heat Transfer and Analysis. P-182,
Society of Automotive Engineers, Warrendale, PA.
Ullman, T.L., C.T. Hare, and T.M. Baines (1984). "Influence of
Maladjustment on Emissions from Two Heavy-Duty Diesel Bus Engines,"
SAE Paper No. 840416, Society of Automotive Engineers, Warrendale,
PA.
U.S. Environmental Protection Agency (1989) "Analysis of the Economic
and Environmental Effects of Methanol as an Automotive Fuel", Ann
Arbor, Michigan.
van der Weide, J., J.J. Seppen, J.A.N, van Ling, and H.J. Dekker
(1988). "Experiences with CNG and LPG Operated Heavy Duty Vehicles
with Emphasis on US HD Diesel Emission Standards," SAE Paper No.
881657, Society of Automotive Engineers, Warrendale, PA.
Wade, W.R., T. Idzikowski, C.A. Kukkonen, and L.A. Reams (1985)
"Direct Injection Diesel Capabilities for Passenger Cars", SAE
Paper No. 850552, Society of Automotive Engineers, Warrendale, PA.
Wade, W.R., C.E. Hunter, F.H. Trinker, and H.A. Cikanek (1987) "The
Reduction of NOx and Particulate Emissions in the Diesel Combustion
Process", ASME Paper No. 87-ICE-37, American Society of Mechanical
Engineers, New York, NY.
Wagner, T. 0., and D. F. Tatterson (1987). "Comparative Economics of
Methanol and Gasoline" SAE Technical Paper Series, No. 871579,
Society of Automotive Engineers, Warrendale, PA.
Wall, J.C. and S.K. Hoekman (1984). "Fuel Composition Effects on
Heavy-Duty Diesel Particulate Emissions," SAE Paper No. 841364,
Society of Automotive Engineers, Warrendale, PA.
Wallace, F.J., D. Howard, E.W. Roberts, and U. Anderson (1986).
"Variable Geometry Turbocharging of a Large Truck Diesel Engine,"
SAE Paper No. 860452, Society of Automotive Engineers, Warrendale,
PA.
Weaver, C.W., C. Miller, W.A. Johnson, and T.S. Higgins (1986).
"Reducing the Sulfur and Aromatic Content of Diesel Fuel: Costs,
Benefits, and Effectiveness for Emissions Control," SAE Paper No.
860622, Society of Automotive Engineers, Warrendale, PA.
Weaver, C.S. and R.F. Klausmeier (1987). A Study of Excess Motor
Vehicle Emissions. Task 1: Heavy Duty Diesel NOx and Particulate
Control Technologies. Draft Final Report submitted to the
California Air Resources Board.
-107-
-------
Weaver, C.S. and R.F. Klausmeier (1988). Heaw-Duty Diesel Vehicle
Inspection and Maintenance Studv: Final Report (4 Volumes), report
under ARB Contract No. A4-151-32, Radian Corporation, Sacramento,
CA.
Williams, P.T., G.E. Andrews, and K.D. Bartle (1987) "Diesel
Particulate Emissions: The Role of Unburnt Fuel in the Organic
Fraction Composition", SAE Paper No. 870554, Society of Automotive
Engineers, Warrendale, PA.
World Bank (1986). Urban Transport. A World Bank Policy Study, The
International Bank for Reconstruction and Development/The World
Bank, Washington, DC.
Yu, R.C. and S.M. Shahed (1981) "Effects of Injection Timing and
Exhaust Gas Recirculation on Emissions from a DI Diesel Engine",
SAE Paper No. 811234, Society of Automotive Engineers, Warrendale,
PA.
-108-
------- |