Draft Regulatory Impact Analysis
Control of Sulfur and
Aromatics Contents of On-Highway
Diesel Fuel
July 1989
U.S. Environmental Protection Agency
Office of Air and Radiation
Office of Mobile Sources
-------
Draft Regulatory Impact Analysis
Control of Sulfur and
Aromatics Contents of On-Highway
Diesel Fuel
July 1989
U.S. Environmental Protection Agency
Office of Air and Radiation
Office of Mobile Sources
-------
TABLE OF CONTENTS
Page
Chapter 1: Introduction 1-1
I. Background 1-1
II. Diesel Fuel Control Options 1-4
III. Structure of Report 1-6
Chapter 2: Refinery Impacts of Controlling Sulfur and
Aromatics in Diesel Fuel 2-1
I. Trends in Commercial Diesel Fuel Properties . 2-1
A. Current Fuel Quality Standards 2-1
B. Commercial Diesel Fuel Survey Results . 2-2
II. Refinery Cost Impacts 2-5
A. Background 2-5
B. Energy Resource Consultants/Sobotka
Study 2-9
C. Bonner and Moore's Study. ....... 2-12
D. National Petroleum Refiners Association
Survey 2-19
E. Best Estimate Refinery Cost Estimates . 2-23
III. Effect of Hydrodesulfurization on Fuel
Aromatics 2-25
IV. Small Refinery Impact 2-29
V. Leadtime Requirements 2-31
Chapter 3: Engine Control Technology and Cost Prior to
Fuel Control 3-1
I. Engine-Out Particulate Emissions - Current
Fuel 3-1
A. Heavy-Duty Diesel Engines 3-1
B. Light-Duty Diesel Emissions 3-27
II. Exhaust Aftertreatment Technology 3-29
A. Exhaust Aftertreatment Types 3-31
B. Exhaust Aftertreatment Efficiencies . . 3-32
C. Exhaust Aftertreatment Device Costs . . 3-41
D. Exhaust Aftertreatment Device
Deterioration 3-48
III. Aftertreatment Technology Mix for Compliance
with Standards 3-49
A. Methodology 3-49
B. Results 3-50
Appendix 3-A: Projected 1991 and 1994
HDDE Particulate Emission
Distribution 3-A
Chapter 4: Effect of Fuel Quality on Emissions and
Engine Cost 4-1
I. Effect of Fuel Properties on Diesel Emissions 4-1
A. Effect of Fuel Sulfur on Highway
Diesel Emissions 4-2
B. Effect of Fuel Aromatics on Highway
Diesel Emissions 4-6
II. Effect of Fuel Quality on Off-Highway Diesel
and No. 2 Fuel Oil Emissions 4-14
-------
TABLE OF CONTENTS (Cont'd)
Page
III. Diesel Aftertreatment Technology With Sulfur
Control 4-15
A. Light-Duty Diesels 4-15
IV. Diesel Aftertreatment Technology Fuel
Aromatics Control 4-18
V. Sensitivity Analysis 4-21
Chapter 5: Effect of Fuel Modification on Engine Wear. . 5-1
I. Introduction 5-1
II. Effect of Fuel Sulfur on Engine Wear 5-1
A. Background 5-1
B. Experimental Results 5-2
C. Used Oil Analyses 5-3
III. Effect of Reduced Wear on Operating Costs . . 5-7
A. Oil Change Cost Reduction 5-10
B. Extension in Engine Rebuild Interval/
Vehicle Life 5-16
Chapter 6: Effect of Fuel Modifications on Air Quality,
Public Health, and Welfare 6-1
I. Emissions Inventories 6-1
A. Emission Sources 6-2
B. General Inventory Estimation Methodology 6-2
C. Fuel Consumption 6-3
D. Emission Factors 6-5
E. Emission Inventory Results 6-15
II. Effects of Emission Changes on Pollutant
Ambient Concentrations 6-22
A. Indirect Sulfate Analysis 6-23
B. Air Quality Impacts 6-30
III. Cancer Risk Assessment 6-40
A. Population Exposure Analysis 6-40
B. Cancer Assessment 6-46
IV. Visibility Assessment 6-47
A. Methodology 6-47
Appendix 6-A: Diesel Fleet VMT, Fuel 6-A
Consumption, Registrations Fuel
Economy, and Annual VMT/Veh
Calculations
Chapter 7: Cost Effectiveness of Fuel Controls
A. Overview
B. Methodology
-------
Chapter 1
Introduction
I. Background
The analyses in this document were originally prepared in
1987 and revised in early 1988. However, since then, due to
more recent technical information and the July 19, 1988
agreement by the oil and engine industries, some of the
assumptions about the use of control technology are now
outdated. These assumptions relate to both when add-on
technology will be needed and the types of technology likely to
be used. This most significantly affects the analysis of the
cancer impacts of sulfur control. The Agency no longer
considers this analysis valid. It is included here only to
allow for comments on the methodology used. The other analysis
in this document are much less sensitive to the revised
assumptions. Economically, these newer assumptions would lead
to slightly better cost effectiveness than is reported here.
This is not considered significant since, as will be shown
later, sulfur control is already considered to be very cost
effective. Given the relative insensitivity of these analyses
to the changing assumptions, and the need to provide sufficient
leadtime for the 1994 model year. This document is being
published without revision; however, all the analyses will be
updated for the final rule.
Diesel engines contribute a significant amount of
particulate emissions to the nation's total particulate
emission inventory. Much of these emissions occur in urban
areas where population density, and thus, personal exposure to
diesel particulate emissions are higher than in non-urban
areas. The particulate emissions from diesel engines are all
less than 10 microns in diameter, which means that when they
are inhaled, the particulates are deposited deep in the lung
tissue. This can cause aggravation of respiratory conditions,
reductions in lung functions, increased susceptibility to
infections, structural changes in lung tissue and alveolar
macrophage damage. Diesel particulate has also been identified
as having probable carcinogenic effects. Finally, diesel
particulate reduces visibility (primarily in urban areas), and
can cause soiling of materials.
Diesel particulate is made up of three basic components,
or fractions — the carbonaceous fraction, the soluble organic
fraction (SOF), and the sulfate fraction. Carbon particulates
are formed from incomplete combustion of hydrocarbon molecules
which make up diesel fuel. The soluble organic fraction is not
really a separate particulate, but consists of unburned
hydrocarbons which are absorbed onto the carbon particulate
(and thereby add to the weight of the carbon particulate). The
sulfate particulate is predominately condensed sulfuric acid
-------
1-2
gas, which is derived from sulfur in diesel fuel. Sulfur in
the fuel oxidizes to form SO2• Some of the SO2 further
oxidizes to form S03, which quickly reacts with water to form
sulfuric acid. Some of the sulfuric acid can react with wear
products (iron, chrominum), forming metal sulfates. Gaseous
products emitted by diesel engines are hydrocarbons, carbon
monoxide (CO), nitrous oxides (NOx), and (as mentioned earlier)
sulfur dioxide (S02).
Due to statutory requirements in the Clean Air Act (CAA)
mandating reductions in NOx and particulate emissions, and the
projected growth in the diesel vehicle population, in March of
1985 EPA promulgated emission standards for heavy-duty trucks
and buses.[1] These standards were expected to bring about
technological changes in diesel engines and exhaust treatment
that would lead to a reduction in particulate emissions. These
technologies are concentrated in two primary areas — the
development of strategies to reduce engine-out emissions
(combustion chamber modifications, turbocharging and
aftercooling, electronic controls, etc.), and the development
of technology to "trap" and burn diesel particulate in the
exhaust system. These technologies affect the various
particulate types in different manners. For example,
combustion chamber modifications which reduce "dead space"
result in a greater reduction in SOF than carbon or sulfate
particulate by reducing the quantity of unburned hydrocarbon
emissions which are available to adsorb onto the carbon
particulate.
Devices used to reduce particulate in the exhaust system
are called aftertreatment devices, and there are three primary
types. Uncatalyzed traps are devices which filter or trap
diesel particulate until the trap becomes loaded to a certain
point. Then the trapped particulate is burned off periodically
by a regeneration system, which is needed because diesel
exhaust is not usually hot enough by itself to burn the
particulate. Next there are catalyzed traps, which function
quite similarly to uncatalyzed traps. The catalyst simply
reduces the ignition temperature of the carbon particulate,
thereby reducing or eliminating the need for a separate
regeneration system. The catalyst can also reduce HC emissions
and thus, SOF. Last, there are flow-through oxidation
catalysts, which are used extensively in light-duty gasoline
vehicles and trucks. Oxidation catalysts are effective in
reducing gaseous HC emissions, which reduces SOF. Generally,
initial and maintenance costs are highest in the trap systems,
and much lower for catalysts. Also, most trap systems increase
exhaust backpressure, and thereby increase fuel consumption
slightly. There is little or no impact on fuel consumption
from oxidation catalysts.
-------
1-3
Although in-cylinder and aftertreatment controls are known
to have some effect on carbon and SOF emissions, neither kind
o£ control has been demonstrated to be effective at reducing
sulfate particulate. Furthermore, some catalyzed traps and
oxidation catalysts are known to promote further oxidation of
SO2 to sulfate particulate. This situation can lead to
sulfate concentrations being higher in the aftertreatment
outlet than the inlet, thereby reducing the overall
effectiveness of the aftertreatment device. So far, the only
demonstrated way to remove sulfate particulate is through
reducing the sulfur content of diesel fuel.
At the time the particulate emission standards were
proposed, several engine manufacturers raised the issue of the
impact of diesel fuel sulfur levels on emissions. According to
their comments, sulfur from fuel could be a problem either
through trap plugging from engine-out sulfate emissions, or
through the generation of significant measurable particulate
sulfate emissions which would make it impossible to meet the
stringent emission standards. In the preamble to the final
rule on particulate standards, EPA's response was that it would
continue to investigate this issue and resolve any continuing
difficulties. EPA further stated, "while it does not appear at
present that regulating sulfur content of diesel fuel is a
prerequisite to the feasibility of traps, if it is shown to be
necessary based on ... further analysis, EPA will investigate
potential action under Section 211(c) of the Act."[l]
Shortly after the particulate standards were finalized,
EPA initiated a preliminary study to examine the impacts of
diesel fuel controls on a number of areas. [2] The study
explored some of the costs and benefits of sulfur and aromatics
controls. Included were estimates of potential emission
reductions, the costs of control to refiners, and engine wear
benefits. Major conclusions of the study were:
0 Reducing sulfur in diesel fuel would significantly
reduce sulfate particulate and SO2 emissions.
The sulfur content of diesel fuel could be reduced
from 0.27 wt. percent to 0.05 wt. percent and
aromatics from about 30 percent to 21 percent for a
cost of about 1.2 cents per gallon. However, this
would require extensive segregation of distillate
and burner fuels.
0 Reducing the sulfur content of diesel fuel would
reduce corrosive wear in engines, bringing about
lower cost oil changes and improved engine life,
leading to a savings of approximately four times the
above refinery cost.
-------
1-4
In order to obtain comments on the technical aspects of
this study, EPA published a Federal Register notice of the
availability of the report and opened a docket (A-86-03) to
receive comments.[3] Many comments were received from industry
and other interested parties. These comments identified a
number of areas where the analysis could be improved. Also,
due to the fact that this was a preliminary study, certain
items were not investigated in the analysis. For example, the
study did not address the effect that fuel controls could have
on the cost to engine manufacturers to comply with the 1991 and
1994 particulate standards. Some manufacturers have stated
that the availability of low sulfur fuel could enable the use
of oxidation catalysts on diesel vehicles, presumably at a
lower cost than trap systems. Also, the study did not address
the impacts fuel controls could have on air quality, health,
and welfare.
EPA therefore initiated a second and more comprehensive
study of diesel fuel controls, which is contained in this draft
Regulatory Impact Analysis (RIA). Like the preliminary study,
this draft RIA examines sulfur and aromatics controls.
However, additional refinery modeling has been conducted to
address concerns raised in the preliminary study. A more
thorough analysis of the effect of reduced sulfur levels on
engine wear and its effects on oil change cost and engine life
has been performed. The analysis also includes an assessment
of the impacts that fuel controls would have on the types of
af tertreatment technology needed to meet the 1991 and 1994
standards, and resultant costs. And finally, the effects of
fuel controls on air quality, health and welfare have been
determined.
II. Diesel Fuel Control Options
Two basic diesel fuel control options are explored in this
RIA. The first one is sulfur control. The current diesel fuel
sulfur content is about 0.25 wt. percent. This study examines
sulfur controls down to as low as 0.05 wt. percent. The second
control option is the addition of aromatics control to the
above sulfur control. Aromatics control alone is not treated
as a separate option, because it is more expensive and less
cost effective than sulfur control, and therefore would not be
viewed as an alternative to sulfur control. The current
national average aromatics level of highway diesel fuel is
about 35 volume percent. This draft RIA examines control
levels in the vicinity of 20 volume percent.
Section 211(c) of the CAA gives the EPA administrator the
authority to regulate fuels used in motor vehicles. Therefore,
the diesel fuel controls discussed in this study would apply
only to on-highway diesel fuel. However, a significant amount
of diesel fuel is consumed by off-highway sources
(agricultural, construction equipment, etc.). Also, fuel oil,
which is used as a burner fuel in residential furnaces and
-------
1-5
commercial and industrial boilers, is very similar (and
identical in many applications) to diesel fuel. Therefore,
many refiners sell diesel fuel and fuel oil from a common
middle distillate pool. Refiners would have a choice of how
much fuel to treat, in the range of their highway diesel fuel
alone up to their entire common distillate pool. The amount of
fuel treated might vary considerably from refinery to refinery
depending on many factors, which are discussed in Chapter 2.
However, it is likely that, in many refineries, more than just
on-highway diesel fuel will be treated. The approach used in
this analysis is to include the costs of treating this "extra"
fuel (above on-highway fuel demand), and also to account for
any benefits that accrue, such as reduced SO2 emissions from
residential furnaces and off-highway vehicles, etc.
Sulfur Control - Removing sulfur from diesel fuel would
require the refinery industry to construct a certain amount of
new sulfur refining capacity. This must be amortized over the
fuel thus treated, thereby raising its price. Thus, the
primary cost of sulfur controls would be observed in the
increase in the price of middle distillate products (diesel
fuel and fuel oil).
Many of the major diesel engine manufacturers have stated
that if the sulfur is removed from diesel fuel, their engines
will be able to meet the 1991 particulate standards without
aftertreatment, and could also meet the 1994 standards with
oxidation catalysts instead of with traps. This would result
in a lower cost per vehicle to meet the 1991 and 1994
standards, an obvious economic benefit of sulfur control.
However, sulfur controls, if promulgated, could not be
implemented until around 1993 because the refinery industry
would need leadtime to construct new capacity. Therefore, one
action also being considered is whether to lower the sulfur
content of certification fuel (in 1991) prior to implementation
of in-use sulfur control.
Removing sulfur from diesel fuel (or middle distillate
fuels) would have a number of emission benefits. First, it
would reduce sulfate particulate from all highway, off-highway
and stationary sources that use the treated fuel. Second,
SO2 emissions from these sources would also be reduced. Much
of the SO2 reacts in the atmosphere to form further sulfate
particulate, and this would also be reduced, leading to lower
ambient particulate levels and improved visibility. If sulfur
controls lead to significant use of oxidation catalysts or
catalyzed traps on diesel trucks these technologies would
reduce HC emissions, leading to lower particulate SOF and
reduced cancer incidences.
-------
1-6
Finally, a potentially sizable benefit of diesel fuel
sulfur control is reduced engine wear. There is a growing body
of evidence that indicates that reduced sulfur levels could
lead to lower corrosive wear in diesel engines, thereby
possibly extending engine and vehicle life. Or, the reduced
sulfur levels could enable lubrication oil producers to produce
oils at a lower cost, or allow truck owners and operators to
extend oil change intervals, thereby reducing oil change costs.
Sulfur and Aromatics Control - Removing aromatics from
diesel fuel would also require the refinery industry to install
additional refinery capacity. However, the removal of
aromatics is more difficult than removing sulfur, so the
refining costs are higher. These costs would have an impact on
diesel fuel prices.
Reducing the aromatics content of diesel fuel reduces
carbon and SOF particulate emissions from diesel engines
somewhat. This has a number of implications. First, reducing
in-use fuel aromatics levels would reduce emissions from trucks
currently on the road, thereby improving ambient SOF and carbon
concentrations. The reduction in ambient SOF could further
lead to a reduction in cancer incidences.
The reduction in certification fuel aromatics levels would
reduce engine-out emissions of engines designed to meet the
1994 particulate standards. With lower engine-out levels, less
aftertreatment control would be needed to meet the standards,
thereby yielding a vehicle cost benefit. In the long term,
ambient levels of carbon and SOF would be essentially equal
with or without aromatics control since under either scenario
engines must meet the same exhaust standard in 1994. The cost
of aromatics control, therefore, must be evaluated against the
potential technology savings in obtaining the net cost of
expected emission reductions from aromatics control.
III. Structure of Report
The primary analysis contained in this report focuses on
the cost effectiveness of diesel fuel sulfur and sulfur and
aromatics control. (Cost effectiveness analysis relates the
control cost to the amount of emission control achieved in
terms of dollars per ton of pollutant reduced.)
Chapter 2 discusses the implications of fuel controls on
the refinery industry. First, fuel composition trends are
discussed. Next, costs are reported from recent contractor
studies and compared with other studies. The impacts on small
refiners are also discussed. Last, the effects of fuel
controls on other fuel parameters such as cetane number and
pour point are addressed.
-------
1-7
Chapter 3 discusses diesel engine technology, cost, and
emissions without fuel controls. Engine-out emissions are
developed for four heavy-duty vehicle classes. Aftertreatment
technologies are then discussed, with the focus being on the
efficiencies of each technology on the various particulate
fractions, and the mix of technologies needed to meet the 1991
and 1994 standards. Finally, the per vehicle technology costs
to meet the 1991 and 1994 standards on current diesel fuel are
estimated.
Chapter 4 presents the effect of fuel controls on diesel
engine technology, cost and emissions. The effect of fuel
controls on engine-out emissions is examined first. Then the
information on aftertreatment efficiencies and costs developed
in the previous chapter is used to develop per vehicle
emissions and control technology costs with the two fuel
control cases.
Chapter 5 presents an analysis of how reducing diesel fuel
sulfur levels can reduce engine wear, thereby extending engine
and vehicle life or reducing lubricant oil costs and improving
oil change intervals. The first part of the chapter discusses
a contractor analysis of experimental and in-use wear data.
The second part of the chapter develops potential operating
cost reductions which could be experienced by truck owners and
operators with low sulfur fuels.
The effect of fuel modifications on air quality, public
health and welfare are discussed in Chapter 6. The first part
presents inputs used in inventory modeling of both mobile and
stationary sources. The second part estimates the effect fuel
controls would have on urban and rural particulate and SO2
concentrations. This is followed by a discussion of SO2
conversion to sulfate particulate in urban areas, and the
resultant effect on urban particulate concentrations. Next,
the impacts of reduced SOF concentrations on cancer incidences
is discussed. The last two sections estimate welfare benefits
such as improved visibility and reduced national SO2
inventories.
The analysis of diesel fuel control alternatives is
presented in Chapter 7. The effect of sulfur and aromatics
control on the costs for heavy duty trucks to meet the 1991 and
1994 particulate standards are examined. Also examined is the
net cost effectiveness of sulfur and aromatics controls,
estimated in total dollars spent per ton of pollutant reduced.
A complete discussion of how the Agency used this study in
selecting specific options for proposal is found in the
preamble to the associated Notice of Proposed Rulemaking itself.
-------
1-8
References (Chapter 1)
1. "Control of Air Pollution From New Motor Vehicles
and New Motor Vehicle Engines; Gaseous Emission and Particulate
Emission Regulations," Federal Register, 10606, Friday, March
15, 1985.
2. "Diesel Fuel Quality Effects on Emissions,
Durability Performance, and Costs," ERC, Sobotka, EPA Contract
#68-01-6543. Available in Docket #A-86-03.
3. "Diesel Fuel Quality Effects on Emissions,
Durability, Performance and Costs; Availability of Preliminary
Study," Federal Register, 23437, Friday, June 27, 1986.
-------
Chapter 2
Refinery Impacts of Controlling
Sulfur and Aromatic in Diesel Fuel
As discussed in Chapter l, two diesel fuel properties have
been identified as affecting diesel particulate emissions. The
control of these two properties, sulfur and aromatics content,
shall be investigated in this chapter. The chapter begins with
a summary of the results of two commercial fuel surveys which
track and identify trends for various diesel fuel parameters
(Section I). Section II begins with background on the
production and distribution of diesel fuel as well as a general
discussion of the refinery methods which can be used to reduce
sulfur and aromatic content. This section also presents and
evaluates a number of studies and surveys concerning the
refinery cost estimates of controlling these diesel fuel
parameters. These studies will be discussed in detail in this
section and the results of the studies will be used to
determine an estimate of the total U.S. refinery cost for
on-highway diesel fuel modifications. Section III deals with
the effects of hydrotreating on the types of aromatic species
in diesel fuel. Section IV discusses the results of an
analysis addressing the cost impact of diesel fuel regulations
on small refineries in the U.S. Section V, the final section,
presents an analysis of the leadtime requirements for
implementation of fuel controls.
I. Trends in Commercial Diesel Fuel Properties
In considering diesel fuel quality standards, it is
important to identify existing diesel fuel regulations, and
also to examine any trends in diesel fuel quality which can be
determined from fuel survey data. A synopsis of the current
diesel fuel quality standards and the available survey data are
presented below.
A. Current Fuel Quality Standards
Several states have established legal requirements for
diesel fuel. In the northeast, regulations on the sulfur
content of heating oil force average distillate sulfur levels
down. California has mandated that no diesel fuel sold in the
South Coast Air Basin shall have a sulfur content which exceeds
0.05 percent by weight. Recently, California has promulgated
regulations limiting highway diesel sulfur and aromatics
content throughout the entire state. These regulations will
take effect in 1993. In most states, however, strict
requirements for diesel fuel quality have not been legislated.
-------
2-2
ASTM document D-975 lists recommended fuel quality
standards for commercial No. 2 distillate fuel oils. Table 2-1
shows some of the specifications for No. 2 diesel fuel and also
for No. 2 fuel oil. It should be noted that these ASTM
specifications are recommended levels and are not legally
binding for refiners. Table 2-1 also shows all of the
specifications which diesel test fuels are currently required
to meet as part of the Federal Test Procedure (FTP) when new
diesel vehicles or engines are officially tested for exhaust
emissions.
In general, the EPA test fuel standards and the ranges for
the various specifications are more stringent, to limit the
variability in test results due to variations in fuel quality.
For example, the ASTM cetane specification for diesel fuel is
40 (minimum) while FTP fuels require 42-50. There is no cetane
specification for No. 2 fuel oil. ASTM recommends that diesel
fuels and fuel oil do not contain greater than 0.50 weight
percent sulfur, compared to a range for the FTP fuel of 0.2-0.5
weight percent. ASTM has no recommendation with regard to
aromatic content while an EPA test fuel must have a minimum of
27 volume percent aromatics.
B. Commercial Diesel Fuel Survey Results
This section will focus on the trends of diesel fuel with
respect to three fuel parameters; sulfur content, aromatic
content, and cetane number. Sulfur and aromatic content are
reported here due to the correlation of these two parameters
with particulate emissions, as discussed in Chapters 1 and 4.
Cetane number will also be shown since it is an indicator of
the fuel's ignition quality and tends to vary inversely with
aromatic content. Results of surveys which have been conducted
by two organizations will be reported. One is the National
Institute for Petroleum and Energy Research (NIPER) survey,
which is conducted annually. The second is a semi-annual
survey conducted by the Motor Vehicle Manufacturers Association
(MVMA).
Table 2-2 lists minimum, average, and maximum levels from
the NIPER survey as well as the MVMA survey. As can be seen,
cetane number has been steadily declining over the past few
decades, while sulfur content, for the most part, has been
relatively constant, with only a very slight increase.
Table 2-2 also shows trends in aromatic content. MVMA
measures aromatic content as part of their survey, and the
results indicate that aromatic content may have increased only
slightly during the 1980's. The NIPER survey does not measure
aromatics, although the reported decline in cetane over the
last 25 years suggests that aromatics levels may have been
increasing.[I]
-------
2-3
Table 2-1
Specifications for Highway Diesel No.2 and No.2 Fuel Oil
No.2 Diesel
Cetane Number
Distillation Range:
IBP, °F
10% point, °F
50% point, °F
90% point, °F
EPA, °F
Gravity, °API
ASTM'
40 (min.)
540-640
Total Sulfur, weight. % 0.50 (max.)
(max.)
Aromatics, Min. Vol. %
Flashpoint, °F, Min.. 125
Viscosity, CST. (40°C) 1.9-4.1
EPA
* *
42-50
340-400
400-460
470-540
550-610
580-660
33-37
0.2-0.5
27
130
2.0-3.2
No.2 Fuel Oil
ASTM*
540-640
30 (min)
0.50
100
1.9-3.4
* ASTM D 975.
** CFR Title 40, Part 86. Applies only to emission test
fuels.
-------
2-4
Table 2-2
Trends in Diesel Fuel Quality
NIPER Survey*
Sulfur Content
Year
(weiqht %)
Cetane Number
1986
. 000-.260-.960
36.8-45.1-54.6
1985
.001-.256-.930
38.0-44.7-52.1
1984
.010-.267-.940
39.0-45.3-53.2
1983
. 040-.275-.960
37.0-45.0-55.4
1982
.020-.272-.970
39.0-44.4-53.0
1981
.010-.283-.950
29.0-45.6-52.4
1980**
.010-.232-.980
39.0-44.9-54.0
1979**
.000-.230-.890
39.0-46.3-55.9
1978**
.010-.230-1.100
39.0-46.3-61.5
1977**
.004-.220-.900
40.0-47.3-66.3
1970***
.22
48. 7
I960***
.23
50.0
MVMA
Survey****
Year/
Sulfur Content
Aromatics
Season
(weiqht %)
Cetane Number
(volume %)
1987/W
. 030-.218-.297
42-44-46
25.9-31.1-38.3
1986/S
.030-.23 8-.3 52
41-44-47
32.4-39.6-43.1
1986/W
.047-.238-.361
42-44-47
26.6-32.3-36.6
1985/S
.034-.237-.389
41-4 5-49
27.9-31.7-38.4
1985/W
.082-.228-.341
42-45-49
26.5-31.6-38.8
1984/S
.140—.277-.403
43-45-48
27.8-31.7-36.1
1984/W
. 119—.220-.338
37-44-50
23.8-31.9-37.7
1983/S
. 168-.344-.498
42-45-53
(not given)
1983/W
.100—.253-.349
42-44-49
24.6-30.4-34.5
1982/S
,151-.299-.462
42-46-53
26.7-32.4-39.3
1982/W
.103-.253-.363
43-47-49
29.4-31.8-35.5
1981/S
. 133-.276-.465
42-47-51
23.3-29.9-36.3
1981/W
. 082-.197-.353
44-46-53
23.9-31.3-35.8
1980/S
.078-.217-.370
44-47-54
23.2-31.6-38.1
* "Diesel Fuel Oils," National Institute for Petroleum and
Energy Research.
** Classification was truck-tractor fuel in these years.
Classification changed to diesel No. 2 in 1980.
*** Approximated from relationship between clear cetane number.
**** "MVMA National Diesel Fuel Survey," MVMA, Inc.
-------
2-5
There are a number of reasons to expect distillate fuel
quality to deteriorate in the future. The processing of
heavier and poorer quality (with regard to sulfur content)
crude oil will likely result in higher distillate sulfur levels
due to the direct blending of more sulfur into the distillate
fraction. Likewise the increased conversion of heavier
components to lighter components should result in lower cetane
and higher aromatics levels in the future. As fluid catalytic
cracking has increased in severity,' in order to produce more
gasoline, the aromatic content of light cycle oil has
increased. Blending of light cycle oil, generally the poorest
quality diesel blending component, into diesel fuel has also
been increasing in recent years resulting in increased aromatic
content and corresponding decreases in cetane number. Without
regulations it is expected that a gradual decline of diesel
fuel quality will occur.[1,2]
11. Refinery Cost Impacts
Several studies have been conducted recently investigating
the refining cost impact of sulfur and aromatic content
restrictions of highway diesel fuel. An initial study of the
issue was performed under contract for EPA by Energy Resource
Consultants and Sobotka and Company (ERC & SCI), and was
released for public comment in June, 1986.[2] SCI responded to
many of the comments which were raised in a follow-up study,
while Bonner & Moore Management Science also prepared a study
under contract for EPA.[3,4] In addition, the National
Petroleum Refiners Association (NPRA) surveyed member refiners
and published a report estimating the cost af diesel fuel
control in 1986.[5] Other studies have also been performed on
the cost of diesel fuel control in limited geographical regions.
In this section, a discussion of the background and key
issues relating to the refining cost of diesel fuel
modifications will be presented, as well as a synopsis of the
aforementioned studies. Following the discussion of each of
the studies, a final analysis will incorporate information from
each in deriving a "best estimate" of refinery costs.
A. Background
Two major issues exist which make estimation of the total
refinery cost of highway diesel fuel control difficult. These
issues were handled differently in each study, and an
understanding of these two issues is required before discussion
of the study results. The first major issue affecting highway
diesel fuel control costs concerns how much fuel will actually
need to be treated if on-highway diesel fuel regulations are
implemented. The light fuel oils distilled during the refinery
process are known as the distillate fuel oils. Included are
products known as No.l, No. 2, and No. 4 fuel oils and No.l,
No.2, and No.4 diesel fuels, conforming to ASTM Specifications
-------
2-6
D396 and D975, respectively. On-highway diesel fuel is one
product within the No. 2 distillate pool, though, as indicated
in Table 2-1, ASTM specifications show No. 2 diesel and No. 2
fuel oil to be essentially the same product. Table 2-3 shows
Department of Energy (DOE) statistics on the end user
categories for the No. 2 distillate pool, including both diesel
and fuel oil.[6] On-highway diesel represents 41 percent of
the No. 2 distillate pool, and 55 percent of the No. 2 diesel
pool. Residential heating oil represents 18 percent of the
total No. 2 distillate pool, but 69 percent of the No. 2 fuel
oil pool.
Comparison of the recommended ASTM specification for No.2
diesel fuel and No. 2 heating oil (Table 2-1) reveals some
differences between the two distillate pools. Diesel fuel must
meet a minimum cetane number of 40 whereas fuel oil has no
cetane number specification. Since fuels which satisfy the
ASTM No. 2 diesel requirements generally satisfy the No.2 fuel
oil specification as well, some refiners choose to produce one
product which is sold to both markets, rather than incur the
cost of segregating the products and blend stocks. The same
storage tanks, loading facilities, and pipelines are used for
the common product. However, if on-highway diesel fuel quality
specifications are mandated, there would be a greater incentive
to separate the products to avoid having to reduce the sulfur
content of the unregulated No. 2 fuel oil, as well as
distillate used for other off-highway applications. If
segregation is not feasible for some or many refineries, some
fuel oil and off-highway distillate may also be reduced in
sulfur or aromatic content if such regulations are imposed on
on-highway diesel, thus increasing the cost of both products.
In its written comments on the ERC study, the American
Petroleum Institute (API) stated that refiners will have to
control the quality of all No.2 diesel products listed in the
left column of Table 2-3, except for railroad, vessel
bunkering, and military use, which are currently
segregated.[7] However, Sobotka and Co. believes that
segregating on-highway diesel fuel through the distribution
system would be possible, but at some additional cost where the
ability does not currently exist.[8] The ability of pipelines
to segregate distillate products would be limited to the
availability of tankage at on-line tank farms, with pipelines
currently shipping products on a fungible basis apparently
having the most difficulty. Additional costs for new tankage
and piping would likely be incurred. The ability of distillate
to be segregated at bulk terminals would be dependent on the
mix of tanks currently used.
In summary, Sobotka and Co. stated that most major
pipelines would have relatively little difficulty in handling
an additional grade of diesel fuel. Bulk terminals would, in
most cases be able to handle another grade of diesel with
relatively low cost as well. If carrying additional grades of
distillate fuel proved to be too costly, bulk terminals would
-------
2-7
Table 2-3
End-Use of No. 2 Distillate Products[6]
Residential
Commercial
Industrial
Oil Company
Farm
Electric Utility
Railroad
Vessel Bunkering
On-Highway Diesel
Military
Off-Highway Diesel
All Other
Total
Percent of Total No.2 Distillate
No.2 Diesel*
0
5
4
2
7
0
7
5
40
2
4
0
76
No.2 Fuel Oil
16
4
2
0
0
1
0
0
0
0
0
1
24
Includes small amounts of No. 1 Distillate and No. 4
Distillate.
-------
2-8
have the option of marketing only one of the distillate
products, depending on the number of terminals nearby and their
response to the added grade.
The segregation options available to the refiner to reduce
sulfur or aromatic content will vary greatly depending on his
raw materials, product slate and the configuration of the
refinery. Where segregation of final products is possible the
refinery may change blending practice to direct low sulfur
blend stocks to the on-highway pool. Directionally, this
practice may have the undesirable side effect of increasing the
sulfur content of the non-highway distillate fuels, although
significant desulfurization will still be required to reduce
the on-highway diesel to 0.05 weight percent. The costs for
fuel control presented in this chapter were generated under the
assumption that the sulfur content of the off-highway pool
would not increase. To the extent the sulfur content of
off-highway distillate does increase in reality, actual
refining costs and many of the resultant environmental benefits
will be commensurately lower.
Since it is impossible to predict at this time the degree
to which segregation would take place if fuel regulations were
promulgated, the effects of fuel segregation on refining cost
has been bracketed for this analysis. As will be presented
later in this Chapter, refining costs were developed under two
segregation scenarios. In the first scenario, it was assumed
that no increase in fuel segregation would take place. In the
second scenario, complete segregation of highway diesel from
all other distillates was assumed. The actual degree of fuel
segregation employed and the actual cost of fuel control will
lie somewhere within this range.
The second major issue concerns the refinery facilities
that are currently available to reduce sulfur in diesel fuels.
The refining unit used to desulfurize fuel is referred to as a
hydrotreater. The hydrotreater removes sulfur by introducing
hydrogen to the fuel and passing it over a catalyst at elevated
pressures and temperatures. Public information on current
industry hydrotreating capacity is difficult to interpret since
the related severity of the process and how the unit is being
used is not reported. Also, little is known publicly about the
capability of existing units, such as the maximum severity an
individual unit can withstand.
Most mid-distillate hydrotreating units currently in
operation may technically be capable of reducing sulfur to
levels as low as 0.05 weight percent, but not without
significant changes in their operation.[9] Running the units
at much lower liquid hourly space velocities (a measure of
residence time) and higher temperatures will increase sulfur
removal, but at a reduction in fuel throughout. A few units
may be able to keep current typical space velocities if they
-------
2-9
can reach the required operating temperature. Either way,
operating costs will be higher. If a significant reduction in
the space velocity is required, or if a unit is unable to reach
the desired operating conditions, then additional units will be
required, and refiners will have to make capital investments in
order to meet the nationwide diesel fuel demand.
Hydrogen required for hydrotreating is generated from the
refinery's reformer (a producer of high octane gasoline) or a
hydrogen manufacturing plant. The hydrogen reacts with the
sulfur to form hydrogen sulfide (H2S) which can be removed in
a subsequent process. The H2S is eventually converted to
elemental sulfur and sold. Thus, increased hydrotreating
requires increases in hydrogen generation, gas treatment, and
sulfur recovery processing.
Hydrodearomatization is similar to hydrodesulfurization
(i.e., hydrotreating), differing mainly by the more severe
operating conditions present in the hydrotreater.
Hydrodearomatization requires greater temperatures and
pressures, and also requires the use of a noble metal
catalyst. Large reductions in aromatic content would force
refiners to invest in new units or significantly revamp
existing units, since most units today cannot operate at
significantly increased pressures.[9] Other potential
processing options for aromatics removal also exist.
Hydrocracking can be used to mildly reduce aromatic levels in
distillate fuels. Solvent extraction can also be used for
aromatics removal but may not be practical in many refining
situations.
In addition to these processing technologies, other
options may be available to refiners who face reduction of
either sulfur or aromatic content of on-highway diesel fuel.
One option would be product sharing. This practice currently
exists in the gasoline market where a refiner in one part of
the country will produce and sell gasoline for a refiner who
produces an equal volume of gasoline for exchange in another
part of the country. It may be possible for a similar practice
to exist for the distillate market. For example, if a refinery
has access to low sulfur crude or has excess hydrotreating
facilities, he may be able to produce low sulfur diesel fuel in
exchange for production of non-regulated fuel oil at another
refinery. Should certain refiners find it impossible to
finance additional processing equipment, this type of approach
may become attractive.
B. Energy Resource Consultants/Sobotka Study
Energy Resource Consultants, Inc. (ERC) prepared a report
under contract to EPA entitled, "Diesel Fuel Quality Effects on
Emissions, Durability, and Performance," dated September
30, 1985.[2] Sobotka and Company, Inc. (SCI) consultants were
-------
2-10
subcontracted to estimate the refining impact support of sulfur
and aromatic content reductions. Sobotka's approach is similar
to Bonner and Moore's (discussed below) in that linear
programming computer models were developed to represent the
nation's refinery capabilities. The model used for the
analysis was a version of the Department of Energy's Refinery
Evaluation Modeling System (REMS).
Two models were developed, one to represent the Eastern
U.S. (PADDs 1 through 4), and one to represent the Western U.S.
(PADD 5). Projections were made of refinery supply and demand,
product qualities, and processing capacities for the modeling
year which was 1987. No new investments, beyond the facilities
currently in place or projected to be in place, were allowed in
the modeling. After the base cases were run, two diesel fuel
control cases were run. The first considered sulfur reductions
from current levels of 0.27 weight percent to 0.05 weight
percent. The second case controlled sulfur to 0.05 weight
percent with simultaneous control of aromatic content to 17
volume percent.
Sobotka used a diesel fuel demand of 1.25 million barrels
per day (42.8 percent of middle distillate demand), assuming
complete segregation of on-highway diesel from the rest of the
distillate pool. Total middle distillate demand was 2.92
million barrels per- day. Crude oil price was assumed to be
roughly $29 per barrel. The Number 2 heating oil pool was not
allowed to serve as a sink for the sulfur removed from the
diesel fuel pool. The heating oil pool was allowed, however,
to serve as a sink for aromatics which were removed from the
diesel fuel.
The results from the ERC/SCI report are shown in Table
2-4. The results show the nationwide average cost to control
sulfur content to 0.05 weight percent to be 1.2 cents per
gallon, The average cost to control sulfur to 0.05 weight
percent and control aromatic content to 17 volume percent was
1.5 cents per gallon.
Due to the assumption of complete segregation of products,
the model was able to alter the refinery's blending practice to
a great degree to help achieve the diesel quality levels. This
blending flexibility accounts for the reductions in aromatics
that was noted when sulfur was the parameter controlled.
Aromatics decreased from 28.7 volume percent to 20.3 volume
percent when sulfur was controlled to 0.05 weight percent.
This was a result of the model directing higher aromatic blend
stocks to the No. 2 fuel oil and not to the diesel pool. In
later work (described below), Sobotka and Co. agreed that the
modelling overstated the ability of refiners to selectively
blend distillate streams, and thus the degree to which
aromatics would be reduced when sulfur is controlled would
likely be less than that presented here.
-------
2-11
Table 2-4
Results of Refining Analysis From
the Energy and Resource Consultants Study[2]
Base 0.05% Sulfur
Case 0.05% Sulfur 17% Aromatic
Costs
Change in Cost (tf/gal) 1.2 1.5
Change in Total Cost(106$/yr) 214 277
Fuel Properties
Sulfur (wt.%) 0.274 0.048 0.048
Aromatics (Vol.%) 28.7 20.3 17.0
API Gravity 34.0 37.5 37.9
Calculated Cetane Index 46.0 48.2 51.B
-------
2-12
Another important modeling assumption made by Sobotka and
Co. was that existing refinery capacity was sufficient to
reduce the sulfur content of the on-highway diesel pool. Thus,
the costs in Table 2-4 reflect only changes in refinery
operating and raw material costs. This was partly the result
of the assumptions that the industry could segregate the diesel
fuel from the other distillate products and thus only treat the
on-highway diesel product. The analysis did not determine the
minor cost of making the piping changes within the refinery nor
the potentially more significant cost of segregating the
product distribution system.
Having recognized the sensitivity of modelling results of
these assumptions, EPA asked Sobotka and Co. to prepare another
study of the issue. [3] Sulfur control to 0.05 percent with and
without aromatics control to 20 percent was analyzed for both
segregated and combined production of highway diesel fuel and
other distillates. Additional constraints were also placed on
the ability to selectively blend refinery streams, in order to
better reflect actual refinery operation. Another important
aspect of the study was that it was performed assuming that
aromatics reduction could only take place via solvent
extraction (i.e., hydrodearomatization technology was not
available).
Two assumptions concerning existing desulfurization
capacity were used in Sobotka"s updated analysis. The first
case assumed that the 1992 base case desulfurization capacity
was utilized at 69 percent. The second case was analyzed
assuming no excess desulfurization, hydrogen, or sulfur
recovery capacity was available.
Results showed that sulfur control to 0.05 weight percent
would cost between 1.4 and 2.3 cents per gallon of controlled
fuel depending on the segregation and investment assumptions.
Cost for controlling both sulfur and aromatics ranged from 1.9
to 3.5 cents per gallon of controlled fuel. Detailed results
are shown in Table 2-5.
C. Bonner and Moore's Study
EPA contracted with Bonner and Moore to use their
proprietary linear programming (LP) computer model, designated
the Refinery and Petrochemical Modeling System (RPMS), to
estimate separately the refinery costs of reducing the sulfur
and aromatic content of diesel fuel.[4] The cost of sulfur
content control to 0.05 weight percent and sulfur control with
aromatic content control to 20 vol. percent was investigated
for each of four geographic refining regions. Region l
represents the East Coast or the Petroleum Administration for
Defense District (PADD)l, Region 2 represents the mid-continent
area, that is, PADD 2, 4 and 5 (excluding California), Region 3
represents the Gulf Coast, PADD 3, and Region 4 represents
-------
2-13
Table 2-5
Updated Sobotka Costs for Diesel Fuel Quality Control[3]
Excess HPS Capacity* Ho Excess HPS Capacity
Controlled
Fuel
Volume Refining Cost** Refining Cost**
Case (1000 bbl/day) ($Million/yr) (g/qal) ($Million/yr) (i/qal)
Segregated-Low Sulfur 1071 232 1.4 340 2.1
Segregated-Low Sulfur/
Aromatics 1071 310 1.9 430 2.6
Combined-Low Sulfur 2351 654 1.8 821 2.3
Combined-Low Sulfur/
Aromatics 2351 1110 3.1 1275 3.5
* Costs were developed assuming existing HDS capacity was
utilized at 69 percent.
** Cost shown are on a per gallon of controlled fuel basis.
-------
2-14
California. Costs were developed based on an average crude oil
acquisition cost of $22 per barrel.
The national average costs for sulfur and aromatic
reductions were estimated by production weighting the PADD
specific costs obtained from the RPMS. The estimated cost for
each region was developed by the RPMS using a single "super
refinery" to represent all of the refining capabilities of that
region (i.e., the average refinery). This super refinery was
required to produce all of the products projected to be
produced by all of the individual refineries in that region in
the time frame of the study, which was 1990. Individual
refineries would be expected to experience costs both above and
below that estimated by the RPMS, but on average, the actual
costs should be close to that projected by the model. The
complexities involved with modeling individual refineries made
such an approach economically infeasible and make the use of a
single refinery for each PADD a necessary limitation of this
study.
The costs estimated by the RPMS are, by design,
incremental in nature and do not attempt to represent the full
costs of refining diesel fuel. This avoids a number of complex
issues associated with valuing capital equipment already in
place. As the desired output is the effect of sulfur and
aromatic control on refining costs, the difference in cost
between an uncontrolled and controlled scenario is fully
satisfactory for this study.
A base 1990 case was run to determine optimal process
requirements and refinery costs associated with producing the
1990 product slate, considering process capacities known to be
available in 1986 along with capacities announced to be built
and available by 1990, and, thus, not requiring capital
investment. The controlled case was run in an analogous
fashion (i.e., a fresh optimization from 1986 capacities), only
with a diesel product of lower sulfur or lower sulfur and
aromatic content.
It should be noted that the RPMS runs tend to project
sizable capital investments between 1986 and 1990 for the base
cases even though the refinery industry as a whole is expected
to invest little for refining capacity. This occurs because
the current capacity of many peripheral processes (e.g.,
cooling towers) is not known and was presumed to be zero in
1986 for modeling purposes. Thus, the required 1990 base
capacity for these processes is considered to be entirely
incremental, though in all likelihood, the vast majority of it
is currently in place. These sizable capital investments have
no direct effect on the estimated sulfur or aromatic content
control costs, nor the estimated capital investment required
for control, since the investments are present in both the base
case and the control case and do not affect the incremental
-------
2-15
control cost. It simply means that the capital investment
shown for either the base or controlled cases should not be
used to estimate the total capital investment required by the
refining industry between 1986 and 1990. However, the current
capacities of some of these peripheral processes may be in
excess of that needed in 1990. If the model is predicting that
additional capacity is needed in the control case it may be
overestimating the additional capacity needed for control. The
degree to which this may be occurring is unknown and not easily
estimated.
Each regional refining model was required to meet product
demand for 1990 using crude supplies projected to be available
in that same timeframe. Five types of distillate fuels were
modeled; naphtha jet fuel, kerosene jet fuel, heating oil,
highway diesel fuel and non-highway diesel fuels. Three types
of residual fuels were modeled; low-sulfur, high sulfur, and
bunker fuels, as well as three grades of gasoline; unleaded
premium, unleaded regular, and leaded regular. The distillate
fuel demand for each region is shown in Table 2-6.
Bonner and Moore evaluated two degrees of distillate
segregation in their study. The first was labeled 100 percent
segregation, where all heating oil was segregated from the rest
of the diesel fuel pool. The second scenario was evaluated
assuming that the current degree of segregation of heating oil
from diesel fuel, as documented by the NPRA survey, would
continue. Bonner and Moore did not, however, evaluate the cost
of completely segregating diesel used in on-highway vehicles
from all other distillates (heating oil and diesel used in
off-highway applications). This will be discussed further in
Section E below.
The base case diesel fuel used in the model had a sulfur
content of 0.25 weight percent and an average aromatic content
of 34 volume percent. Bonner and Moore did not allow the
quality of other products to degrade in order to accommodate
the control imposed on highway diesel fuel. More specifically,
heating oil was not allowed to be a sink for sulfur which would
no longer be allowed in the highway diesel fuel.
The results from the Bonner and Moore modeling are shown
in Table 2-7. Bonner and Moore performed modelling runs for
each of the four regions described above, and aggregated the
results to yield national cost estimates. After issuing a
draft report, a follow up to the original study was performed
to account for several changes. These included reducing the
estimated efficiency of the aromatics removal process used by
the model, adjusting capital investments required for
incremental process capacity, and restricting the internal
segregation capability of the refinery to a more realistic
level. Cases were run that investigated the sensitivity of the
model results to two parameters: distillate segregation, and
-------
2-16
Table 2-6
Bonner & Moore Refinery Distillate
Production Recruirements (1000 bbl/day)[4]
100 Percent Segregation*
Region
1
2
3
4
Totals
Naphtha Jet Fuel
34.0
54 . 0
61. 0
33 . 0
182 . 0
Kerosene
4.0
41. 0
76. 0
1. 0
122.0
Kerosene Jet Fuel
37 . 0
253 . 0
595 . 0
222. 0
1107 . 0
Region
1
2
3
4
Totals
Heating Oil
90.8
136.8
141. 6
4.6
373.8
Highway Diesel Fuel
155. 6
724.0
975.3
251.0
2105.9
Other Diesel Fuel
12.6
126.2
312. 1
43 .4
494.3
Subtotal: No. 2 pool
2974 . 0
Total
4385.0
NPRA Segregation*
Region
1
2
3
4
Totals
Naphtha Jet Fuel
34.0
54.0
61.0
33.0
182 . 0
Kerosene
4.0
41.0
76 . 0
1.0
122 . 0
Kerosene Jet Fuel
37.0
253 . 0
595 . 0
222 . 0
1107 . 0
Heating Oil
31.81
20.5
2.8
0.0
55 . 1
Highway Diesel Fuel
214 . 6
840.3
1114 . 1
255 . 6
2424.6
Other Diesel Fuel
12.6
126.2
312. 1
43 .4
494 .3
Subtotal: No. 2 pool
2974 . 0
Total
4385.0
Bonner and Moore designations.
-------
2-17
Table 2-7
Bonner and Moore Costs of Diesel Fuel Quality Control
(4/ga.l of Controlled Fuel) [4]
Added Capital
National ($ million)
100% Segregation with investment
Sulfur to .05 wt.%
Sulfur & Aromatic
NPRA Segregation with investment
Sulfur to .05 wt.%
Sulfur & Aromatic
Volume of fuel controlled (MBPD)
100% Segregation
NPRA Segregation
2.57 1065.9
5.20 3023.2
2.69 2023.1
5.16 4877.7
2105.9
2424 . 6
-------
2-18
capital investment. Bonner and Moore's final adjusted national
costs are also shown in Table 2-7.
The assumption that the refiners are able to segregate all
heating oil from other distillates and then only treat the
highway fuel is referred to as the 100% segregation case. The
costs under this scenario include only refinery costs and do
not include any cost that may be required to change the
distribution system downstream from the refinery to further
segregate fuels. The volume of fuel treated under the
assumption is shown in Table 2-6 and also at the bottom of
Table 2-7. The other sensitivity case that was run with regard
to segregation is referred to as NPRA segregation. This case
involved using NPRA's survey results to determine current
segregation practices. This involved including some of the
home heating oil into the on-highway diesel pool and treating a
common fungible product. Since the NPRA segregation data
reflects current segregation practices, no changes in the
distribution system would be required.
It should be noted that this segregation analysis tried to
bracket the cost by modeling the extremes of segregation.
Bonner and Moore estimated total diesel fuel demand to be 2.106
million barrels per day, a volume which included all diesel
fuels required to meet a minimum cetane specification, not
merely that fuel which is used in highway vehicles. This
volume is just slightly lower than the volume of fuel which
NPRA projected would require treatment (2.46 million barrels
per day). The actual projected 1990 demand for diesel fuel for
highway vehicles is only 1.29 million barrels per day, however
(Table 6-1). Because Bonner and Moore's estimate of on-highway
diesel fuel demand included all fuel for applications requiring
a minimum cetane specification, the control pool size did not
vary to a significant degree under the two segregation
extremes. Therefore, Bonner and Moore did not find a strong
cost sensitivity to segregation due to the assumed volume of
on-highway diesel fuel.
The amount of investment required by refiners to meet the
diesel quality specifications is also an important factor
affecting cost. Base case capacity for distillate
hydrotreating includes current industry capacity plus announced
capacity. The maximum throughput for this process depends on
the operating severity, but the publicly available data are
published without defining the related severities, thus there
is uncertainty concerning the real capacity limits. Bonner and
Moore only ran cases under the assumption that current
operations utilize hydrotreating capacity to their limits, and
that any desulfurization or dearomatization beyond current
practice would require new capacity to be built. Thus, the
refinery costs shown include capital recovery costs for new
desulfurization equipment. Should additional desulfurization
-------
2-19
equipment currently be available, costs would be somewhat
lower. This will be discussed further in Section E.
Assuming NPRA segregation, and that any additional
desulfurization beyond the base case requires capital
investment, Bonner & Moore estimated the nationwide average
cost of control sulfur content to 0.05 weight percent to be 2.7
cents per gallon of controlled fuel. For the maximum
segregation case, with investment required, Bonner and Moore
estimated the sulfur control cost to be 2.6 cents per gallon of
controlled fuel. The total cost to control aromatics to 20
volume percent and sulfur to 0.05 weight percent, was
approximately 5.2 cents per gallon for the both the NPRA
segregation and maximum segregation case. All of the costs in
Table 2-7 are expressed on a cents per gallon of controlled
fuel basis.
The added capital requirements for treating diesel is also
shown in Table 2-7. Capital requirements range from 1.1 to 2.0
billion dollars for sulfur control, and from 3.0 to 4.9 billion
dollars for sulfur and aromatics control, assuming no excess
desulfurization capacity currently exists.
D. National Petroleum Refiners Association Survey
During the final quarter of 1986, the National Petroleum
Refiners Association (NPRA) conducted a survey of its members
in order to assess the capability of U.S. refiners to produce
diesel fuels of lower sulfur and aromatic content. As noted on
the transmittal letter from the NPRA to its members, the survey
was conducted "...to ensure the complete awareness of the
associated costs and logistical implications to the industry
are made known to the government." [5] The survey was also
performed in order to gather data which could be used to check
Bonner & Moore's industry model. Responses were gathered from
139 refineries which represent 98 percent of the total U.S.
operating capacity.
The NPRA survey consisted of five parts. Part I
determined current and future practices under present product
quality specifications. Part II determined the minimal sulfur
content achievable with the use of existing facilities only.
Parts III thru V investigated the costs associated with
manufacturing diesel fuel under various diesel fuel
specifications.
The NPRA survey recognized the importance of the
segregation of distillate fuel issue and asked the refiners to
list their projected (1991) product volumes in the following
categories: base diesel fuel, common diesel/distillate No.2,
distillate No.2 (fuel oil), and other diesel. Base diesel fuel
was defined as that fuel intended specifically for highway,
off-highway construction, farm, industrial, and commercial
-------
2-20
uses. The common diesel/distillate No. 2 pool refers to
finished stocks which may be marketed as either diesel fuel or
fuel oil. Other diesel fuel includes railroad, marine, and
military fuel. The refiners projected product volumes for each
of these categories for 1991 are shown in Table 2-8. The
average sulfur content is also shown for each product pool.
Table 2-8 also shows findings by NPRA concerning how much
additional segregation is possible by the refiners. The NPRA
asked refiners how much fuel could be segregated, above current
practices, "...to the limit of perceived capability to maintain
segregation through distribution and marketing systems."
Although 30 percent of the refiners indicated they would
initiate segregation of final products, the actual volumes of
segregated products changed very little as indicated in Table
2-8.
It should be noted that ideally the volumes in each column
of Table 2-8 should be equal but in fact they are not due to
inaccuracies involved in conducting a survey of many parties.
It is also interesting to note the trends in Table 2-8. One
would expect the volumes of highway diesel and distillate No. 2
to increase as the specifications become more stringent and the
industry attempted to break up the common pool into two
distinct pools. The table shows the highway diesel pool
increases, but the distillate No. 2 pool actually decreases.
Also, in the base case there is a small amount of "other"
diesel fuel which is high in sulfur content. In the control
cases, this volume was included with other products. The table
shows 441,000 barrels per day of segregated highway diesel and
2,022,000 barrels per day of common diesel/distillate fuel
would have to be controlled in order to lower sulfur content of
the on-highway fuel. Examination of the detailed results of
the NPRA survey, however, revealed that 59,000 barrels of fuel
in the highway diesel pool was fuel which was designated for
either railroads, military, or vessel bunkering use. In their
comments on the ERC report, API stated that fuel designated for
these uses was currently segregated from on-highway diesel
pools and should not be included in the control pool
volume.[3] Subtracting this fuel from the total yields a total
control pool volume of 2,404,000 barrels per day. This volume
represents the refining industry's estimate of their capability
to separate the distillate products, as determined by the NPRA
survey. However, as shown in Table 6-1, on-highway diesel fuel
consumption is expected to be only 1.30 million barrels per day
in 1991.
NPRA's estimate of the nationwide average increase in
refining costs to meet lower on-highway diesel fuel
specifications are shown in Table 2-9. The total cost to
reduce sulfur to 0.05 weight percent is 3.11 cents per gallon
(on a dollar per gallon of controlled fuel basis). The Bonner
and Moore result which corresponds to this cost is the NPRA
segregation case, which was 2.69 cents per gallon. The primary
-------
2-21
Table 2-8
Distillate Product Volumes From NPRA Survey (10Q0 bbl/day)[5l
Base Diesel
Common Distillate
Distillate No.2
Other Diesel
Total
Base
363(.23)*
2,053(.28)
240(.29)
59(.55)
2,715
Sulfur Control
to 0.05 wt. %
441<.05)*
2,022(.05)
219(.12)
N.R.***
2,682
Sulfur Control
to 0.05 wt. % &
Aromatics to 20 Vol %
406 (.05,19.6)**
2,029 ( .04,20.0)
221 (.12,27.0)
N.R.
2,656
* Sulfur content, weight percent, shown in parentheses.
** Sulfur content (wt. percent), aromatics content (volume
percent), shown in parentheses.
*** Not reported, volume included with other products.
-------
2-22
Table 2-9
NPRA Survey Results of Nationwide Diesel Fuel Quality Control Costsf51
Sulfur Control
.05 wt. %
,05 wt Aromatics Control
to 20 vol. %
Additional Manufacturing Costs (£/gal)*
Operating Expense
Capital Cost
Total Added Cost
Total Capital Investment ($ million)
.15 wt. %
0.48
1.33
1.81
2028
0.92
.920
3.11
3313
1.90
4.45
6.35
6651
* On a per gallon of controlled fuel basis.
-------
2-23
difference between the two studies is in the capital cost.
NPRA projects capital expenditures of $3.3 billion, versus $1.0
to $2.0 billion as calculated by Bonner and Moore. The NPRA
survey did find that a small amount of distillate fuel could be
produced at 0.05 weight percent sulfur without further capital
investment. If existing facilities were used to their fullest
potential, it was estimated that 316,000 barrels per day of
0.05 weight percent sulfur diesel fuel could be produced by
increasing operating severity only.
NPRA's estimated costs of controlling on-highway diesel
fuels sulfur content to 0.05 weight percent and aromatic
content to 20 volume percent is also shown in Table 2-9. The
total cost of control is estimated at 6.35 cents per gallon, 20
percent higher than Bonner Moore's estimate of 5.2 cents per
gallon. Total capital investment was estimated to be $6.65
billion, versus $3.0 to $4.9 billion as estimated by Bonner and
Moore.
Twenty seven percent of the refiners, representing 17
percent of distillate production, stated they could not
implement the low-sulfur diesel specifications within 3 years,
due to a lack of capital or to environmental permitting
constraints. It should be noted the NPRA survey did not
address an implementation schedule longer than three years.
Despite the fact that these refiners could not meet the
specifications, their cost estimates are included in all of the
cost data presented above.
The survey also instructed the responding refiners to keep
their product slate constant from the base case to the control
case. This implies that a refiner who has facilities to
hydrotreat 75 percent of his current product must make an
investment to treat the remaining 25 percent, while a refiner
with excess hydrotreating facilities could not produce any
additional low sulfur fuel. The production volume lost by
those refiners who cannot produce low sulfur diesel fuel could
presumedly be made up by the refiners who do have excess
hydrotreating facilities, thus decreasing the total capital
investment required by the industry.
E. "Best Estimate" Refinery Cost
This section will outline how a "best estimate" of the
refinery cost impact was determined for sulfur and aromatic
control of diesel fuel. While the analysis used information
which was provided by each of the studies presented above, the
control cost estimates were derived primarily from the Bonner
and Moore study. The Bonner and Moore study is an objective
analysis which was conducted after the ERC/SCI study was
complete, thus being able to address and incorporate relevant
comments on the ERC/SCI study. It is worth noting that the
updated SCI study's sulfur control costs (Table 2-5) are very
close to those of the Bonner and Moore study, once corrections
are made by EPA to adjust capital cost assumptions.
-------
2-24
The costs presented in the NPRA survey were not used
directly, since the respondents to the NPRA survey were the
parties which would be regulated and the survey was conducted
to provide information specifically to regulatory authorities.
Although the information is valuable, and some of it is
incorporated below, the cost estimates can not be used directly
since it is difficult to confirm the accuracy of the data
provided by each respondent.
The "best estimate" control costs were determined as
follows. Per gallon control costs, as determined by Bonner and
Moore, were first adjusted to reflect a lower after-tax rate of
return and were then applied to the volume of fuel controlled
in 1991 under two extreme distillate segregation scenarios. In
the first scenario, 100 percent segregation, it was assumed
that only diesel fuel used by on-highway diesel vehicles would
require treatment. In the second scenario, NPRA segregation,
it was assumed that current segregation practice would be
maintained. The amount of existing excess desulfurization
capacity was also accounted for in the cost estimates, as will
be described below.
An upper and lower bound to the per gallon control cost
was determined for both the sulfur and subsequent aromatics
control scenarios. The range in cost is dependent largely on
the segregation capability of the industry, as this affects the
volume of fuel to which the control costs are to be applied.
If on-highway diesel fuel specifications are mandated, the
economic incentive to segregate the products will be strong.
The minimum amount of fuel that would need to be treated is the
on-highway diesel pool itself, which is estimated to be 1.30
million barrels per day in 1991. This volume of fuel was used
in estimating a lower bound for the refining cost. The NPRA
control volume estimate (2,404,000 barrels per day) was used to
determine the upper bound.
The costs developed by Bonner and Moore were modified as
follows. First, the Bonner and Moore costs were computed using
a capital recovery factor of 0.226 (Table 2-7 in their
report). In other words, supporting one dollar of investment
would cost 0.226 dollars per year. When allowance for
maintenance, local taxes, insurance, and overhead is included,
this capital recovery factor becomes 0.286. These capital
recovery factors are based on a 15 percent after-tax cost of
capital (in real, not nominal, dollars). However, a recent
report prepared by Jack Faucett Associates under contract with
EPA indicates that a 10 percent after-tax cost of capital is
more appropriate for the refining industry. [10] When this
figure is used the capital recovery factor, as calculated by
Bonner and Moore's methodology, becomes 0.171 (0.231 including
maintenance, local taxes, etc.).
-------
2-25
Using these adjustments, sulfur control costs for the
maximum segregation case were determined to be 0.857 dollars
per barrel of controlled fuel (2.04 cents per gallon) assuming
HDS investment is required, and 0.482 dollars per barrel (1.15
cents per gallon) for the "no investment" case.
An additional adjustment was made to Bonner and Moore's
100 percent segregation case results. In the Bonner and Moore
report, control costs for desulfurization were determined by
comparing investment and operating costs for two cases:
maximum segregation with 0.25 percent sulfur fuel (baseline)and
maximum segregation with 0.05 percent sulfur fuel. However, to
accurately estimate the cost of processing changes as actually
perceived by the industry, investment and operating costs for
the "maximum segregation with 0.05 percent sulfur" case should
be compared with a baseline case representing current industry
practices (i.e., the NPRA segregation case). Differential
processing investment and operating costs were recalculated
accordingly. Bonner and Moore's fuel control costs for the 100
percent segregation and NPRA segregation cases were adjusted,
and are shown in Table 2-10.
These costs were generated under the assumption that no
excess desulfurization equipment was available. However, the
NPRA survey found that if existing facilities were used to
their fullest potential, 316,000 barrels per day of 0.05 weight
percent sulfur diesel fuel could be produced. Thus, 316,000
barrels per day of fuel could be treated at a lower cost. This
cost was determined by subtracting the per gallon capital
recovery cost associated with new desulfurization equipment
from the per gallon "HDS Investment Required" costs shown in
Table 2-10. These "No HDS Investment Required" costs are also
shown in Table 2-10.
As stated previously, the volume of fuel which will be
controlled ranges from 1.30 to 2.404 million barrels per day in
1991. A total of 316,000 barrels per day can be controlled at
the "no HDS investment" cost shown in Table 2-10. The
remainder of the pool will be controlled at the "HDS Investment
Required Cost." The average national control costs for control
of sulfur and subsequent control of aromatics were thus
calculated and are shown in Table 2-11. Sulfur control costs
range from 1.8 to 2.3 cents per gallon of controlled fuel ($360
to $830 million per year), while subsequent control of
aromatics costs from 2.1 to 2.4 cents per gallon of controlled
fuel ($470 to $770 million per year).
Ill. Effect of Hydrodesulfurization on Fuel Aromatics
The refinery processing required to remove the sulfur from
diesel fuels involves introducing hydrogen to the fuel and
passing it at elevated temperature and pressures over a
catalyst. This process is similar to that used to saturate
-------
2-26
Table 2-10
Adjusted Bonner and Moore Costs of Diesel Fuel Control
(Cents per Gallon of Controlled Fuel)
Cost (^/gallon)
100% Segregation (w/HDS investment required)
Sulfur to .05 wt.% 2.06
0.05% Sulfur & Aromatic to 20 Vol % 4.34
NPRA Segregation (w/HDS investments required)
Sulfur to .05 wt.% 2.39
0.05% Sulfur & Aromatic to 20 Vol % 4.44
100% Segregation (no HPS investment reguired
Sulfur to .05 wt.% 1.04
0.05% Sulfur & Aromatic to 20 Vol % 3.58
NPRA Segregation(no HPS investment reguired)
Sulfur to .05 wt.% 1.32
0.05% Sulfur & Aromatic to 20 Vol % 3.69
-------
2-27
Table 2-11
"Best Estimate" Refinery Cost Cor Controlling On-Hiqhway Diesel Fuel
Sulfur Control to 0.05 Weight Percent
Lower Bound (100 Percent Segregation) Upper Bound (NPRA Segregation)
Volume
Control Cost
Control Cost
Volume
Control Cost
Control Cost
(1000 bbl/day)
(*/gal)
($ million/yr)
(1000 bbl/day)
(g/gal)
($ million/yr)
316
1.04
50.4
316
1.32
63.9
962
2.06
310.1
2088
2.39
765.0
1298
1.81
360.5
2404
2.25
829.0
Sulfur
Control to 0.05
Weight Percent With
Aromatic Control to
20 Volume Percent
Lower Bound
Upper Bound
Volume
Control Cost
Control Cost
Volume
Control Cost
Control Cost
(1000 bbl/day)
(tf/gal)
($ Million/yr)
(1000 bbl/day)
U/gal)
($ Million/yr)
316
3.58
173.4
316
3.69
178.8
9B2
4.34
653.3
2088
4.44
1,421.2
1298
4.16
826.7
2404
4 .34
1,600.0
Incremental Aromatics
Control:
2.35
466.2
2.09
771.0
-------
2-28
aromatic compounds in the fuel, although larger-scale aromatics
saturation requires more severe hydrotreating conditions.
However, although desulfurization units serve primarily to
remove sulfur from the fuel, some mild ring saturation does
take place, and consequently, the amount of aromatic carbon is
reduced.
Typically desulfurization does not result in the
saturation of monocyclic aromatic structures. Dicyclic and
polycyclic aromatic compounds present in diesel fuel, however,
are often partially saturated. This is because the second and
third aromatic rings are less stable than the first ring, and
are therefore saturated more easily during hydrodesulfurization.
The result of this partial saturation is that the amount of
aromatic carbon in the fuel is reduced, while aromatics levels
measured by Fluorescent Indicator Adsorption (FIA) analysis
(which counts mono-, di-, and tricyclic compounds equally)
remain the same. Fuel analysis for aromatic content by mass
spectroscopy, however, shows the shift from polycyclic to
monocyclic structures.
Hydrodesulfurizing experiments were carried out on various
fuels by Jack R. Yoes and Mehmet Y. Asim of Akzo Chemie America
with the purpose of determining hydrotreater performance and
aromatic saturation.[11] Two cracked feeds and a high sulfur
virgin stock were hydrotreated at hydrogen partial pressures of
400, 800, and 1200 psig, at temperatures designed to produce of
0.0 5 wt. percent sulfur product. Both cobalt-molybdenuir
(KF-742-1.3Q) and nickel molybdenum (KF-843-1.3Q) catalyst,
were evaluated to determine differences in sulfur removal ar. j
aromatics saturation. Although conditions were severe enouc;
to produce reductions in aromatics as measured by FIA as wel"
as sulfur content, what is interesting to note is the shift .i
aromatic distribution upon hydrotreating. For the three fuels,
using the nickel-molybdenum catalyst, the fraction of aromatics
falling in the monocyclic and dicyclic categories shifted m
average from about 70 and 18, respectively, to about 77 and 1J,
respectively upon hydrotreating. In other words, approximately
30 to 40 percent of the dicyclic aromatics were partially
saturated.
Another evidence of this shift took place in the
hydrotreating of fuels for the CRC VE-1 project
Hydrotreating was used to produce a low sulfur fuel (0,06
weight percent) from a high sulfur feed (0.32 weight percent).
Although FIA aromatics levels remained constant; at
approximately 40 volume percent, the distribution of aromatic
species changed drastically. The mass percent of mono- , di-,
and tricyclic aromatic carbon was 8.4, 14.2, an 1 1.6,
respectively, in the feed, and 13.7, 7.9, and 0.8,
respectively, in the hydrotreated product. This represents a
reduction of roughly 50 percent in dicyclic and tricyclic
compounds.
-------
2-29
The increase in monocyclic structures indicates that
nearly all those di- and tricyclic molecules became monocyclic
aromatic structures. It also appears that little or no
monocyclic aromatic saturation took place. Treatment of
another fuel used in the CRC test program to reduce both FIA
aromatics (to 10 percent) and sulfur levels resulted in
eliminating nearly all dicyclic (92 percent) and tricyclic (100
percent) structures.
Another study of the effect of hydrotreating on diesel
fuel aromatics was funded by the National Research Council of
Canada.[13] A coal-derived middle distillate fuel was
subjected to three levels of severity of hydrogenation. A
reduction in ftal aromatics of approximately 30 volume percent
showed a corresponding decrease of 70 and 85 percent in
dicyclic and tricyclic aromatic structures. More severe
treatment showed reductions in excess of 80 and 95 percent of
dicyclic and tricyclic structures, respectively.
What the first two examples suggest is that even when
aromatics levels measured by FIA stay constant,
hydrodesulfurization may result in a reduction in aromatic
carbon, due to a shift the distribution of aromatic compounds
greatly in the direction of monocyclic compounds. The more
severe hydrodearomatization process may result in the partial
saturation of nearly all di- and tricyclic aromatic species.
As will be discussed in Chapter 4, emissions may be dependant
on the type of aromatics present. Further experimental
quantification of the shift in aromatic structure of fuels with
hydrodesulfurization may be warranted as emission models become
more sophisticated and are able to utilize this information.
In this report, it will be estimated that: 1) a reduction
in fuel sulfur levels to 0.05 weight percent will result in the
partial saturation of approximately 50 percent of the di- and
tricyclic aromatic species in the fuel to monocyclic aromatics
and 2) processing to reduce fuel sulfur to 0.05 weight percent
and aromatics to 20 volume percent will result in the partial
saturation of 80 and 90 percent of the di- and tricyclic
aromatics respectively.
IV. Small Refinery Impact
Of particular interest in this study is the impact on
small refiners of a regulation of diesel fuel quality. To
investigate this, EPA commissioned Sobotka and Company Inc. to
study the issue.[14]
In Sobotka's report, small refiners were categorized
according to the type and configuration of processing
technology employed, and then simulated by aggregate refining
models. A comparison of costs for fuel control for small
versus large refiners was then made.
-------
2-30
Three types of small refineries were identified for
purposes of the study; cracking, hydroskimming, and topping.
Cracking refers to refineries employing both catalytic cracking
and reforming technology. Hydroskimming refers to those
refineries with reforming but without catalytic cracking
facilities. Refineries with neither catalytic cracking or
reforming facilities were defined as topping facilities.
Sobotka found that 74 refineries operating at the
beginning of 1986 were classified as small refineries*
representing about 8.4 percent of the total U.S. crude oil
distillation capacity. Modeling runs for fuel sulfur control
and fuel sulfur and aromatics control were performed, assuming
both segregated and combined production of diesel fuel and
other distillates.
The study concluded that small topping refineries would
have little ability to alter the quality of their product.
Instead of making the necessary investment to produce a low
sulfur fuel, these facilities would most likely discontinue to
produce highway diesel fuel.
For hydroskimming refineries, the study concluded that
costs for sulfur control would range from 1.5 to 2.2 cents per
gallon of controlled diesel, depending on what is assumed
regarding current excess desulfurization capacity and the
ability to segregate on-highway diesel fuel. Cost for
producing fuel with 0.05 wt percent sulfur and 20 percent
aromatics ranged from 2.7 to 4.5 cents per gallon of controlled
fuel.
Sobotka estimated that the cost for sulfur control at
cracking refineries would range from 1.3 to 2.5 cents per
gallon. Costs for control of sulfur and aromatics were
estimated to range from 3.1 to 5.2 cents per gallon.
Sobotka compared these small refinery cost estimates to
their cost estimates for the aggregate U.S., which ranged from
1.4 to 2.3 cents per gallon of controlled fuel for sulfur
control, and from 1.9 to 3.6 cents per gallon for sulfur and
aromatics control. As can be seen, sulfur control costs for
small refiners are similar to those of the entire U.S. refining
industry. However, as the Sobotka report states, reducing
sulfur would require investments of $49 to $98 million, a level
which may be difficult for small refiners to finance. The cost
Crude oil feedstock of 50 MBPD or less, owned or
controlled by a refiner with total capacity less than
137.5 MBPD.
-------
2-31
of subsequently reducing aromatics is substantially higher for
small refiners than for large refiners.
V. Leadtime Requirements
One important aspect of the leadtime requirement question
is the close proximity of this regulation and the regulation of
gasoline volatility. Investment requirements on the order of
$1 billion for volatility regulations in conjunction with the
requirements placed on refiners due to diesel fuel regulations
may make it more difficult to make both changes simultaneously
than separately. Consideration of this issue was given by
Sobotka and Company in an analysis of leadtime requirements for
diesel fuel control.[15]
SCI concluded that, if on-highway and off-highway diesel
fuel could be segregated, capital requirements would range from
$1 to $2 billion, and compliance could be achievable by late
1993, given promulgation of a final rule by mid-1990. If
segregation of diesel fuels proved infeasible, and investments
of $3.3 billion, as estimated by NPRA, were required, SCI
concluded that compliance might not be feasible until 1995.
SCI could not estimate the leadtime requirements for aromatics
control, but given the large amount of capital investment which
would be required, compliance would likely not be feasible in
time for the 1994 diesel engine particulate regulations.
In a joint industry proposal submitted to EPA, members of
the oil refining and engine manufacturing industries proposed
that diesel fuel containing no greater than 0.05 percent sulfur
by weight and meeting a minimum cetane index specification of
40 could be supplied commercially by October 1, 1993.[16] The
proposal indicated that segregation of on-highway and
off-highway diesel would likely occur to some extent. Thus,
both those organizations responsible for the joint industry
proposal and Sobotka agree that with an increase in fuel
segregation, compliance with regulations requiring sulfur
control to 0.05 percent could take place by late 1993.
Even if domestic refiners were unable to install the
necessary processing equipment by the effective date of low
sulfur regulations, it is likely that foreign refiners could
supply some amount of low sulfur to the United States to
supplement any shortage. Indeed, as indicated in a preliminary
analysis performed by Sobotka and Company, Inc., foreign
refiners may even be able to produce low sulfur fuel at a
competitive advantage over their U.S. counterparts.[17]
Sobotka predicted a cost advantage of 0.6 to 1.4 cents per
gallon for sulfur control for foreign refiners. This is due to
the fact that only a small portion of their distillate pool
would be subject to sulfur regulations. The flexibility of
foreign refiners in minimizing production costs would therefore
be greater. This analysis did not, however, investigate the
-------
2-32
cost to foreign refiners of segregating the production of a
relatively small portion of their distillate pool, nor the
costs associated with transporting low sulfur distillate to the
U.S. Although these factors could reduce the competitive
advantage of foreign refiners, Sobotka suggested that current
imports of distillate fuel of 200,000 barrels per day could
double or even triple if diesel fuel sulfur control regulations
are implemented.
-------
2-33
References (Chapter 2)
1. "Diesel Fuel Quality—Refining Constrictions and the
Environment," George H. Unzelman, HyOx Inc., presented at the
1987 NPRA Annual Meeting, San Antonia, Texas, March 29-31, 1987.
2. "Diesel Fuel Quality Effects on Emissions,
Durability, and Performance," Craig Miller and Christopher
Weaver, Energy and Resource Consultants, Inc., and William
Johnson and Terry Higgins, Sobotka and Co., Inc., EPA contract
68-01-6543 report, September 30, 1985.
3. "Cost and Feasibility of Lowering Diesel Fuel Sulfur
and Aromatic Content," Sobotka & Co., Inc., draft final report
for EPA contract #68-01-7288, November 2, 1987.
4. "A Study on Restriction of Sulfur and Aromatics
Content of Highway Diesel Fuel - An Estimate of Economic Impact
on the U.S. Refining Industry," Franklin P. Frederick, Bonner
and Moore Management Science, final report for EPA Contract
68-03-3353, June 9, 1988.
5. "U.S. Refining Industry Capability to Manufacture
Ultra Low Sulfur Diesel Fuels," National Petroleum Refiners
Association, Washington D.C., 1986.
6. "Petroleum Marketing Monthly," DOE/EIA-0380 (88/06),
June, 1988.
7. "Comments from the American Petroleum Institute in
Response to EPA's Federal Register Request (51 F.R. 23437, June
27, 1986)," October 27, 1986, (Available in Public Docket No.
A-86-03).
8. "Effects of Diesel Fuel Standards on Transportation
and Bulk Terminal Storage of Distillate Fuels," Memorandum from
Doug Koplow and Eric Hillenbrand, Sobotka & Co., Inc., to
willard Smith, EPA, December 17, 1987.
9. "Higher Severity Diesel Hydrotreating," D.C.
McCulloch, M.D. Edgar, and J.T. Pistourius, American Cyanamid
Company, presented at the 1987 NPRA Annual Meeting, San
Antonio, Texas, March 29-31, 1987.
-------
2-34
10. "Butane Suppliers: An Industry Profile and Analysis
of the Impacts of Decreased Market Prices Caused by Gasoline
Volatility Control," Jack Faucett Associates, Final report
prepared for Work Assignment #16, EPA Contract #68-03-3244,
February 1988.
11. "Confronting New Challenges in Distillate
Hydrotreating," Jack R. Yoes, Mehmet Y. Asim, Akzo Chemie
America, presented at 1987 NPRA Annual Meeting, San Antonio,
Texas, March 29-31, 1987.
12. "Investigation of the Effects of Fuel Composition
and Injection and Combustion System Type on Heavy-Duty Diesel
Exhaust Emissions", Terry L. Ullman, Southwest Research
Institute, Coordinating Research Council Contract CAPE-32-80,
Project VE-1, March 1989.
13. "Production and Analysis of EDS Coal-Derived Middle
Distillate Test Fuels from Hydrogenation at Three Levels of
Severity," J. Erwin and N. Sefer, Southwest Research Institute,
and B. Glavincevski, National Research Council of Canada, SAE
Paper no. 872038, 1987.
14. "Cost and Feasibility of Lowering Diesel Fuel Sulfur
and Aromatic Content—Impact on Small Refiners," Sobotka & Co.,
Inc., draft report for EPA contract #68-01-7288.
15. "Evaluation of the Leadtime to Comply with Gasoline
Volatility and Diesel Fuel Regulations," Sobotka & Co., Inc.
May 5, 1988.
16. "Recommended Federal On-Highway Diesel Fuel
Specifications to Assist Engine Manufacturers in Meeting the
1991 and 1994 Particulate Standards," submitted by API, NPRA,
EMA, and NCFC, July 19, 1988.
17. "Assessment of the Relative Impact of the Proposed
Diesel Fuel Regulations on Domestic and Foreign Refiners,"
Sobotka and Company, Inc., draft, April 14, 1988.
-------
Chapter 3
Engine Control Technology
and Cost Prior to Fuel Control
The purpose of this chapter is to develop and examine a
scenario which represents the response of the diesel industry
to emissions standards set for upcoming model years assuming no
change in diesel fuel quality. An analysis of the likely
developments in engine-out emissions will be presented, as well
as a determination of the emission target levels expected to be
used by industry. The feasibility, application, and cost of
the various types of exhaust aftertreatment devices which may
be necessary for compliance with standards for diesel engine
particulate emissions will also be examined. A discussion of
the influence of fuel sulfur on each type of aftertreatment
device, an estimation of the costs and the emission reduction
efficiencies of each technology, and a projection of the types
of aftertreatment devices which will be used assuming no fuel
control, are presented below.
I. Engine-Out Particulate Emissions - Current Fuel
A. Heavy-Duty Diesel Engines
This section will present engine-out particulate emissions
for heavy-duty diesel engines (HDDE) using baseline fuel. It
is convenient to break this analysis into four timeframes
(pre-1988, 1988-1990, 1991-1993, 1994+) due to progressively
more stringent particulate standards effective in 1988, 1991
and 1994. While it is sufficient for air quality modelling
purposes to analyze average emission levels of all engines
within a HDDE subclass, beginning in 1991 it is necessary to
address the distribution of emissions about the average in
order to model aftertreatment compliance strategies under the
emissions averaging program. It is also necessary to analyze
the exhaust particulate composition for all timeframes for two
reasons: 1) to quantify the environmental effect of fuel
modifications, since each type of exhaust particulate has a
different environmental effect, and 2) beginning in 1991 to
correctly apply the various aftertreatment devices since their
control efficiency can differ dramatically between the various
types of particulate. Thus, this discussion of engine-out
emissions will be structured as follow: First, an emission
target level will be established for each standard. Second,
engine-out particulate compositions will then be estimated for
each of the four timeframes. Third and fourth, average
emission levels of pre-1988 and 1988-1990 engines will be
presented, respectively. Fifth and sixth, the rationale and
derivation of the 1991-1993 and 1994+ projected emission
scenarios will be presented. Finally, a discussion of how the
emission distributions for these projected scenarios were
derived will be presented.
-------
3-2
1. Selection of an Emission Target Level for Heavy-Duty
Diesel Engines
In order to meet the particulate standard in 1991-93 and
1994 and beyond, manufacturers need to select a target level
below that of the standard to allow for emission deterioration,
variability in testing, and engine to engine production
variability. A methodology for selecting design target levels
was developed in the Draft Regulatory Impact Analysis for the
NOx/Particulate Rulemaking. [1] In that document, the equation
for determining the zero-mile target level (TL) was developed:
ZMTL = MPL-DF
AQL
where
MPL = Maximum emission level passing the standard,
DF = the particulate deterioration factor, and
AQL = Selective Enforcement Audit Adjustment Factor.
In the aforementioned document, the AQL for particulate
was estimated to be 1.10-1.15. For this analysis the higher
level of 1.15 was used, since the lower emission levels may
produce some upward pressure on measurement variability (as a
percentage of the standard). Using the deterioration factors
developed below and an AQL of 1.15, particulate emission target
levels were calculated for the 1991 and 1994 particulate
standards. These target levels were then used in estimating
the cost and technological requirements for compliance with the
particulate standards.
2. Particulate Composition
The particulate emitted by diesel engines is made up of a
variety of components whose relative amounts vary depending on
a number of factors such as engine size, engine type (direct or
indirect injection) and fuel and oil consumption level. When
analyzing the effects of fuel modifications and exhaust
aftertreatment on particulate levels it is helpful to know the
relative amounts of each component in the particulate because
fuel modification and aftertreatment devices tend to impact
some portions of the particulate much more than others.
For the purposes of this study the standard breakdown of
the particulate into soluble organic fraction (SOF), sulfate
and bound water, and residual carbonaceous portion (RCP) will
be used. The SOF consists primarily of hydrocarbons coming
from unburned or partially burned fuel and lubricating oil. A
small amount (about 2 percent, as discussed in Chapter 4) of
the fuel sulfur is converted to sulfates which attract water.
Finally, the RCP consists primarily of solid carbon but also
includes trace amounts of ash and other components from the
fuel and lubricating oil.
-------
3-3
The standard method of determining the particulate
composition is to determine the SOF through a solvent
extraction using methylene chloride as the solvent. The
particulate filter can be weighed before and after the
extraction or the extract can be dried and weighed to determine
the weight of the SOF. Following the soxhlet extraction the
particulate filter, which now contains only the RCP and sulfate
and bound water components, is washed with water and isopropyl
alcohol to remove the sulfate and bound water leaving the RCP
on the filter.
With this technique the three components of the
particulate can be measured fairly accurately. However, this
technique is time-consuming and costly. As a result, more
testing is being done using the vacuum oven sublimation
technique, which consists of baking the particulate filter in a
vacuum at 200-225°C for 16-18 hours. This evaporates all of
the SOF. However, the literature suggests that roughly 25-40
percent of the sulfate portion is also removed during vacuum
sublimation.[2,3] Also, since vacuum sublimation is generally
used to get a quick estimate of SOF, an additional analysis of
the sulfate is generally not performed and the results are
reported only in terms of volatile and nonvolatile fractions.
Since it is desirable to utilize as much data as possible in
analyzing particulate composition, data obtained using the
vacuum sublimation technique will be incorporated into this
analysis after being adjusted to account for the inclusion of a
portion of the sulfate emissions. To do this, it will be
assumed here that 30 percent of the sulfate is baked off during
the vacuum sublimation process based on the range of roughly
25-40 percent reported in the literature.
As was mentioned previously, the level of sulfate and
bound water can easily be calculated. It is primarily a
function of fuel sulfur level, brake-specific fuel consumption
(BSFC), sulfur to sulfate conversion, and relative humidity.
The equation used to calculate the sulfate and bound water
level for a transient test is as follows:
S04 + H20 = 0.316 x (BSFC x S x CONV) (eq.3-1)
Where:
SO4 + H2O
BSFC
S
CONV
Mass of sulfur plus bound water
(g/BHP-hr)
Brake specific fuel consumption
(lb/BHP-hr)
Weight percent sulfur in fuel
Percent conversion of sulfur to sulfate
in the engine
When any of the inputs to this equation are known for a
specific piece of data they will be used to calculate the
-------
3-4
sulfate level for that engine. In the absence of
engine-specific data, average values are assumed as follows:
1) a fuel sulfur level of 0.25 weight percent, which is
representative of typical diesel test fuels, 2) the
class-average BSFC value found in the MOBILE4 conversion factor
analysis (0.54 lb/BHP-hr for light HDDEs, 0.43 lb/BHP-hr for
medium HDDEs and 0.39 lb/BHP-hr for heavy HDDEs), and 3) a
sulfur to sulfate conversion of 2.0 percent as discussed in
detail in the next chapter.
Figure 3-1 shows how the adjusted vacuum sublimation data
compares to data obtained through soxhlet extraction. This
data set is made up of both current technology engines as well
as future development engines. At low particulate levels it is
difficult to make a direct comparison due to a lack of soxhlet
extraction data on low particulate engines. However, in the
mid-level particulate range the results from the two techniques
seem to relate fairly well. Therefore, this technique will be
used on vacuum sublimation data to be used in this analysis.
Data on a number of current production and development
engines were collected in order to analyze the particulate
composition of both current and future engines. Because much
of the data submitted by the manufacturers for this analysis
are confidential, it will only be presented graphically without
identification with respect to specific manufacturer or engine
family. These data are shown in Figure 3-2. It should be
noted that ten of the twenty-one heavy-duty data points were
obtained through vacuum sublimation while all ten of the
medium-heavy data points come from soxhlet extraction and only
one of the four light-heavy data points was obtained through
vacuum sublimation. It should also be noted that some of the
data presented in this table were obtained using low sulfur
fuel and, as will be discussed in the next chapter, the overall
particulate levels would be higher on baseline fuel due to an
increased sulfate level. However, the SOF and RCP levels would
remain unchanged.
As was previously mentioned, sulfate and bound water
levels are primarily a function of fuel consumption, sulfur to
sulfate conversion and fuel sulfur level and do not tend to
vary with total particulate level when these other parameters
are held constant. Therefore, it is convenient to treat the
subject of sulfates separately from that of SOF and RCP. The
data from Figure 3-2 are replotted in Figure 3-3 showing SOF
vs. total non-sulfate particulate. It may also be helpful to
examine SOF as a percentage of non-sulfate particulate as shown
in Figure 3-4. Examining the medium and heavy HDDE classes in
Figure 3-3 it appears as though the average SOF level decreases
slightly as the non-sulfate particulate level decreases. In
Figure 3-4 it appears that the average percent SOF increases
slightly at lower particulate levels. In both instances the
-------
FIGURE 3-1
HHDDE SOF LEVEL vs. PARTICULATE LEVEL
o_
x
m
D)
LLI
>
LU
O
CO
~
VACUUM OVEN SUB.
PARTICULATE LEVEL (g/BHP-hr)
+ SOXHLET EXTRACTION
-------
FIGURE 3-2
SOF LEVEL vs. PARTICULATE LEVEL
0.3 0.4
PARTICULATE (g/BHP-hr)
~ HHDDE
+ MHDDE
o LHDDE
-------
0.24
FIGURE 3-3
SOF LEVEL vs. NON-SULFATE PARTICULATE
0.22 -
0.2 -
0.18 -
0.16 -
0.14 -
0.12 -
0.1 -| £)
cP +D D°
0.08 - ~
0.06 -I D
~
0.04 -
0.02 -
0.15 0.25
+
+
~
~ +
~ ~
NON-SULFATE PARTICULATE (g/BHP-hr)
~ HHDDE + MHDDE o LHDDE
+
D ~ n ~
0.&5 '
-------
FIGURE 3-4
NON-S04 PM:SOF PERCENT VS. LEVEL
0.15 0.25 0.35 0.45 0.55
NON-SULFATE PARTICULATE (g/BHP-hr)
~ HHDDE + MHDDE o LHDDE
-------
3-9
change is very slight and a case could be made to represent SOF
as a constant level or a constant percentage over a range of
particulate levels. However, SOF levels must decrease in order
for engines to reach the very low particulate levels projected
for the 1994 timeframe. Figure 3-3 demonstrates that some
engines are already capable of very low SOF levels and
therefore shows that this assumption is not unreasonable.
Thus, a constant percentage SOF will be used for all projected
particulate levels. Averaging the data in Figure 3-4 shows the
SOF percentage of non-sulfate particulate to be 51 percent for
light HDDEs, 44 percent for medium HDDE's, and 24 percent for
heavy HDDEs. Urban bus engines, generally DDA's, generally
fall into the category of medium HDDEs as defined in this
study. Thus, 44 percent will also be used for the urban bus
class.
In summary, the particulate composition for all cases in
this study will be determined as follows: first, the level of
sulfate and bound water will be calculated using appropriate
input information for the scenario bring considered. Second,
this calculated sulfate level will be subtracted from the total
particulate level to obtain the non-sulfate particulate level.
The appropriate SOF percentage for that weight class will then
be applied to the non-sulfate particulate to determine the
level of SOF with the remainder being RCP. With respect to
deterioration, the deterioration factor (DF) will apply only to
the SOF and RCP (i.e., sulfates will remain constant throughout
the useful life of the engine) and will be applied
proportionally to the SOF and RCP such that their relative
amounts remain constant throughout engine life. This
assumption is made due to the limited knowledge of how the
factors that affect particulate deterioration impact
particulate composition.
3. Pre-1988 HDDE Particulate Emissions
The data used to obtain average HDDE particulate emission
rates for pre-1988 engines were taken from the Draft Regulatory
Impact Analysis for the HDDE NOx/Particulate Rulemaking and is
reproduced in Table 3-1.[1] At that time transient test data
were available on a number of current engines (representing
approximately 66 percent of HDDE sales of that timeframe) and
therefore the data are fairly representative of engines
produced in the early 1980s. It should be noted that these
engines were very low mileage and could essentially be
considered to represent zero-mile emission levels. A
discussion of in-use deterioration will be presented shortly.
Sales weighting these data (using actual sales data from
1985) yield average zero-mile particulate emission levels of
0.46 g/BHP-hr for light HDDEs (LHDDEs), 0.66 g/BHP-hr for
medium HDDEs (MHDDEs) and 0.63 g/BHP-hr for heavy HDDEs
(HHDDEs). While engines from the 1970's probably emit at
-------
3-10
Table 3-1
HDDE Emission Levels - Pre-1988 Engines
Manufacturer/
Engine
Light HDDE:
Detroit Diesel
V8-6.2
Medium HDDE:
Caterpillar
3208D1T
3208D1NA
3208D1NA
3208D1NA
Cummins
NH250
NH250
Detroit Diesel
8.2T
8.2T
8.2T
8 . 2T
8 . 2T
8 . 2T
8 . 2T
Particulate
(g/BHP-hr)
0 . 46
0 . 59
0.63
0 .88
1. 09
0. 72
0.83
0 . 43
0.43
0 .
0
0
1
1
44
67
69
06
42
Navistar (International Harvester)
DTI466B 0.81
DTI466C 0.62
DTI466C 0.62
DTI466C 0.57
DTI466C 0.57
DT466C 0.56
DT466C 0.53
DT466C 0.43
NOx
(g/BHP-hr)
3 .01
10. 00
8.49
4.30
4.13
8.11
6.80
5,
6,
6.
5,
5
4
4
99
80
88
04
60
17
73
05
64
62
50
95
50
, 05
, 65
-------
3-11
Table 3-1 Cont'd
HDDE Emission Levels - Pre-1988 Engines
Manufacturer/
Engine
Heavy HDDE:
Caterpillar
3306DITA
3406DITA
3406DITA
Cummins
NTCC240
NTC350
NTC350
NTCC400
VTB903
Detroit Diesel
6V92TA
6V92TA
8V71N
8V71TA
8V92TA
8V92TA
8V92TA
8V92TA
8V92TAS
Mack
EC6-235
EM6-250
EM6-250R
EM6-285
EM6-300
EM6-300R
E6-350
Particulate
(g/BHP-hr)
0.73
0.81
0.58
0 . 77
0 . 70
0.58
0.85
0. 72
NOx
(q/BHP-hr)
0,
0,
0
0
0
0
0
0
0
83
83
69
42
46
41
34
52
44
0.81
0 .40
0
0
0
0
0
49
61
55
55
36
9 . 02
4.78
8.20
78
00
23
,31
,27
4 .92
4 . 59
6.31
7 . 03
8 . 68
9 . 53
11 . 65
5 . 61
9 . 24
9 . 73
9 . 11
8.81
6.97
8.20
7.98
8 . 61
-------
3-12
somewhat higher levels than this due to higher fuel consumption
and less control of HC emissions and smoke, these levels will
be used for all pre-1988 engines for two reasons. First, an
accurate assessment of older engine emission levels is
difficult due to the limited amount of transient test data
available on them. Second, due to the declining number of
these older engines still in operation, their impact on future
air quality projections is small. Because actual certification
data for 1988 LHDDEs show particulate levels greater than 0.46
g/BHP-hr (i.e., 0.514 g/BHP-hr), and because it is not likely
that particulate emission levels from LHDDE increased
coincidentally with the introduction of particulate standards,
it will be assumed that SOF and RCP emissions from pre-control
(pre-88) LHDDE* s are the same as those from 1988-1990 model
year engines. Average end of life emissions from pre-'88 model
year vehicles are shown in Table 3-2.
4. 1988-1990 HDDE Particulate Emissions
The data used to determine 1988-1990 model year HDDE
particulate emissions are presented in Table 3-3 and
graphically in Figure 3-5. This data set is made up entirely
of 1988 certification data and represents nearly every major
manufacturer and engine line. Using manufacturer's projected
1988 sales (which cannot be reproduced here due to their
confidential nature), average 1988 particulate certification
levels for the three HDDE subclasses were calculated to be
0.514 g/BHP-hr for LHDDEs, 0.479 for MHDDEs, and 0.436 g/BHP-hr
for HHDDEs. Similarly, sales weighting the certification
deterioration factors (DF) yields DFs of 0.009 g/BHP-hr for
light HDDEs, 0.033 g/BHP-hr for medium HDDEs and 0.016 g/BHP-hr
heavy HDDEs. These DFs will be subtracted proportionally from
the SOF and RCP as discussed in the previous section on
particulate composition to derive the zero-mile particulate
levels and compositions. Using the BSFC values and SOF
percentages presented in the composition action, the end of
life particulate compositions were calculated. The particulate
levels and compositions for all HDDE subclasses for the
1988-1990 timeframe are presented in Table 3-4.
5. 1991-1993 HDDE Particulate Emissions Scenarios
EPA has been monitoring HDDE manufacturers' progress
toward meeting the 1991 NOx and particulate standards. Based
on progress to date it is EPA's assessment that manufacturers
on average will be able to meet the 1991 HDDE particulate
standard without aftertreatment using low sulfur fuel (0.05
percent sulfur by weight) and possibly even with current
(baseline) fuel. Two future particulate emission scenarios
were developed for this study. The first, which will be
considered the nominal scenario, assumes that manufacturers on
average will be able to meet the 0.25 g/BHP-hr standard with
engine-out reductions alone (i.e., without aftertreatment)
-------
3-13
Table 3-2
Average Pre-1988 HDDE End of Life Particulate Levels
and Estimated Compositions
Class
Light HDDE
Medium HDDE
Heavy HDDE
Urban Bus
Total
(q/BHP-hr)
0.5156
0.6946
0.6444
0.6931
SOF
(q/BHP-hr)
0.2195
0.2744
0.1399
0.2737
S04+H20
(q/BHP-hr)
0.0852
0.0678
0.0615
0.0710
RCP
(q/BHP-hr)
0.2109
0.3492
0.4430
0.3484
-------
3-14
Manufacturer/
Engine
Light HDDE:
Cummins
4BT3.9
Detroit Diesel
V8-6.2L
Hino
W04C-TE
W04C-TD
Mitsubishi
4D31-OAT
Navistar
7.3L
Medium HDDE:
Caterpiller
3208T
3208 ATAAC
Cummins
6BT5.9
6BTA5.9
6CT8.3
6CTA8.3
Daimler-Benz
OM366-190
OM366-170
Detroit Diesel
8.2N
8.2T
Ford
7.8L-210
7.8L-215
7.8L-240
7.8L-215C
7.8L-185F
6.6L-165F
6.6L-170
Table 3-3
1988 HDDE Certification Data
NOx Particulate Particulate DF
(g/BHP-hr) (g/BHP-hr) (g/BHP-hr)
5.2 0.50 0.04
3.2 0.537 0.02
6.8 0.54 0.00
5.4 0.51 0.00
5.72 0.54 0.04
5.67 0.57 0.04
4.7 0.50 0.00
5.2 0.60 0.128
4.4 0.60 0.128
5.6 0.47 0.04
4.7 0.55 0.04
4.6 0.58 0 . 12
5.7 0.50 0.04
5.6 0.41 0.04
5.4 0.44 0.02
7.8 0.513 0.02
8.3 0.406 0.00
5.9 0.376 0.00
5.4 0.40 0 . 00
8.5 0.34 0.00
5.3 0.42 0.00
7.6 0.415 0.042
8.4 0.597 0.042
5.2 0.578 0.042
-------
3-15
Medium HDDE:
Table 3-3 Cont'd
1988 HDDE Certification Data
Manufacturer/
Engine
NOx Particulate
(q/BHP-hr) (q/BHP-hr)
Particulate DF
(q/BHP-hr)
Hino
EM100-E 7.6 0.59 0.04
H07C-G 9.4 0.47 0.00
H07C-F 5.5 0.40 0.00
H06C-TJ 5.3 0.44 0.00
H06C-TH 5.7 0.52 0.02
Isuzu
65A1T 5.6 0.31 0.028
Navistar
DT-360 8.2 0.46 0.02
DTA-360 5.9 0.52 0.02
DT466-210 8.4 0.45 0.02
DT466-185F 8.6 0.43 0.02
DT466-185C 5.6 0.53 0.02
DT466-240 5.0 0.47 0.02
DT466-245 7.8 0.44 0.02
Nissan
FE6-M 4.9 0.46 0.04
FE6-A 5.0 0.48 0.04
FE6T-M 5.3 0.30 0.00
FE6T-A 5.5 0.26 0.00
NE6T-M 5.0 0.44 0.00
NE6T-A 5.1 0.41 0.00
Perkins
180TI 9.46 0.582 0.04
Volvo
TD71-230 10.3 0.52 0.04
TD71-227 10.7 0.56 0.04
TD71-FCQ 5.4 0.54 0.04
-------
3-16
Table 3-3 Cont'd
1988 HDDE Certification Data
Manufacturer/
Engine
NOx
(q/BHP-hr)
Particulate
(q/BHP-hr)
Particulate DF
(q/BHP-hr)
Heavy HDDE:
Caterpillar
3306 PCTA
3306B
3306B ATAAC
3406B ATAAC
Cummins
NHC-250
NTC 093E
NTC 093J
KTA-19
LTA10
Detroit Diesel
6V92TA DDEC
S60-11.1L
S60-12.7L
Mack
E9-500
EM6-300
EM6-300L
EM6-275L
Volvo
TD102FBQ
TD102FGQ
TD102FCQ
TD122FAQ
TD122FCQ
TD122FEQ
TD122FGQ
5.5
7.3
5.8
6.0
9,
7,
5,
7,
5
9
8
7
74
0
78
7 . 5
5.6
6.6
5.1
9
10
5
6
9
5
5
0 . 59
0.56
0.46
0 . 51
0
0
0
0
0
0
0
0
43
46
38
60
38
363
42
42
0.38
0 .51
0 .39
0 . 47
0
0
0
0
0
0
0
37
38
40
43
39
39
47
0 . 04
0 . 04
0 . 04
0 . 00
0 . 00
0.00
0 . 01
0.04
0.00
0. 07
0 . 17
0.17
0 . 04
0 . 00
0 .01
0 .01
0
0
0
0
0
0
0
04
04
04
00
00
00
00
-------
11
10
9
8
7
6
5
4
3
FIGURE 3-5
NOx LEVEL vs. PARTICULATE LEVEL
~
Dn ~
~
~
+
~
+ + + +
+
+
~
+ +
~
+ + n
j + [p
~
+ oD D + s + ~+
+ +a+++o + + +°
+
o
~~1 oS 1 (M5 1 0.^5
PARTICULATE LEVEL (g/BHP-hr)
~ HHDDE + MHHDE o LHHDE
-------
3-18
Table 3-4
Average 1988 HDDE Certification Particulate Levels
and Estimated Compositions
Class
Light HDDE
Medium HDDE
Heavy HDDE
Urban Bus
Total
(q/BHP-hr)
0.5140
0 . 4790
0 .4360
0.4790
SOF
(q/BHP-hr)
0.2195
0.1809
0.0899
0.1796
S04+H20
(q/BHP-hr)
0.0836
0.0678
0.0615
0.0709
RCP
(q/BHP-hr)
0.2109
0.2303
0.2846
0.2285
-------
3-19
using low sulfur fuel. The second scenario, the low emissions
scenario, assumes that manufacturers on average will be able to
meet the standard with engine-out reductions alone using
current (baseline) fuel.
In order to meet the 0.25 g/BHP-hr standard manufacturers
must establish a target level somewhat below the standard in
order to account for engine to engine production variability,
as discussed earlier in this Chapter. According to the
methodology outlined there, meeting the 0.25 g/BHP-hr standard
essentially means meeting a zero-mile target level of about
0.20 g/BHP-hr, or an end of life target level of about 0.22,
assuming a DF of 0.02 g/BHP-hr, as described below. For the
low emissions scenario this target level will represent the
average certification (end of life) emissions rate for all
three HDDE subclasses which would result using current sulfur
level (0.25 percent by weight) fuel. However, for the nominal
scenario, this target level represents the average emission
rate which would result from using low sulfur (0.05 percent by
weight) fuel. The average emission level on baseline fuel in
the nominal emissions case is higher, due to the higher amount
of sulfates produced. The difference between sulfate (and
bound water) levels using low sulfur fuel and baseline fuel was
added to the target level to yield the average emission rates
for the nominal scenario on baseline fuel (0.29 g/BHP-hr for
LHDDE's, 0.27 g/BHP-hr for MHDDE's and HHDDE's, and 0.28
g/BHP-hr for urban buses).
Due to the small amount of composition data available on
future development engines, as well as uncertainty as to how
new engine technology will impact particulate composition, two
projections of composition will be used for all 1991 and 1994
scenarios. Instead of using the average percent SOF of
non-sulfate particulate as calculated in Section I-A-2, this
value was bracketed by +10 percent. This was done since the
benefits of catalysts will depend on the level of SOF and it is
important to determine if the results of this analysis are
sensitive to the level of SOF. Thus, for light HDDEs each case
will be run using non-sulfate SOF levels of 61 and 41 percent.
For medium HDDEs these values will be 54 and 34 percent and for
heavy HDDEs and urban buses 34 and 14 percent will be used.
The average end of life emission levels and compositions for
both the nominal and low emission scenarios are presented in
Table 3-5.
In order to meet a certification level of 0.22 g/BHP-hr
the zero-mile level must be lower than this to account for
deterioration. Due to a lack of durability testing data on
1991 development engines, there is some uncertainty as to what
DFs will look like in this timeframe. However, they are not
expected to change significantly from 1988 levels. In looking
at the 1988 DFs there is a great variability among the HDDE
subclasses. This variability is generally due to the small
-------
RCP
/BHP-]
. 1211
. 1376
. 1801
. 1371
. 0801
.0959
. 1382
. 0956
. 0824
. 1026
. 1378
. 1005
. 0545
. 0715
. 1057
. 0701
3-20
3-20
Table 3-5
Projected 1991-1993 Average End of Life
Emission Levels and Compositions
Total SOF SO4H2O
(g/BHP-hr) {g/BHP-hr) (g/BHP-hr)
Nominal Scenario - Low SOF Case
0.2873
0.0842
0.0820
0.2747
0.0709
0.0662
0 .2709
0.0293
0.0615
0.2772
0.0707
0.0694
Nominal
Scenario -
High SOF Case
0 .2873
0.2747
0.2709
0.2772
0.1252
0.1126
0.0712
0.1122
0.0820
0.0662
0.0615
0.0694
Low Emission Scenario - Low SOF Case
0.2217 0.0573 0.0820
0.2217 0.0529 0.0662
0.2217 0.0224 0.0615
0.2217 0.0518 0.0694
Low Emission Scenario - High SOF Case
0.2217
0.2217
0.2217
0.2217
0.0852
0.0840
0 .0545
0.0822
0.0820
0.0662
0 .0615
0.0694
-------
3-21
number of engines tested and the fact that the standard is not
very stringent and manufacturers have not yet had to develop
low DF engines. It is uncertain whether this variability will
remain in the future due to likely changes in product offerings
by some manufacturers. Therefore, a sales-weighted average of
the 1988 HDDE subclass DFs yields a DF of 0.020 which will be
used for all HDDE subclasses in the 1991-1993 timeframe.
6. 1994 and Later HDDE Particulate Emissions
The same scenarios presented for 1991-1993 emissions
(nominal and low) including the two composition cases (high and
low SOF) will be retained for modelling 1994 and later
emissions. Due to uncertainties in the particulate reduction
potential of advanced engine technology which will be
implemented in 1994, an accurate assessment of 1994 particulate
levels is difficult. However, rapid advances are being made in
engine-out particulate controls, as evidenced by the large
decrease in emissions expected by 1991, and EPA believes
continued progress beyond 1991 is likely. To model this
progress, it was assumed for all emissions scenarios and HDDE
subclasses that there would be a 35 percent reduction in
non-sulfate particulate from 1991 end-of-life levels by 1994.
Sulfate levels tend to be primarily a function of fuel
consumption and will not likely be reduced by engine technology
but only by reductions in fuel consumption (conservatively
assumed here to be zero). Thus, a 35 percent reduction in
non-sulfate particulate translates into a 23-30 percent
reduction in overall particulate, depending on the engine
subclass and emissions scenario. Due to uncertainties in the
effect of future control technologies on particulate
composition, this 35 percent reduction was applied
proportionally to both the SOF and RCP portions of the
particulate (i.e., a 35 percent reduction will be applied to
each). The average 1994 end of life particulate levels for all
HDDE subclasses and emissions scenarios are presented in Table
3-6.
For the 1991-1993 timeframe an additive DF of 0.020
g/BHP-hr was used. However, some advances are expected to be
made by 1994 in the areas generally associated with particulate
deterioration such as oil control. To model this progress a 25
percent reduction in deterioration was assumed. Therefore, a
DF of 0.015 g/BHP-hr was used for all 1994 and later model year
HDDE emission scenarios.
7. Engine-out Emissions Distributions
In the preceding sections average particulate levels were
projected for the 1991 and 1994 timeframes for all three HDDE
subclasses. However, in order to model aftertreatment
compliance strategies more accurately under the emissions
averaging program, the distribution of emissions from various
-------
3-22
Table 3-6
Projected 1994 and Later Average
End of Life Emission Levels and Compositions
Engine
Subclass
Total
(q/BHP-hr)
SOF
(q/BHP-hr)
S04H20
(q/BHP-hr)
RCP
(q/BHP-hr)
Nominal Scenario - Low SOF Case
Light HDDE
Medium HDDE
Heavy HDDE
Urban Bus
0.2159
0.2041
0.1985
0.2053
0.0562
0.0474
0.0196
0.0468
0.0788
0.0646
0.0583
0.0678
0.0809
0.0921
0.1206
0.0908
Nominal Scenario - High SOF Case
Light HDDE
Medium HDDE
Heavy HDDE
Urban Bus
0.2159
0.2041
0.1985
0.2053
0 . 0836
0.0753
0.0477
0.0743
0.0788
0.0646
0.0583
0.0678
0.0535
0.0642
0.0925
0.0633
Low Emission Scenario - Low SOF Case
Light HDDE 0.1727
Medium HDDE 0.1692
Heavy HDDE 0.1660
Urban Bus 0.1691
0.0385
0,0356
0 .0151
0.0344
0.0788
0.0646
0.0583
0.0678
0.0554
0.0690
0.0926
0.0669
Low Emission Scenario - High SOF Case
Light HDDE
Medium HDDE
Heavy HDDE
Urban Bus
0 .1727
0.1692
1660
1691
0
0
0 .0573
0 . 0565
0.0366
0.0547
0.0788
0.0646
0.0583
0.0678
0.0366
0.0481
0.0711
Q.0466
-------
3-23
engines about the average level must also be projected. This
section describes the method used to arrive at a standardized
distribution and then discusses its application specifically to
the 1991 and 1994 emission scenarios.
In the absence of sufficient data on the emission
distributions of future engines the next step is to examine the
distribution of current engine emissions. The 1988
certification data presented in the discussion of 1988-1990
emissions are shown in Figure 3-6 in a sales vs. emission level
format. It can be seen from this figure that the majority of
HDDEs are clustered around the 0.5 g/BHP-hr range with some as
high as 0.6 g/BHP-hr (the level of the 1988 standard) and a
larger number spread out below the 0.5 g/BHP-hr level all the
way down into the 0.3 g/BHP-hr range. Analyzing the subclasses
separately, the heavy HDDEs also follow this trend, while the
medium HDDEs tend to be more evenly spread out and light HDDEs
are all clustered around the 0.5 g/BHP-hr level. While it is
apparent from this graph that there is a definite skewed
distribution in 1988 there are two main factors which lead EPA
to conclude that this distribution is not appropriate for the
1991/1994 timeframes. First, in 1988 all engines are required
to meet the 0.60 g/BHP-hr standard whereas in 1991 and 1994
under the emissions averaging program engines can be produced
which do not meet the standards as long as all engines on
average meet the standards. For this reason EPA expects some
manufacturers to continue to produce some engines lines which
emit well above the standard based on their proven performance
in the marketplace. Second, the 1991 and 1994 standards are
significantly more stringent than the 1988 standards and EPA
does not expect to see a large number of engines emitting at
levels substantially below the level of the 1991 and 1994
standards. For these reasons a log-normal distribution skewed
toward higher emissions was used to represent the 1991 and 1994
emissions distributions.
The first step in arriving at the 1991 log-normal emission
distributions was the construction of a discrete standardized
normal distribution. The discrete distribution was assumed to
contain eleven intervals. This number is small enough to keep
the analysis to a manageable level yet large enough to allow
sufficient resolution of the distribution to adequately model
aftertreatment compliance strategies. Also, a total range of
+3 standard deviations (+3Z), which covers 99.74 percent of a
normal population was used. Dividing the range of +3Z into
eleven discrete segments yielded the twelve segment end points
and their corresponding population fractions (here sales
fractions) as shown in Table 3-7. Choosing the midpoint for
each segment to represent that segment yields the Z values and
the corresponding fraction of engines represented by each of
the eleven intervals, also shown in Table 3-8.
-------
400
350
300
250
200
150
100
50
0
FIGURE 3-6
ENGINE SALES vs. PARTICULATE LEVEL
\// / / / / ;-%z\
0.27
033
HHDDE
0.39 0.45
PARTICULATE (g/BHP-hr)
057
MHDDE
^ LHDDE
-------
3-25
Table 3-7
Standardized Discrete Normal Distribution
Standardized Normal Deviate
(Z-value) Cummulative Sales Fraction
-3 . 00
0.0013
-2.45
0.0071
-1.91
0.0281
-1.36
0.0869
-0.82
0.2061
-0 . 27
0.3936
0.27
0.6064
0.82
0 . 7939
1.36
0.9131
1.91
0.9719
2.45
0.9929
3.00
0.9987
-------
3-26
The relationship needed to move from this standardized
normal distribution to a log-normal distribution is:
log(x) = Z log (°) + i0g (y) (eq, 3-2)
or:
x = exp (Z log(a)+log (M))
where:
log (V) = mean value of normal distribution (i.e.,
logarithm of emissions)
log (°) = standard deviation of the normal distribution
z = standardized normal deviate
x = mean value of interval point in the
log-normal distribution (i.e., emissions)
In order to make use of this equation both the mean value
and standard deviation of the normal distribution (log x) must
be known. The mean value for all scenarios were derived in the
previous sections and the log of these values can simply be
found. Standard deviations were chosen which provided
reasonable upper and lower bounds for the scenario being
considered. This is described below in more detail for the
1991, nominal emissions case. Since the distribution of only
the non-sulfate portion of the particulate is being
constructed, here the ratio of the maximum values for the low
emission scenarios was assumed to be the same as the ratio of
the mean emissions of the two scenarios.
In light of progress being made on particulate control
technology, for the nominal, 1991 scenario, an upper bound of
0.40 g/BHP-hr total particulate appears to give reasonable
lower bounds for all cases. For the lowest emission case (low
emission scenario for HHDDEs) this yields a zero-mile total
particulate level of 0.1405 g/BHP-hr. All the other vehicles
classes in the 1991 low emission scenario and all vehicles
classes in the nominal emission scenario have higher lower
bounds.
First, the calculated sulfate levels were subtracted from
this value to give the class-specific non-sulfate maximum
values. Class-specific non-sulfate maximum values for the low
emission scenario were then calculated, as described above.
Finally, Equation 3-2 was used along with the non-sulfate
maximums and means (Z = 2.725) to yield the standard deviations
for all vehicle classes of the two 1991 emission scenarios.
However, because the urban bus market is so limited in nature
with respect to the number of engines available, the
distribution of emissions about the mean is expected to be much
narrower than for the medium HDDE class. To model this
difference the standard deviation for the urban bus scenarios
-------
3-27
was assumed to be one half of the standard deviation of the
corresponding medium HDDE scenarios.
These standard deviations were used in Equation 3-2 along
with the non-sulfate means and the Z values from Table 3-8 to
calculate the non-sulfate end of life 1991 emission
distributions. The DF of 0.02 g/BHP-hr was then subtracted
from each point in each distribution to yield the non-sulfate
zero-mile distributions for 1991. Each point in each
distribution was then broken down into its SOF and RCP
components according to the compositions presented in Section
II-A-5. Following this the calculated sulfate level for each
scenario was added to each point in the non-sulfate
distributions to obtain the zero-mile total particulate
engine-out emission distributions for 1991. The zero-mile
distributions for all 1991 emission scenarios and HDDE
subclasses are presented in Appendix 3-A.
As was discussed in Section II-A-6 on 1994 emissions, both
emission scenarios for 1994 assume a 35 percent reduction in
non-sulfate particulate from 1991. To achieve this, the SOF
and RCP values for each point in each 1991 zero-mile
distribution will be reduced by 35 percent. The 1994 zero-mile
distributions for all emission scenarios and HDDE subclass are
also presented in Appendix 3-A.
B. Light-Duty Diesel Emissions
Although the vast majority of highway diesel fuel is
consumed by engines classified as heavy-duty, any modification
to diesel fuel quality will also have an effect on emissions
from light-duty diesel vehicles (LDDVs) and trucks (LDDTs). In
order to determine the magnitude of this effect, the first step
to be performed is to establish baseline emission values for
both classes of vehicles.
A number of studies of emissions from light-duty diesel
vehicles have been performed in which the fraction of fuel
sulfur converted to particulate sulfate has been
measured.[4-7] These studies show that typical sulfur to
sulfate particulate conversions range from 1 to 2 percent of
the sulfur in the fuel. The mean conversion level of 1.5
percent was used in conjunction with fuel economy data to
determine emissions of sulfate, bound water, and sulfur dioxide
for LDDV and LDDT.
Baseline total particulate emissions for LDDV's sold in
model years prior to 1987 were estimated from pre-1987
certification test results. Sulfate emissions were calculated
based on the 1.5 percent conversion of fuel sulfur to sulfate
and model year specific fuel economy data. SOF was estimated
to be 18 percent of the total particulate, the average of
several LDDV engines tested in a particulate trap evaluation
-------
3-28
Table 3-8
Standardized Normal Distribution Interval Parameters
Mid-Point Z-value Fraction of Engine Sales
-2.725
0.0058
-2.180
0 . 0210
-1.635
0.0588
-1.090
0.1192
-0.545
0.1875
0 .000
0.2128
0.545
0.1875
1.090
0.1192
1.635
0.0588
2.180
0.0210
2.725
0.0058
-------
3-29
program.[8] The remainder of the particulate by definition was
assumed to be RCP. For 1987 and later LDDV's, 1987
certification test results were used to estimate non-sulfate
particulate emissions emissions. SOF was estimated to be 18
percent of total particulate. Sulfate emissions were then
calculated using MOBILE3 fuel economy projections. As a
simplifying assumption, particulate levels for all LDDV's were
assumed to remain constant over the full life of the vehicle at
the end-of-life level. Emissions of HC were taken from
AP42.[9] Baseline end-of-life emissions for LDDVs are shown in
Table 3-9.
Baseline emissions for LDDTs less than 6,000 pounds GVWR
(LDDTls) were determined in the same way as LDDVs. Once again,
total particulate end of life emissions were estimated from
certification test results. For LDDT1, it was assumed that the
SOF represented 50 percent of total particulate, based on test
data obtained on a light-duty diesel truck.[10] Sulfate, bound
water, and sulfur dioxide emissions were determined assuming
I.5 percent sulfur to sulfate conversion as explained above.
RCP emissions were calculated by subtracting SOF and sulfate
emissions from total particulate levels. As with LDDV's
particulate emissions were assumed to remain constant at
end-of-useful-life levels over the life of the vehicle.
Emissions of HC were taken from AP42.[9] Baseline emissions
for LDDTls are shown in Table 3-9.
Total particulate emissions for LDDTs over 6,000 pounds
GVWR (LDDT2s) for model years prior to 1988 were determined
from certification test data. Sulfate and sulfur dioxide
emissions were determined from fuel economy data assuming a 1.5
percent sulfur to sulfate conversion as described above. The
remainder of the particulate was assumed to be SOF and RCP, in
relative proportions the same as those of light heavy-duty
diesel engines because the only engine used in LDDT2s is also a
LHDDE. Total engine-out particulate emissions from engines
certifying for model year 1988 and beyond were estimated to be
0.295 grams per mile, as projections by manufacturers
indicated. No reduction in engine-out emissions beyond 1988
levels is anticipated (i.e., reductions necessary to comply
with the 1991 standard of 0.13 gpm will be achieved by the use
of trap-oxidizers). Baseline emissions for LDDT2 are shown in
Table 3-9.
II. Exhaust Aftertreatment Technology
In this section, the various types of exhaust treatment
devices which may be employed to meet the 1991 and 1994
heavy-duty particulate standard, both trap-oxidizer and flow
through catalyst type systems, will be presented. Both the
emission reduction potential and the costs of such devices will
be examined.
-------
3-30
Table 3-9
Baseline Emissions - LDDV
Model Fuel Econ Emissions (q/mi)
Year
(MPG)
SOF
RCP
S04
TPM
S02
HC*
LDDVs
1979
25. 1
0.046
0. 178
0.035
0.260
0 . 651
0.42
1980-81
25. 1
0.046
0 . 178
0.035
0.259
0.651
0.29
1982-84
27. 1
0 . 046
0 .178
0 . 032
0 .256
0.603
0.29
1985-86
28.4
0 . 046
0. 178
0.031
0 .255
0 . 575
0 .29
1987
28.4
0.024
0.079
0.031
0.134
0.575
0 .29
1988-90
29.8
0.024
0 . 079
0.029
0.132
0 . 549
0.29
1991-93
31.4
0.024
0.079
0.028
0 .131
0 . 520
0 .29
1994 +
34.2
0.024
0.079
0.025
0.128
0.477
0.29
LDDT1
1979-80
21.0
0.135
0.099
0.041
0 .275
0 . 779
0.86
1981
21.0
0.135
0.099
0.041
0 .275
0 . 779
0 . 43
1982-84
23.0
0.135
0 . 099
0.038
0 .272
0 . 711
0 . 43
1985-86
23.8
0.135
0.099
0 .036
0 .270
0 . 687
0.43
1987
23.8
0.108
0 . 072
0 .036
0 .216
0 . 687
0.43
1988-90
24.7
0 .108
0.072
0 .035
0 .215
0 . 662
0 . 43
1991-93
25.9
0.108
0.072
0 .033
0 .213
0. 631
0.43
1994 +
27.8
0.108
0.072
0.031
0.211
0 . 588
0 . 43
LDDT2
1979-80
21. 0
0 .188
0 .167
0 . 041
0.396
0 . 778
0.86
1981
21.0
0 .188
0 . 167
0 . 041
0 .396
0 . 778
0.43
1982-84
23 .0
0 . 188
0 . 167
0.038
0 .393
0.710
0.43
1985-86
23.8
0.188
0. 167
0. 036
0.391
0 . 686
0.43
1987
23.8
0 .180
0 . 159
0 . 036
0.375
0.686
0.43
1988-90
24 . 7
0 . 138
0 . 122
0 . 035
0.295
0 . 661
0.43
1991-93**
25. 4
0. 043
0 . 012
0. 034
0.089
0.637
0 .15
1994+**
27.3
0 . 044
0 . 013
0.032
0.089
0. 594
0 .15
Zero-mile emissions.
Emissions levels reflect the use of trap oxidizers.
-------
3-31
A. Exhaust Aftertreatment Types
1. Trap-Oxidizer Systems
A trap-oxidizer system is an exhaust treatment device
which collects solid particulate matter in the exhaust and
subsequently oxidizes the collected particulate (i.e.,
regenerating itself) at certain intervals. Possible devices
used for trapping of the particulate matter include ceramic
wall-flow monolith, ceramic fiber coil, and wire mesh media.
Possible methods of regeneration include the use of burner
assemblies, electrical heaters, fuel additives, and catalyst
formulations for passive regeneration. A brief overview of
these different systems is presented in this section.
a. Ceramic Wall-Flow Monolith
The wall-flow monolith trap is designed similarly to the
ceramic honeycomb substrate used in gasoline-fueled vehicle
catalytic converters. Alternate upstream and downstream cells
of the honeycomb matrix are plugged, causing particulate-laden
exhaust gases to filter through the porous walls. The result
is a high efficiency capturing of the carbonaceous particulate
components which must in turn be periodically regenerated or
burned off. The system can be either uncatalyzed or may be
catalyzed with either base or noble metals. A more complete
description of this device can be found in "An Updated
Assessment of the Feasibility of Trap Oxidizers."[11]
b. Johnson-Matthey Catalyzed Wire Mesh
The catalyzed wire mesh trap is a passive regeneration
system which traps particulate and subsequently oxidizes it
under certain operating modes with the aid of a catalyst. The
wire mesh system has the attraction of not requiring an
expensive regeneration system, but does not trap particulate as
efficiently as the ceramic monolith system. Problems with
increased sulfate emissions due to the catalyst remain
unresolved for current high sulfur fuels. For a more detailed
description of the operation of such a system the reader is
referred to the Diesel Particulate Study.[12]
c. Ceramic Fiber Coil
This type of trap was developed as part of a cooperative
effort by Daimler-Benz and Mann & Hummel.[13] The ceramic
fiber coil trap consists of a number of perforated stainless
steel cylinders wound with silica fibers (the trapping medium)
and arranged in parallel in a stainless steel housing. The
number of windings affects the filtration efficiency as well as
the back pressure associated with the trap. Regeneration of
the trap takes place via some type of oxidant injection
system. [14] For more details on this type of trap, the reader
-------
3-32
is referred to the RIA for the 1985 NOx/Particulate
Rulemaking.[1]
2. Flow-through Catalyst
A low-cost alternative to the trap-oxidizer type system is
a flow-through oxidation catalyst similar to those used on
gasoline vehicles. Work aimed at developing both ceramic
monolith type and pellet type catalysts for diesel application
is currently underway. The operational requirements are such
that the catalyst formulation must satisfactorily oxidize the
organic particulate while minimizing the oxidation of SO2 to
particulate sulfate if high sulfur fuel is used. Typically,
those catalysts which are the most active on the organic
particulate also tend to oxidize sulfur dioxide to the greatest
extent. Determining the optimal catalyst formulation and
loading is the subject of a considerable research effort.
B. Exhaust Aftertreatment Efficiencies
Accurately estimating the efficiency of the performance of
the aftertreatment systems discussed above on the various
components of diesel emissions is a critical link in projecting
their application on vehicles certifying with the 1991 and 1994
particulate standards. A great deal of test data exist in the
literature, much of which is contained in several SAE special
publications on diesel particulate control. Both transient and
steady state testing on engine and chassis dynamometer have
been performed on these systems. A discussion of system
efficiencies by type is presented below.
1. Trap Oxidizer Systems
a. Ceramic Wall-Flow Monolith (Uncatalyzed)
The analysis of the test data was approached with some
degree of selectivity; particularly in the areas of ceramic
monolith testing, where a substantial body of data exist.
Since emission characteristics of heavy-duty engines are
different in steady-state and transient modes, and because the
federal test procedure for heavy-duty diesel engines is a
transient test, only data generated in the FTP transient mode
were considered. Data generated from light duty vehicles were
excluded in this analysis, due to the difference in trap sizes
and particulate composition for these vehicles. After these
criteria were applied, the data used in the heavy-duty analysis
consisted of SAE papers, EPA reports and confidential engine
manufacturer submissions.[13,15,16] A summary of the results
of the tests performed on heavy-duty engines is shown in Table
3-10. Though confidential data received from engine
manufacturers are not presented, consideration of these data
were included in determining the "Best Estimate Efficiencies"
shown in the table. Many ceramic monolith traps of varying
-------
lft
Heavy-Duty Uncatalyzed Ceramic Monolith Efficiency-Transient Testing
Source
Enqme{ Vehicle)
Trap
Vol.
(Cu In)
Cat
SOF
Effici
RCP
ency (Percent Reduction)
S04 TPM* HC
CO
EPA
[15]
NTC400
EX47
1193
N
—
—
82
8
-50
EPA
[16]
DDAD 6V71
EX47
1193
N
37.5
93.5
66.8 61.3
0.5
14.1
EPA
[16]
(GMC RTS II)
EX47
1193
N
87.1
93.0
15.4 92.1
34.6
16.6
SAE
[13]
(MB 0 405 Bus)
—
—
N
—
70-95
51-82
4-15
0
Best Estimate
=
N
55
90
0 71*
13
-6
Contingent on engine emission characteristics.
-------
3-34
size, porosity and pitch have been tested. Efficiencies and
fuel economy penalties vary with design parameters of the
monolith system, although general conclusions about the
efficiencies of the ceramic monolith trap can be drawn.
As illustrated in Table 3-10 the ceramic particulate
filters design acts primarily on the solid carbonaceous portion
of the diesel particulate. Solid carbon (RCP) trapping
efficiencies reported generally range from 70 to 95 percent.
The average efficiency yielded from the available data is
approximately 90 percent. The efficiency of the trap on the
SOF fraction of particulate is significantly less than this, as
most passes through the trap in the gaseous phase. However,
some of the SOF particulate is adsorbed onto the trapped RCP,
and by this mechanism is controlled to a certain extent.
Although the SOF efficiency varies widely from test to test,
the data indicate that a SOF reduction of 55 percent on average
can be expected.
Many references in the literature also report decreases in
the sulfate portion of the particulate with the use of ceramic
monolith traps. However, no mechanism by which the trap might
reduce the sulfate to SO2 seems likely, and theoretically no
reduction in the amount of sulfate formed in the engine is
expected. It is probable that the apparent reduction in sulfate
particulate levels by the trap is attributable to temporary
sulfate storage in the trap, which, upon the eventual release
of the sulfate from the trap will disappear. It is therefore
concluded that sulfate particulate emission will not be lowered
by using an uncatalyzed ceramic monolith trap.
A reduction in gaseous HC emission with ceramic trap use
has also been reported in some of the literature [13, 15, 16],
although the actual degree of reduction varies widely. The
preponderance of data suggest that on average a 13 percent
reduction in HC emissions will result from the use of an
uncatalyzed ceramic monolith trap, although actual reductions
may range from 0 to 50 percent depending on operating
conditions. The data also suggest an increase in emissions of
CO of 6 percent. Actual test data reported range from a 50
percent increase to a 17 percent decrease.
b. Ceramic Monolith (Catalyzed)
The emission reduction efficiencies described above are
those which can be expected for an uncatalyzed ceramic monolith
trap. Some attention has been given to ceramic monolith trap
designs implementing base and/or noble metal catalysts in order
to facilitate passive regeneration and/or increase overall
efficiency. Based on data in the literature as well as
confidential data submitted, the efficiency of the catalyzed
trap on the RCP fraction of the particulate appears to be about
the same as that of the uncatalyzed ceramic trap. The data
-------
3-35
show that the efficiency of the catalyzed trap on the SOF
fraction is much higher than the uncatalyzed version.
As expected, however, the use of high-activity catalysts
with the ceramic monolith may greatly increase emissions of
sulfate, up to seven times the engine out level with the highly
active catalyst formulations which appear to be necessary for
dedicated passive regeneration. This is a major problem
preventing the use of high activity catalysts in these devices
to meet 1991 and 1994 particulate standards without fuel sulfur
control.
A passive regenerating ceramic monolith system not
requiring a supplemental active regeneration system as a backup
has not yet been satisfactorily demonstrated for heavy-duty
engines, and therefore can not currently be considered a
feasible option. It may, however, in some cases be desirable
to use a catalyst with a active regeneration ceramic trap in
order to increase the SOF reduction capability of the trap.
This type of device would be particularly attractive for
light-heavy-duty engines applications, where the SOF comprises
a substantial portion of the particulate. While this type of
trap would still require an auxiliary regeneration system, the
resulting increase in particulate collection efficiency above
and beyond an uncatalyzed ceramic trap might in some cases be
the most cost effective approach to particulate control.
In order to estimate the efficiency of such a device, the
particulate reduction efficiency of an uncatalyzed particulate
trap was combined multiplicatively with the efficiency of the
flow-through oxidation catalyst devices described later in this
chapter. Efficiencies for two types of catalyzed traps ("high"
and "low" cost) were determined, analogous to the two types of
flow-through catalysts systems described below. The efficiency
of the "high cost" catalyzed trap was determined by combining
the efficiency of an uncatalyzed ceramic trap with the
efficiency of the "high cost" or "high activity" flow-through
catalyst. The efficiency of the "low cost" catalyzed trap was
determined likewise. These emission reduction efficiencies are
shown in Table 3-11.
It is worthy to note that, while some data are available
in the literature on the particulate collection efficiency of
catalyzed ceramic monolith traps, these data were not used
directly in this analysis. Due to the proprietary nature of
the research, in many instances it is difficult to determine
much about the catalyst formulation and loading corresponding
to any specific piece of test data beyond whether the catalyst
was all noble metal or a mixture of base and noble metals. To
assure consistency between the efficiency and cost of the
catalyzed traps used in this study and the efficiency and cost
of the other aftertreatment devices used, efficiencies were
determined by combining the efficiencies of uncatalyzed traps
and flow through catalysts as described above.
-------
3-36
Table 3-11
Efficiency of Catalyzed Ceramic Monolith Traps
Efficiency (Percent Reduction)
Type SOF RCP _S04_ HC CO
Low Cost-Low 69 90 0 39 -6
Efficiency
High Cost-High 91 91 -100 83 90
Efficiency
-------
3-37
c. Johnson Matthey Catalyzed Wire Mesh
Trapping efficiencies of Johnson Matthey catalyzed wire
mesh traps as found in the available literature are shown in
Table 3-12.[15,17,18,19] Total particulate trapping
efficiencies for this trap type (excluding the type I trap
tested by Ulman [17]) range from 46 to 86 percent. It should
be pointed out, though, that some of the higher total
particulate efficiencies reported were taken from high SOF
engines. Only two reports in the literature gave a
compositional breakdown of the particulate
efficiencies.(17,19] The average of particulate collection
efficiencies reported were 87 percent on SOF, 47 percent on
RCP, and -136 percent on sulfate (i.e., additional 136 percent
over engine-out sulfate).
d. Ceramic Fiber Coil
Little test data exist in the literature on the ceramic
fiber coil particulate trap. However, results of some
development work and demonstration programs have been
documented by Hardenberg and others in the SAE
literature.[13,14] The transient test results that have been
reported show that these devices' total particulate and HC
emission reduction efficiencies are similar to that of the
ceramic monolith system. Soot trapping efficiency of the
ceramic fiber coil is somewhat higher than that of the ceramic
monolith, approaching 100 percent as the trap becomes loaded.
A slight increase in CO emissions (about 10 percent) occurs
with the use of the ceramic fiber coil.
2. Flow Through Catalyst
Little recent information exists in the literature on the
performance of flow through catalysts on diesel emissions,
although the results of steady-state testing of an Engelhard
catalyst on a single-cylinder engine are available.[20] The
majority of the information on catalyst efficiency was obtained
through confidential submittals and conversations with engine
and catalysts manufacturers. Because of the confidential
nature of the material, little discussion can be presented here.
Of major concern in developing an oxidation catalyst for
diesel engines is determining a catalyst formulation which
satisfactorily oxidizes the organic particulate while
minimizing the oxidation of sulfur dioxide to sulfate
particulate. Typically, the formulations which provide the
highest reduction in organic particulate also tend to produce
the most sulfate.
Based on the data received, two representative types of
flow through catalysts have been used in this analysis. One is
a "high efficiency, high cost" system. Theoretically, assuming
-------
Table 3-12
HDD Johnson Matthey Wire Mesh Efficiency
Efficiency (Percent Reduction)
Vol.
Source
Enqine
Trap
(Cu In) Catalyst
SOF
RCP
S04
TPM*
HC
CO
[15]
NTC400
JM
Y
46
86
-9
[17]
DDAD 8V-71TAC
JM I
Y(aged)
82
-100
-250
25
80
60
[17]
It
JM II
Y(aged)
90
32
-136
64
85
46
[18]
Scania DS11 15
JM
Y
—
—
—
71**
40
71
[19] Mining JM — Y 72-97 60-63 Stor. 65-86
Best Estimate 87 47 -136 64 70 36
Contingent on engine emission characteristics.
Operation on low sulfur fuel.
-------
3-39
that the conversion of the catalyst is reaction rate limited
{i.e., that the system can be designed so that mass transfer is
not the limiting factor), it should be possible to design a
system for a diesel vehicle achieving the same conversion of
gaseous hydrocarbons (or SOF) as is seen in gasoline vehicles.
One would merely have to design volume and catalyst loading for
the different reaction rates and reaction concentrations in the
exhaust. The concentration of SOF in diesel exhaust is
generally less than the HC concentration in gasoline exhaust,
and temperatures are generally lower in diesel exhaust, and so
a larger catalyst may be necessary in a diesel vehicle for the
same percent conversion. The concentration of oxygen in diesel
exhaust is typically higher than in gasoline exhaust, however,
which may enhance the performance of the catalyst and reduce
the size reguirement (or at least eliminate the need for an air
pump).
In summary, developing a "high efficiency" system capable
of oxidizing most of the SOF fraction of the particulate should
be feasible. Reductions of 80 percent in SOF (similar to the
reduction of HC in gasoline vehicles) should be achievable with
proper design. The use of such a catalyst should also increase
sulfate emissions. An increase in sulfate of 100 percent is
believed to be reasonable. A nominal reduction in RCP of 10
percent should also be seen, as evidenced by some of the data.
A low cost, low efficiency oxidation catalyst should also
be feasible. Using a less active catalyst formulation, lower
particulate reduction efficiencies would be achieved. Based on
confidential submittals, a reduction in SOF of 30 percent with
no effect on the RCP or sulfate portion of the particulate
would be expected. Efficiency estimates for of the flow
through catalyst technologies used in this study are summarized
in Table 3-13.
The effect of oxidation catalyst use on gaseous
hydrocarbon emissions is not clear. Test results published in
the literature as well as confidential submittals indicate that
the HC reduction efficiency of all flow through catalysts are
on the order of 60 percent. There is no reason, however, to
expect that the effect of the catalyst on SOF emissions should
be significantly different than the effect of catalysts on HC
emissions. Both HC and SOF emissions should be in the gaseous
phase when they pass over the catalyst, and should be acted
upon by the catalyst in a similar manner. Therefore, in this
analysis, it will be assumed that the HC reduction efficiency
of the catalyst will be the same as the SOF reduction
efficiency. As shown in Table 3-13, HC efficiency is estimated
to be 30 and 80 percent for the low cost and high cost flow
through catalysts, respectively.
CO emission reduction characteristics of an Englehard
flow through catalyst have also been reported in the
-------
3-40
Table 3-13
Efficiency of Flow Through Oxidation Catalysts
Efficiency (Percent Reduction)
Type SOF RCP _S04_ HC CO
Low Cost-Low 30 0 0 30 0
Efficiency
High Cost-High 80 10 -100 80 90
Efficiency
-------
3-41
literature.[20] Under high power application, the CO reduction
efficiency was high, approximately .90 percent. Tests results
on other, lower activity, catalyst systems showed no effect on
CO emissions. For this analysis, therefore, it was estimated
that the low cost and high cost flow through catalysts would
have CO reduction efficiencies of 0 and 90 percent,
respectively.
C. Exhaust Aftertreatment Device Costs
Information on the costs of the emission control
technologies considered in this analysis were taken from the
March 1985 Regulatory Impact Analysis accompanying the
NOx/Particulate Rulemaking, and a 1978 report prepared under
contract for EPA by Rath and Strong.[1,2] Costs developed from
these studies were adjusted to 1986 dollars according to the
Producers Price Index for Internal Combustion Engines.
1. Trap-Oxidizer Systems
a. Uncatalyzed Ceramic Monolith
The 1985 NOx/Particulate RIA[1] included cost estimates
for trapping devices and regeneration systems for three classes
of heavy-duty vehicles, roughly equivalent to the three classes
of heavy-duty vehicles used in this study. Cost estimates for
a ceramic monolith trap, with a bypass burner regeneration
system, were generated assuming trap volumes of 11, 21, and 39
liters for LHDDE, MHDDE, and HHDDE respectively. Costs for the
trap and regeneration system are included in Table 3-14.
The NOx/Particulate RIA also quantified increased
maintenance costs to be born by the purchaser of a trap
equipped vehicle. Those costs, discounted back to the year of
purchase were determined to be $68 for LHD, $110 for MHD, and
$136 for HHD diesels.
In addition to hardware and maintenance costs, it was
estimated in the NOx/Particulate RIA that each of the seven
largest HDDE manufacturers would spend about $2.8 million ($2.9
million in 1986 dollars) in general trap research and
development. For the purpose of this study, it was assumed
that approximately one half of this money has already been
spent, and that only half of this ($1.45 million) remains to be
spent. Also, in the NOx/Particulate RIA it was assumed that
specific system designs would cost an additional $230,000
($235,000 in 1986 dollars) in R&D per engine family.
The 1987 Model Year Certification Test Results show a
total of 101 heavy-duty diesel engines families certified; 9
LHDDEs, 49 MHDDE s, and 43 HHDDE s. Of the 49 MHDD engine
families, it was projected based on indications given by engine
manufacturers that two families would be used in urban buses,
-------
3-42
Table 3-14
Trap System Costs
LHDDE MHDDE HHDDE
Uncatalyzed Ceramic Monolith
Trap Cost* 103 176 299
Regen System* 275 282 289
Maintenance 68 110 136
R & D _10 _85 79
Total 456 653 803
"High-Efficiency" Catalyzed Ceramic Trap
Trap* 103 176 299
Regeneration* 275 282 289
Catalyst* 56 111 141
Maintenance 68 110 136
R&D _10 _85 79
Total 512 764 944
"Low Efficiency" Catalyzed Ceramic Trapp
Trap* 103 176 299
Regeneration* 275 282 289
Catalyst* 9 19 24
Maintenance 68 110 136
R&D _10 _85 79
Total 465 672 827
"Low Cost" Trap Cost Estimate
Trap Cost* 75 108 143
Regen System* 154 17 0 209
Maintenance 40 83 141
R&D _10 _85 79
Total 279 446 572
Retail price equivalent (RPE)
-------
3-43
while the remaining 47 families would be used in non-urban bus
MHDDVs. While uncertainties in the future bus market render
this projection to be speculative at best, this breakdown of
MHDDV and bus engine families seems to be a reasonable
projection for the purpose of apportioning R&D costs between
vehicle types. Accordingly, in this analysis, general trap
research and development costs were divided among the four
classes according to the relative number of engine families in
each. Engine family specific R&D expenditures were determined
according to the number of families in each subclass projected
to require traps under a given fuel scenario. R&D costs were
assumed to be spent on a schedule over four years preceding the
production of the trap (two years assuming introduction in
1991). The schedule for R&D expenditures, analogous to that
used in the NOx/Particulate RIA is shown in Table 3-15. These
research and development costs were then discounted to their
present valued in the year of the initial trap application, and
the cost was assumed to be recovered by the manufacturer over
the first three model years of sales.
Using this accounting procedure, the per-vehicle increase
in cost due to research and development expenditures can be
calculated. For example, assuming traps are introduced in
1994, and that 50 percent of vehicles would require traps, the
per vehicle equipped R&D costs will be $10, $85, and $79 for
LHDDVs, MHDDVs, and HHDDVs vehicles, respectively. The sum of
the maintenance cost, the trap cost, and the R&D costs for the
burner regeneration ceramic monolith trap-oxidizer system is
shown in Table 3-14.
b. Catalyzed Ceramic Monolith
As described previously, it was assumed that two different
types of catalyzed ceramic monolith systems would be
available. One system would be loaded with a catalyst similar
to that of the "high efficiency" flow through catalyst. The
hardware cost of this device was assumed to be equivalent to
the cost of a complete uncatalyzed trap oxidizer system with a
burner regeneration system plus the catalytic metal cost of the
high efficiency flow through catalyst system described below,
Maintenance and R&D costs were assumed to be the same as that
of an uncatalyzed trap. Costs for the "high efficiency"
catalyzed ceramic trap are also shown in Table 3-14.
The cost of the "low efficiency" version of the catalyzed
ceramic monolith trap was determined by the same method as the
high efficiency version, catalytic metal process were taken
from the cost estimates of the "low efficiency" flow through
catalyst system described later in this chapter. These costs
are also shown in Table 3-14.
-------
3-44
Table 3-15
R&D Cost Expenditures for Aftertreatment Devices
Introduction in 1991
Year Percent of R&D Spent
1987 0
1988 0
1989 88
1990 12
Introduction in 1994
Year Percent of R&D Spent
1990 19
1991 47
1992 30
1993 4
-------
3-45
c. Johnson Matthey Catalyzed Wire Mesh
No attempt to determine the cost of this device was made
for this study. While this system remains a viable option for
particulate control, current unresolved problems related to
increased sulfate formation make it difficult to base the
results this study on the feasibility of this device.
Furthermore, due to the relatively low efficiency of this trap
on carbonaceous particulate (approximately one half of that of
the ceramic monolith), it is doubtful that this system will be
able to compete economically with the ceramic monolith even if
sulfate production problems are resolved. In any event, the
cost effectiveness of this device should not be substantially
greater than that of the ceramic monolith under any scenario.
Since the purpose of this study is to determine the cost
savings associated with fuel control, and not necessarily to
appraise in detail every possible type of aftertreatment
device, there should be little error involved in using the
ceramic monolith as the primary candidate for a particulate
trap.
d. Ceramic Fiber Coil
As with the catalyzed wire mesh system, it is difficult to
determine whether the ceramic design will be available for
particulate control in 1991 and 1994. Durability problems
still need to be resolved. The rational for not basing the
results of this study on the feasibility of such a device are
the same as in the case of the wire mesh trap.
It may, of course, be the case that a cheaper system such
as the ceramic fiber coil system described in the
NOx/Particulate RIA will be available. To examine the impact
of such a development, a sensitivity case was run (Chapter 4)
using a low cost trap alternative. For simplicity, it was
assumed that the emission reductions characteristic of the low
cost trap were the same as the uncatalyzed ceramic monolith,
but that the cost would be similar to that developed for the
ceramic fiber coil in the NOx/Particulate RIA. As explained
therein, costs for the regeneration system would be
significantly lower, due to the different regeneration
technique employed. Since no trap bypass would be needed,
there would also be a credit equivalent to the cost of a
standard exhaust pipe realized, as well as a decrease in
maintenance costs due to the replacement of the standard steel
exhaust pipe with a stainless steel pipe. There would likely
be an increase in maintenance costs, however, due to the need
to periodically replenish the catalyst supply. The R&D cost
were determined in the same way as those of the ceramic
monolith trap. The "low-cost" trap oxidizer system consumer
cost estimate is shown in Table 3-14.
-------
3-46
2. Flow Through Oxidation Catalyst System
The costs of the flow through oxidation catalyst systems
were calculated in accordance with the procedure developed in a
cost estimation report prepared under contract for EPA.[21]
The methodology was altered by using 1986 noble metal prices
and allowing for the effects of inflation on other costs
according to the Producer Price Index. The markup on vendor
costs originally used in the referenced report was also changed
to more accurately reflect industry practice.[22]
The type, size, geometry, and loading characteristics of
the flow through catalysts which may be applied to heavy-duty
diesels was not readily available, as this technology (for
diesel engines) is still in the development stage and most
information is highly proprietary. The direct application of
gasoline catalyst technology to diesel engines is not
necessarily feasible, due to differences in exhaust
composition, exhaust flow rate, temperature, etc. A catalyst
formulation designed to minimize sulfate formation while
satisfactorily oxidizing organic particulate is required. The
low temperature (relative to gasoline) of the exhaust and the
high volume of exhaust flow need to be designed for as well.
Also of concern is the carbonaceous particulate in the exhaust,
which may reduce the efficiency of the catalyst by occupying
some of the active sites. To summarize, the design of such a
system requires a detailed evaluation of many parameters.
While all of the inputs necessary to determine the cost of
a flow-through catalyst system for heavy-duty diesels are not
known, conversations with members of industry and confidential
information received made it possible to size and cost out two
systems for heavy-duty diesels. Using information received on
catalyst loading and volume on prototype systems, the necessary
catalyst requirements were determined for each subclass.
Catalyst loadings (in gram/cubic foot) were assumed to be the
same for all subclasses. Catalyst volumes for LHD, MHD, and
HHD were determined according to relative exhaust flow rates.
Using this information, the "high-efficiency" oxidation
catalyst system cost was calculated and is presented in Table
3-16. As shown in the table, the flow-through catalyst device
cost alone was determined to be $94 for LHDDVs, $159 for
MHDDVs, and $194 for HHDDVs. In addition to the hardware cost,
to insure the proper functioning of the catalyst for the full
useful life of the vehicle, stainless steel exhaust pipes
between the exhaust manifold and catalyst will be necessary.
The cost for this alteration, including a credit for the
standard steel exhaust pipe replaced, as estimated in the
NOx/Particulate RIA will be $34 for LHDDVs, $43 for MHDDVs, and
$73 for HHDDVs.[l] Chassis heat shields will also be required,
estimated to be $11 per vehicle equipped. [231 The use of this
type of control technology is not expected to increase
maintenance costs for vehicle owners. Conversely, the
-------
3-47
Table 3-16
Catalyst System Costs
LHDDE MHDDE HHDDE
High Efficiency Catalyst
Catalyst 94 159 194
SS Exhaust 34 43 73
Maintenance -42 -52 -83
R&D 5 43 39
Heat Shields 11 11 ll
Total 102 204 234
Low Efficiency Catalyst
Catalyst 48 67 77
SS Exhaust 34 43 73
Maintenance -42 -52 -83
R&D 5 43 39
Heat Shields 11 11 11
Total
56
112
117
-------
3-48
improvements in exhaust system materials due to the use of a
catalyst system will actually decrease maintenance costs. The
maintenance savings from the elimination of the need to replace
part of the exhaust system when discounted back to the time of
purchase were determined to be $42 for LHD, $52 for MHD and $83
for HHD[1].
As in the case of particulate traps, expenditures for
Research and Development will be recovered over the first three
years of sales after the technology is marketed. R&D expenses
for catalyst systems are expected to be approximately one-half
of that required for traps (i.e., each of the seven largest
manufacturers would spend a total of $1.45 million in general
research and development, only half of which remains to be
spent, plus $117,500 per engine family using a catalyst.) The
R&D funds are assumed to be spent according to the schedule
shown in Table 3-15. Table 3-16 also shows an example of
estimated R&D costs for each vehicle class (generated under the
assumption that 50 percent of engine families require catalysts
in 1994.)
The consumer cost of the low cost, low efficiency flow
through catalyst system was determined by similar methodology
and is also shown in Table 3-16.
D. Exhaust Aftertreatment Device Deterioration
The performance deterioration of the three types of
systems which have been indicated as primary strategies, the
burner regeneration uncatalyzed ceramic monolith system, the
burner regeneration catalyzed ceramic monolith, and the flow
through oxidation catalyst, will be addressed in this section.
1. Uncatalyzed Ceramic Monolith Deterioration
Durability testing of a Corning noncatalyzed trap on a
1980 Mercedes 300SD engine was performed by SwRI for EPA.[8]
Emissions tests over 80,000 km of operation showed no
deterioration in particulate collection efficiency of the trap,
with overall particulate collection efficiency ranging from 87
to 93 percent. Based on these data, it was assumed that no
deterioration in trap collection efficiency would occur over
the full life of the vehicle.
2. Flow through Catalyst Deterioration
The loss of efficiency of automotive catalysts is
primarily due to the presence of lead in gasoline. Since lead
is present in much lower forms in diesel fuel, it will be
assumed that catalyst performance will not deteriorate
significantly due to lead.
-------
3-49
A study of the effect of fuel sulfur on the operation of a
gasoline automotive catalyst compared hydrocarbon conversion
efficiency over 15,000 miles of operation on fuels with sulfur
contents of 0, 0.03, and 0.09 weight percent. The data show
that when fuel containing 0.03 and 0.09 weight percent sulfur
is used, HC oxidation efficiency is reduced to 80 percent of
baseline (zero-mile) efficiency. The effect appears to take
place during the first 5,000 miles, after which no additional
deactivation due to sulfur appears to take place. Since diesel
fuel sulfur levels under consideration in this study are 0.05
weight percent or greater, it is expected that this same
deactivation will take place in diesel flow through oxidation
catalysts. For calculation of emissions in this analysis,
therefore, the end-of-life efficiency of a flow through
catalyst device will be assumed to be 80 percent of baseline
(zero-mile) oxidation efficiency.
3. Catalyzed Ceramic Monolith System
With the catalyzed ceramic monolith system, one would
expect the deterioration in efficiency to be some combination
of the 20 percent decrease in efficiency of the flow-through
catalyst and the zero percent efficiency reduction of the
uncatalyzed ceramic trap. The efficiency reduction factor of
this system type was therefore calculated according to the
relative reductions in particulate via trapping and catalysis.
This calculation yields an average end-of-life particulate
reduction efficiency of 91 percent of the zero-mile efficiency
for the "high efficiency" catalyzed system and 98 percent of
the zero-mile efficiency for the low efficiency system.
Ill, Aftertreatment Technology Mix for Compliance with Standards
A. Methodology
1, Light-Duty Vehicles
The analysis of exhaust aftertreatment requirements for
light-duty vehicles was not necessary for LDDV and LDDT1, since
no aftertreatment devices are needed to comply with particulate
standards under present conditions. In the case of the
Mercedes-Benz 30 0D, the recent recall of these systems and
discontinuation of its sale in the U.S. make it difficult to
predict much about future use of the system.
Rulemaking proceedings currently in progress on the LDDT2
class indicate that a 0.13 gram per mile particulate emission
level will be required for these vehicles beginning in the 1991
model year. It was assumed that this standard would be met by
applying ceramic monolith traps to 100 percent of the vehicles
beginning in 1991. Because of the similarity of these engines
to those in LHDDE class, the same trap costs as those developed
for LHDDE vehicles were used. No research and development
-------
3-50
costs were included, as the same amount will be spent whether
fuel control is implemented or not.
The emission level determined by applying 100 percent
traps to projected 1991 engine-out levels was defined as the
emission target level for achieving the 0.13 gpm standard for
this vehicle class. This target level was subsequently used to
determine the fraction of LDDT2 requiring traps under any given
fuel control scenario (only 99 percent traps would be needed in
1994, due to expected fuel economy improvements).
2. Heavy-Duty Vehicles
The lowest cost scenario for meeting particulate emission
standards was determined for each heavy-duty vehicle class, for
model years 1991 and 1994, assuming no fuel control. Using the
engine-out emission distributions described previously for each
vehicle class, a compliance strategy model was used to apply
the various types of aftertreatment devices described in this
chapter to the highest emitting engines in each vehicle class
until the end of life emission target level was reached.
Aftertreatment devices were applied to the highest emitting
engines according to the criterion that the lowest cost per ton
of emission control was achieved. Modeling was performed using
both the "nominal-high SOF" and the "nominal-low SOF" emission
distributions discussed earlier in this chapter. The model
results for both nominal emission distributions were then
averaged together. These averaged results were used in the
cost effective analysis of this report. It should be noted
that, according to this methodology, not every vehicle was
required to comply with the particulate standards. Rather,
emission averaging was allowed over the entire vehicle class.
In the modelling it was assumed that trap-oxidizers would
have a two percent fuel economy penalty associated with their
use (one percent due to trap backpressure, one percent for the
regeneration system), as estimated in the Draft NOx/Particulate
RIA.[25] While a lower fuel economy penalty (1.5 percent) was
used in the NOx/Particulate RIA, trap volumes currently under
development are significantly smaller than those assumed
therein.[1] Since the large trap volumes in the Final RIA used
as a basis for adjusting the estimated fuel economy penalty for
a burner regenerated ceramic monolith system downward from 2.0
to 1.5 percent no longer seem likely, the 2.0 percent fuel
economy penalty used in the Draft RIA will be used here.
B. Results
1. Light Duty Vehicles
As described in Section A, 100 percent of the LDDT2 fleet
will be equipped with traps in 1991, at a cost of $455 per
vehicle. Emission factors and aftertreatment costs for 1991
light duty diesels are shown in Table 3-17.
-------
le
Light-Duty Diesel Aftertreatment Costs and Emissions
Vehicle
Class
1991-1993
Percent
Traps
LDDV 0
LDDT1 0
LDDT2 100
Aftertreatment
Cost($/vehicle)
0
0
455
Fuel Econ
(mpq)
31.4
25.9
25.4
SOF
Emissions (q/rtti)
RCP
§P4_
TPM
§02_
HC*
0.024 0.079 0.028 0.131 0.520 0.29
0.108 0.072 0.033 0.213 0.631 0.43
0.043 0.012 0.034 0.089 0.637 0.15
1994 and Later
LDDV 0
LDDT1 0
LDDT2 99
0
0
450
34.2 0.024 0.079 0.025 0.128 0.477 0.29
27.8 0.108 0.072 0.031 0.211 0.588 0.43
27.3 0.044 0.013 0.032 0.089 0.594 0.15
*
Zero mile levels.
-------
3-52
The 1994 and later MY emission factors and aftertreatment
costs for light-duty diesels are also shown in Table 3-17. As
shown, only 99 percent of LDDT2s required traps to meet the
emission standard. This is lower than the 100 percent
utilization in 1991, due to projected increases in fuel
economy. Since less fuel is consumed per mile, less sulfate
will be produced, and overall engine-out emissions will be
slightly lower. The slight decrease in the nuntber of traps
required in 1994 is the result.
2. Heavy-Duty Vehicles
Results for 1991-93 heavy-duty vehicles are shown in Table
3-18. Aftertreatment costs (not including fuel economy
penalties) range from $157 per vehicle for LHDDVs to $205 for
HHDDVs to meet the 0.25 g/BHP-hr standard. Costs for urban
buses are significantly higher ($656) due to the more stringent
particulate standards. Due to the contribution of sulfate to
total particulate, it was not possible for urban bus engines
which convert 2 percent of fuel sulfur to sulfate particulate
(see Chapter 4 for detailed analysis) to meet the 0.10 g/BHP-hr
standard, even with 100 percent traps. Certifying to the 0.10
g/BHP-hr standard therefore will require the use of an engine
which converts less than 1.5 percent of fuel sulfur to sulfate
and still require the use of traps costing $656 per vehicle.
The urban bus cost estimates in Table 3-18 reflect the use of a
"low sulfate producing" engine, the availability of which is
difficult to predict.
Table 3-18 also shows the modeling results for heavy-duty
vehicles sold in 1994 and beyond. Aftertreatment costs per
vehicle range from $462 for LHDDVs to $702 for HHDDVs (not
including fuel economy penalties). Because of the low emission
standard and the large contribution of sulfate to total
particulate, it was not possible for LHDDVs, MHDDVs, and urban
buses to meet the standards, even with 100 percent trap
application. Certifying to the 0.10 standard with current
sulfur fuel will require the use of vehicles which convert less
than 2.0 percent of fuel sulfur to sulfate; as low as a 1.5
percent conversion would be required for LHDDVs. Table 3-18
reflects the use of these "low sulfate producing" engines.
-------
Table 3-18
Heavy-Duty Diesel Aftertreatment Usage and Emissions
Veh.
Class
1991-93
LHDDE
MHDDE
HHDDE
BUS
Uncatly.
Traps
0
0
11
0
Low Cost
Catalyzed
Traps
27
21
9
100
Low Cost
Flow
Through
Catalysts
50
39
29
0
Average
Aftertreat-
ment Cost
($/Vehicle) SOF
157
192
205
656
0.070
0.063
0.038
0.027
End of Life
Emissions (q/BHP-hr)
RCP
0.072
0.095
0.125
0.015
_S04
0.082
0.066
0.061
0.052
TPM
0.224
0.224
0.224
0.094
sg2_
1.16
0.94
0.87
1.00
HC
0.66
0.82
1.04
0.44
CO
4.0
4.4
4.9
4.6
1994 and Later
LHDDE 0
MHDDE 0
HHDDE 43
BUS 0
100
99
45
100
0
0
0
0
462
642
702
659
0.023
0.018
0.014
0.018
0.008
0.011
0.020
0.010
0.063
0.064
0.059
0.066
0.093
0.093
0.093
0.093
1.14
0.93
0.84
0.98
0.36
0.37
0.67
0.36
3.6
3.6
3.6
3.6
-------
3-54
References (Chapter 3)
1. "Regulatory Impact Analysis, Oxides of Nitrogen
Pollutant Specific Study and Summary and Analysis of Comments -
Control of Air Pollution from New Motor Vehicles and New Motor
Vehicle Engines: Gaseous Emission Regulations for 1987 and
Later Model Year Light-Duty Vehicles, and for 1988 and Later
Model Year Light-Duty Trucks and Heavy-Duty Engines;
Particulate Emission Regulations for 1988 and Later Model Year
Heavy-Duty Diesel Engines," EPA, OAR, OMS, March 1985. Docket
A-80-18.
2. "Determination of a Reliable and Efficient Diesel
Particulate Hydrocarbon Extraction Process," Shirish A Shimpi
and Ming Li Yu, SAE Paper 811183, 1981.
3. "An Improved Method for Determining the Hydrocarbon
Fraction of Diesel Particulates by Vacuum Oven Sublimation," R.
Halsall, M.L. McMillan, B.J. Schwartz, SAE Paper No. 872136,
1987.
4. "Measurement of Sulfate and Sulfur Dioxide in
Automotive Exhaust," Melvin N. Ingalls, and Karl J. Springer,
Southwest Research Institute, prepared for EPA, OAWM, OMSAPC,
ECTD, EPA-460/3-76-015, August 1976.
5. "Characterization of Particulate and Gaseous
Emissions From Two Diesel Automobiles as Functions of Fuel and
Driving Cycle," Charles T. Hare, Southwest Research Institute,
Thomas M. Baines, U.S. Environmental Protection Agency, Paper
790424 present at SAE congress and Exposition, Detroit,
February 26 - March 2, 1979. (Available in Docket # A-80-18,
II-I-18).
6. "Characterization of Sulfates, Odor, Smoke, POM and
Particulates From Light and Heavy Duty Engines - Part IX,"
Springer, Karl J., prepared for U.S. Environmental Protection
Agency by Southwest Research Institute, EPA-460/3-79-007, June
1979. (Available in Docket A-80-18, II-A-6).
7. "Characterization of Exhaust Emissions From Diesel
Powered Passenger Cars With Particular Reference to Unregulated
Components," Lies, K.H., Postulka, A., and Gring, H. ,
Vokswagenwerk A.G., SAE Paper #840361.
8. "Diesel Car Particulate Control Methods," C.M.
Urban, L.C. Landman, R.D. Wagner, SAE Paper No. 830084, 1983.
9. "Compilation of Air Pollutant Emission Factors,
Volume II: Mobile Sources," AP-42 4th Edition, September 1985.
-------
3-55
10. "Effects of Fuel Properties and Engine Design
Features on the Performance of a Light-Duty Diesel Truck - A
Cooperative Study," E.G. Barry, J.C. Axelrod, L.J. McCabe, T.
Inoue, and N. Tsuboi, SAE Paper No. 861526, 1986.
11. "An Update Assessment of the Feasibility of Trap
Oxidizers," Regulatory Support Document, J. Alson and R.
Wilcox. U.S. EPA, OANR, OMS, ECTD, SDSB, June 1983, Public
Docket No. A-82-32.
12. "Diesel Particulate Study," SDSB, ECTD, OMS, OAR,
EPA, November 1983. Available in Docket KA-80-18.
13. "Experiences in the Development of Ceramic Fiber
Coil Particulate Traps," H.O. Hardenberg, H.L. Daudel, and H.J.
Erdmannsdorfer, SAE Paper No. 870015, 1987.
14. "Particulate Trap Regeneration Induced by Means of
Oxidising Agents Injected Into the Exhaust Gas," H.O.
Hardenberg, H.L. Daudel, H.J. Erdmannsdorfer, SAE Paper No.
870016, 1987.
15. "Heavy-Duty Engine Exhaust Particulate Trap
Evaluation," Charles M. Urban, Southwest Research Institute,
EPA 460/3-84-008, September, 1984.
16. "Emission Characteri2ation of a 2-Stroke Heavy-Duty
Diesel Coach Engine and Vehicle with and Without a Particulate
Trap," Terry L. Ullman and Charles T. Hare, Southwest Research
Institute, EPA 460/3-84-015, prepared for U.S. E.P.A., OMSAPC,
ECTD, March, 1985.
17. "Emissions Performance of Two Catalyzed Trap
Oxidizers on a Bus Engine," Terry L. Ullman, SAE Paper No.
860132, 1986.
18. "Chemical Analysis and Biological Testing of
Emissions From a Heavy Duty Diesel Truck With and Without Two
Different Particulate Traps," SAE Paper No. 860014, 1986.
19. "Investigation of the CTO Emission Control System
Applied to Heavy-Duty Diesel Engines Used in Underground Mining
Equipment," J.P. Mogan, E.D. Dainty, H.C. Vergeer, A. Lawson,
K.C. West away, J.K. Weglo, and L.R. Thomas, SAE Paper No.
850151, 1985.
20. "Diesel Particulate SOF Emission Reduction Using an
Exhaust Catalyst," G.E. Andrews, I.E. Iheozor-Ejiofor, and S.W.
Pang, SAE Paper No. 870251, 1987.
-------
3-56
21. "Cost Estimations for Emission Control Related
Components/Systems and Cost Methodology Description," LeRoy H.
Lindgren, Rath & Strong, Inc., March 1978, EPA-460/3-78-002.
(Available in Docket # A-80-18, II-A-18).
22. "Update of EPA's Motor Vehicle Emission Control
Equipment Retail Price Equivalent (RPE) Calculation Formula,"
prepared for U.S. EPA by Jack Faucett Associates, September 4,
1985.
23. "Regulatory Analysis and Environmental Impact of
Final Emission Regulations for 1984 and Later Model Year Heavy
Duty Engines," U.S. EPA, OMSAPC, December 1979. (Available in
Docket # A-80-18, II-A-8).
24. "Deactivation of Three-way Catalysts by Fuel
Contaminants:—Lead, Phosphorus and Sulfur," W.B. Williamson,
H.S. Gandhi, M.E. Heyde, and G.A. Zawacki, SAE Paper No.
790942, 1979.
25. "Draft Regulatory Impact Analysis, Oxides of
Nitrogen Pollutant Specific Study and Summary and Analysis of
Comments - Control of Air Pollution from New Motor Vehicles and
New Motor Vehicle Engines: Gaseous Emission Regulations for
1987 and Later Model Year Light-Duty Vehicles, and for 1988 and
Later Model Year Light-Duty Trucks and Heavy-Duty Engines;
Particulate Emission Regulations for 1988 and Later Model Year
Heavy-Duty Diesel Engines," EPA, OAR, OMS. (Docket #A-80-18).
-------
Appendix 3A
Projected 1991 and 1994
HDDE Particulate Emission Distributions
-------
1991 LHP NOMINAL SCENARIO
Percent Total Sulfates and
of Particulate SOF bound water Residual
Engines (q/BHP-hr) (q/BHP-hr) (q/BHP-hr) (q/BHP-hr)
LOW SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0 .1945
0.2066
0.2198
0.2343
0.2500
0.2673
0.2860
0.3065
0.3289
0.3533
0.3800
0.0461
0.0511
0.0565
0.0624
0.0689
0.0760
0 . 0837
0.0921
0.1012
0.1112
0.1222
0,0820
0.0820
0.0820
0.0820
0.0820
0.0820
0.0820
0,0820
0.0820
0.0820
0.0820
0.0664
0.0735
0.0813
0.0898
0.0991
0.1093
0.1204
0.1325
0.1457
0.1601
0.1758
HIGH SOF CASE
0.5800
2.1000
5 . 8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0. 1945
0.2066
0.2198
0.2343
0 . 2500
0.2673
0.2860
0.3065
0.3289
0 .3533
0 .3800
0.0686
0.0760
0 . 0841
0.0929
0.1025
0.1130
0.1245
0.1370
0.1506
0.1655
0.1818
0.0820
0.0820
0.0820
0.0820
0.0820
0.0820
0.0820
0.0820
0.0820
0.0820
0.0820
0.0439
0.0486
0.0538
0.0594
0.0655
0 . 0722
0.0796
0.0876
0.0963
0.1058
0.1162
-------
1991 MHD NOMINAL SCENARIO
Percent Total
of Particulate SOF
Engines (q/BHP-hr) (q/BHP-hr)
Sulfates and
bound water
(q/BHP-hr)
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1764
0.1892
0.2034
0.2189
2359
2547
2752
2979
3227
3500
3800
0
0
0
0
0
0
0
LOW SOF CASE
0.0342
0 .0381
0.0425
0.0473
0526
0584
0648
0718
0795
0880
0973
0
0
0
0
0
0
0
0
0
0,
0,
0,
0
0
0
0
0662
0662
0.0662
0.0662
0662
0662
0662
0662
0662
0662
0662
Residual
(q/BHP-hr)
0,
0,
0
0
0
0
0
0
0
0
0
0760
0849
0946
1053
1171
1300
1442
1598
1770
1958
2165
HIGH SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1764
0.1892
0.2034
0.2189
0.2359
0 .2547
0.2752
0.2979
0.3227
0 .3500
0 . 3800
0.0562
0.0627
0 . 0699
0.0779
0.0866
0 .0961
0.1066
0.1181
0 .1308
0.1447
0.1600
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0540
0.0603
0.0672
0.0748
0.0832
0.0923
0.1024
0.1135
0.1257
0.1391
0.1538
-------
1991 HHD NOMINAL SCENARIO
Percent Total Sulfates and
of Particulate SOF bound water Residual
Engines (g/BHP-hr) (g/BHP-hr) (g/BHP-hr) (g/BHP-hr)
LOW SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5 . 8800
2.1000
0 . 5800
0.1710
0.1840
0.1984
0.2142
0.2317
0.2509
0.2720
0.2952
0.3208
0 .3490
0.3800
0.0153
0.0172
0.0192
0.0214
0.0238
0.0265
0.0295
0.0327
0.0363
0.0402
0.0446
0.0615
0.0615
0.0615
0.0615
0.0615
0.0615
0.0615
0.0615
0.0615
0.0615
0.0615
0 . 0942
0.1054
0.1177
0.1314
0.1463
0.1628
0.1810
0.2010
0.2230
0.2472
0.2739
HIGH SOF CASE
0.5800
2.1000
5 . 8800
11.9200
18 . 7500
21 .2800
18.7500
11.9200
5.8800
2 . 1000
0.5800
0.1710
0.1840
0.1984
0.2142
0.2317
0.2509
0.2720
0.2952
0.3208
0 .3490
0.3800
0.0372
0.0417
0.0466
0,0519
0.0579
0.0644
0.0716
0.0795
0.0882
0.0977
0.1083
0.0615
0.0615
0.0615
0.0615
0.0615
0.0615
0.0615
0.0615
0.0615
0.0615
0.0615
0.0723
0.0809
0.0904
0.1008
0.1123
0.1250
0.1389
0.1543
0.1711
0.1897
0.2102
-------
1991 BUS NOMINAL SCENARIO
Percent Total
of Particulate SOF
Engines (q/BHP-hr) (q/BHP-hr)
Sulfates and
bound water
(q/BHP-hr)
LOW SOF CASE
Residual
(q/BHP-hr)
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
,2136
2215
2298
2385
.2476
2572
,2672
,2777
, 2887
,3002
,3123
0 . 0447
0.0471
0.0497
0.0524
0.0552
0582
0613
0646
, 0680
0715
0
0
0
0
0
0.0753
0694
0694
0694
0694
0694
0694
0694
0694
, 0694
, 0694
, 0694
0
0
0
0,
0995
1049
1107
1167
0. 1230
0.1295
1365
1437
1513
. 1593
, 1676
0,
0,
0
0
0
HIGH SOF CASE
0 . 5800
2.1000
5.8800
11.9200
18 . 7500
21.2800
18.7500
11.9200
5.8800
2. 1000
0.5800
0 . 2136
0.2215
0.2298
0.2385
0.2476
0.2572
0.2672
0.2777
0.2887
0.3002
0.3123
0.0735
0.0776
0.0818
0.0862
0.0909
0.0958
0.1009
0.1062
0 .1118
0.1177
0.1239
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0 . 0706
0.0745
0 . 0786
0.0829
0.0873
0.0920
0.0969
0.1021
0.1074
0.1131
0.1190
-------
1991 LHP LOW EMISSIONS SCENARIO
Percent Total Sulfates and
of Particulate SOF bound water Residual
Engines (g/BHP-hr) (g/BHP-hr) (g/BHP-hr) (g/BHP-hr)
LOW SOF CASE
0 . 5800
2.1000
5.8800
11.9200
18 . 7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1521
0.1604
0.1694
0.1792
0.1899
0.2017
0.2144
0.2284
0.2436
0.2602
0.2784
0.0288
0.0321
0.0358
0.0399
0.0443
0.0491
0.0543
0.0600
0.0663
0.0731
0.0805
0.0820
0.0820
0.0820
0.0820
0.0820
0.0820
0.0820
0.0820
0.0820
0.0820
0.0820
0.0414
0 . 0462
0 . 0516
0.0574
0 . 0637
0.0706
0.0781
0.0864
0.0953
0.1052
0.1159
HIGH SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1521
0.1604
0.1694
0.1792
0.1899
0 . 2017
0 .2144
0.2284
0.2436
0.2602
0.2784
0.0428
0.0478
0.0533
0.0593
0.0658
0.0730
0.0808
0.0893
0.0986
0.1087
0.1198
0.0820
0.0820
0.0820
0.0820
0.0820
0.0820
0,0820
0.0820
0.0820
0.0820
0.0820
0.0274
0.0306
0.0341
0.0379
0.0421
0.0467
0.0516
0.0571
0.0630
0.0695
0.0766
-------
1991MHD LOW EMISSIONS SCENARIO
Percent Total Sulfates and
of Particulate SOF bound water Residual
Engines (g/BHP-hr) (g/BHP-hr) (g/BHP-hr) (g/BHP-hr)
LOW SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1433
0.1529
0.1634
0.1750
0.1877
0.2017
0.2170
0.2339
0.2524
0.2728
0.2951
0.0239
0.0269
0.0301
0 .0337
0.0377
0.0420
0.0467
0.0520
0.0577
0.0640
0.0710
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0532
0.0598
0.0671
0.0750
0.0838
0.0935
0. 1041
0.1157
0.1285
0.1425
0.1580
HIGH SOF CASE
0.5800
2.1000
5.8800
11.9200
18 . 7500
21 .2800
18.7500
11 .9200
5.8800
2. 1000
0 . 5800
0.1433
0.1529
0.1634
0.1750
0.1877
0.2017
0.2170
0.2339
0.2524
0.2728
0.2951
0.0393
0.0442
0 . 0496
0.0555
0.0620
0.0691
0.0769
0.0855
0.0950
0.1053
0.1168
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0378
0.0425
0.0476
0,0533
0 . 0595
0.0664
0.0739
0.0822
0.0912
0.1012
0.1122
-------
1991 HHP LOW EMISSIONS SCENARIO
Percent Total
of Particulate SOF
Engines (q/BHP-hr) (q/BHP-hr)
Sulfates and
bound water
(q/BHP-hr)
LOW SOF CASE
Residual
(q/BHP-hr)
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1405
0.1505
0.1615
0.1736
0.1870
2017
2178
2356
2552
2767
3004
0
0
0
0
0
0
0
0
0,
0
0
0
0
0
0
0
0
0111
0125
0140
0157
0176
0196
0219
0244
0271
0301
0335
0
0
0
0
0
0
0
0
0
0
0
0615
0615
0615
0615
0615
0615
0615
0615
0615
0615
0615
0
0
0,
0
0
0
0
0
0
0
0
0680
0766
0860
0964
1079
1205
1344
1497
1666
1851
2055
HIGH SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21. 2800
18 . 7500
11.9200
5.8800
2.1000
0.5800
0.1405
0 . 1505
0 . 1615
0.1736
0.1870
0.2017
0.2178
0.2356
0.2552
0 .2767
0.3004
0.0269
0.0303
0.0340
0.0381
0.0427
0.0477
0.0531
0.0592
0.0658
0.0732
0.0812
0 . 0615
0.0615
0 . 0615
0 .0615
0.0615
0.0615
0 . 0615
0.0615
0.0615
0.0615
0.0615
0.0522
0.0588
0.0660
0.0740
0.0828
0,0925
0.1032
0.1149
0.1278
0.1420
0.1577
-------
1991 BUS LOW EMISSIONS SCENARIO
Percent Total Sulfates and
of Particulate SOF bound water Residual
Engines (q/BHP-hr) (g/BHP-hr) (g/BHP-hr) (g/BHP-hr)
LOW SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1697
0.1755
0.1816
0.1880
0.1946
0.2017
0.2090
0.2167
0.2248
0.2332
0.2421
0.0311
0.0329
0.0348
0.0368
0.0388
0 . 0410
0.0433
0.0457
0.0482
0.0508
0.0535
0 .0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0692
0 . 0732
0.0774
0.0818
0.0864
0.0913
0.0963
0.1016
0.1072
0.1130
0.1191
HIGH SOF CASE
0 . 5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1697
0.1755
0.1816
0.1880
0.1946
0.2017
0.2090
0.2167
0.2248
0.2332
0.2421
0.0512
0.0541
0.0572
0.0605
0.0639
0.0674
0.0712
0.0751
0 . 0792
0.0835
0 . 0881
0.0694
0.0694
0 . 0694
0.0694
0 . 0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0 . 0492
0.0520
0.0550
0.0581
0 . 0614
0 . 0648
0.0684
0.0722
0.0761
0.0803
0.0846
-------
1994 LHP NOMINAL SCENARIO
Percent Total Sulfates and
of Particulate SOF bound water Residual
Engines (q/BHP-hr) (q/BHP-hr) (q/BHP-hr) (g/BHP-hr)
LOW SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1519
0.1598
0.1684
0.1778
0.1880
0.1992
0.2114
0.2247
0.2393
0.2552
0.2725
0.0300
0.0332
0.0367
0.0406
0.0448
0.0494
0.0544
0.0598
0.0658
0.0723
0.0794
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0431
0.0478
0.0529
0.0584
0 . 0644
0 . 0710
0.0782
0 .0861
0 . 0947
0.1041
0.1143
HIGH SOF CASE
0.5800
0.1519
0.0446
0.0788
0.0285
2.1000
0.1598
0.0494
0.0788
0.0316
5.8800
0.1684
0.0547
0.0788
0.0349
11.9200
0.1778
0.0604
0.0788
0.0386
18.7500
0.1880
0.0666
0.0788
0.0426
21.2800
0.1992
0.0735
0.0788
0.0470
18.7500
0.2114
0.0809
0.0788
0.0517
11.9200
0.2247
0.0890
0,0788
0.0569
5.8800
0.2393
0.0979
0.0788
0 . 0626
2.1000
0.2552
0.1076
0.0788
0 . 0688
0.5800
0.2725
0.1182
0.0788
0.0755
-------
1994 MHD NOMINAL SCENARIO
Percent Total Sulfates and
of Particulate SOF bound water Residual
Engines (g/BHP-hr) (g/BHP-hr) (g/BHP-hr) (g/BHP-hr)
LOW SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11 . 9200
5.8800
2. 1000
0.5800
0.1362
0.1446
0.1537
0.1638
0.1749
0.1871
0.2005
0.2152
0.2313
0.2491
0.2686
0.0222
0.0248
0.0276
0.0308
0.0342
0.0380
0.0421
0.0467
0.0517
0.0572
0.0632
0.0646
0.0646
0.0646
0.0646
0.0646
0 . 0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0494
0.0552
0.0615
0.0685
0.0761
0.0845
0 . 0938
0.1039
0.1150
0.1273
0.1407
HIGH SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21. 2800
18.7500
11.9200
5.8800
2 . 1000
0.5800
0.1362
0.1446
0.1537
0.1638
0.1749
0.1871
0.2005
0.2152
0.2313
0.2491
0.2686
0.0365
0 . 0408
0 . 0455
0.0506
0.0563
0.0625
0.0693
0.0768
0.0850
0.0941
0.1040
0 .0646
0 .0646
0 . 0646
0.0646
0 .0646
0 .0646
0 .0646
0 .0646
0.0646
0.0646
0.0646
0.0351
0.0392
0.0437
0.0486
0 .0541
0.0600
0.0666
0 .0738
0.0817
0.0904
0.0999
-------
1994 HHD NOMINAL SCENARIO
Percent Total Sulfates and
of Particulate SOF bound water Residual
Engines (g/BHP-hr) (g/BHP-hr) (g/BHP-hr) (g/BHP-hr)
LOW SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1295
0.1380
0.1473
0.1576
0.1689
0.1814
0.1951
0.2102
0.2269
0.2452
0.2653
0.0100
0.0112
0.0125
0.0139
0.0155
0.0172
0.0192
0.0213
0.0236
0.0262
0.0290
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0612
0.0685
0.0765
0.0854
0.0951
0.1058
0.1177
0.1306
0.1450
0.1607
0.1780
HIGH SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0 . 5800
0 . 1295
0.1380
0.1473
0.1576
0.1689
0.1814
0.1951
0.2102
0.2269
0.2452
0.2653
0.0242
0.0271
0.0303
0.0338
0.0376
0.0418
0.0465
0.0517
0.0573
0.0635
0.0704
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0470
0.0526
0.0587
0.0655
0.0730
0.0812
0.0903
0.1003
0.1112
0.1233
0.1366
-------
1994 BUS NOMINAL SCENARIO
Percent Total Sulfates and
of Particulate SOF bound water Residual
Engines (g/BHP-hr) (q/BHP-hr) (g/BHP-hr) (g/BHP-hr)
LOW SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21 .2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1615
0.1667
0.1721
0.1777
0.1836
0.1898
0.1963
0.2032
0 .2103
0.2178
0.2257
0.0291
0.0306
0 . 0323
0.0341
0 . 0359
0 . 0378
0.0399
0.0420
0 .0442
0.0465
0 . 0489
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0647
0.0682
0.0719
0.0758
0.0799
0.0842
0.0887
0 . 0934
0.0983
0.1035
0.1089
HIGH SOF CASE
0.5800
2.1000
5.8800
11.9200
18 . 7500
21.2800
18.7500
11.9200
5. 8800
2.1000
0 . 5800
0.1615
0.1667
0.1721
0.1777
0.1836
0.1898
0.1963
0.2032
0.2103
0.2178
0.2257
0.0478
0.0504
0.0532
0.0561
0.0591
0.0622
0.0656
0.0690
0.0727
0.0765
0 . 0805
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0459
0.0484
0.0511
0.0539
0.0568
0.0598
0.0630
0.0663
0.0698
0.0735
0.0774
-------
1994 LHP LOW EMISSIONS SCENARIO
Percent Total Sulfates and
of Particulate SOF bound water Residual
Engines (g/BHP-hr) (q/BHP-hr) (q/BHP-hr) (q/BHP-hr)
LOW SOF CASE
0.5800
2. 1000
5.8800
11.9200
18.7500
21.2800
18 . 7500
11.9200
5.8800
2.1000
0.5800
0.1244
0.1298
0.1356
0.1420
0.1490
0.1566
0.1649
0.1739
0.1838
0.1946
0.2064
0.0187
0.0209
0.0233
0.0259
0.0288
0.0319
0.0353
0.0390
0.0431
0.0475
0.0523
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0269
0.0301
0.0335
0.0373
0.0414
0.0459
0.0508
0.0561
0.0620
0.0683
0.0753
HIGH SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1244
0.1298
0.1356
0.1420
0.1490
0.1566
0.1649
0.1739
0.1838
0.1946
0.2064
0.0278
0.0311
0.0347
0.0385
0.0428
0.0474
0,0525
0.0580
0.0641
0.0707
0.0779
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0178
0.0199
0.0222
0.0246
0.0274
0.0303
0.0336
0.0371
0.0410
0.0452
0.0498
-------
1994 MHD LOW EMISSIONS SCENARIO
Percent Total Sulfates and
of Particulate SOF bound water Residual
Engines (g/BHP-hr) (g/BHP-hr) (g/BHP-hr) (g/BHP-hr)
LOW SOF CASE
0 .5800
2.1000
5 . 8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0 . 5800
0 .1147
0.1209
0.1278
0 . 1353
0 .1436
0.1526
0.1626
0 . 1736
0.1856
0.1989
0.2134
0.0155
0.0175
0 .0196
0.0219
0.0245
0.0273
0.0304
0.0338
0.0375
0 . 0416
0.0461
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0346
0.0389
0.0436
0.0488
0.0545
0.0608
0.0676
0.0752
0 . 0835
0.0926
0.1027
HIGH SOF CASE
0 . 5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1147
0.1209
0.1278
0.1353
0.1436
0.1526
0.1626
0.1736
0.1856
0.1989
0.2134
0.0256
0.0287
0.0322
0.0361
0.0403
0.0449
0.0500
0.0556
0.0617
0.0685
0.0759
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0 . 0646
0.0245
0.0276
0.0310
0.0346
0.0387
0.0431
0.0480
0.0534
0.0593
0.0658
0.0729
-------
1994 HHP LOW EMISSIONS SCENARIO
Percent Total Sulfates and
of Particulate SOF bound water Residual
Engines (g/BHP-hr) (g/BHP-hr) (g/BHP-hr) (g/BHP-hr)
LOW SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1097
0.1162
0.1233
0.1312
0.1399
0.1494
0.1599
0.1715
0.1842
0.1982
0.2136
0.0072
0.0081
0.0091
0 . 0102
0.0114
0.0128
0.0142
0.0158
0.0176
0.0196
0.0217
0.0583
0 .0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0 . 0442
0 . 0498
0.0559
0.0627
0.0701
0.0783
0.0874
0.0973
0.1083
0.1203
0.1336
HIGH SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1097
0.1162
0.1233
0.1312
0.1399
0.1494
0.1599
0.1715
0.1842
0.1982
0 .2136
0.0175
0 .0197
0.0221
0 . 0248
0.0277
0. 0310
0.0345
0.0385
0.0428
0.0476
0 . 0528
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0339
0.0382
0.0429
0 . 0481
0 .0538
0.0601
0.0671
0.0747
0.0831
0.0923
0.1025
-------
1994 BUS LOW EMISSIONS SCENARIO
Percent Total Sulfates and
of Particulate SOF bound water Residual
Engines (g/BHP-hr) (g/BHP-hr) (g/BHP-hr) (g/BHP-hr)
LOW SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2. 1000
0.5800
0. 1330
0.1368
0.1407
0.1449
0. 1492
0.1538
0.1585
0.1635
0.1688
0.1743
0.1800
0.0202
0.0214
0.0226
0.0239
0.0252
0.0266
0.0281
0.0297
0 . 0313
0.0330
0.0348
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0450
0 . 0476
0.0503
0.0532
0.0562
0.0593
0.0626
0.0661
0.0697
0.0735
0.0774
HIGH SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1330
0.1368
0.1407
0.1449
0.1492
0.1538
0.1585
0.1635
0.1688
0.1743
0.1800
0.0333
0.0352
0.0372
0.0393
0.0415
0.0438
0.0463
0.0488
0.0515
0.0543
0.0572
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0320
0,0338
0.0357
0 . 0378
0.0399
0.0421
0.0445
0.0469
0.0495
0.0522
0.0550
-------
Chapter 4
Effect of Fuel Quality on Emissions and Engine Cost
One of the major purposes of this study is to estimate
the impact a modification in diesel fuel quality will have
on engine technology requirements. In this chapter, the
engine-out emission estimates, aftertreatment technology
costs, and aftertreatment device efficiencies developed in
Chapter 3 will be used to estimate the cost of complying
with the particulate standards under various fuel control
scenarios. The difference between compliance costs with
baseline fuel quality (Chapter 3) and the compliance costs
with fuel quality control developed here will then be used
in Chapter 7 to evaluate the cost effectiveness of the fuel
quality regulation.
The two primary fuel control scenarios for which
exhaust aftertreatment technologies mixes were determined
are: a sulfur reduction to 0.05 weight percent, and a
sulfur reduction to 0.05 weight percent accompanied by an
aromatics reduction from the current average level of 34.2
to 20 volume percent. Hereafter, these scenarios will be
referred to simply as sulfur control and aromatics control,
respectively. Other degrees of fuel control will be
discussed in section IV of this chapter.
The first subject to be discussed in this chapter will
be the effect of fuel sulfur and aromatics on the emissions
of diesel vehicles. Second will be a presentation of the
aftertreatment requirements and costs for compliance under
the two fuel control scenarios. This will be followed by a
sensitivity analysis of the aftertreatment requirement
estimates to some of the key assumptions that were made
about emissions, aftertreatment efficiency, and
aftertreatment cost in Chapter 3.
I. Effect of Fuel Properties on Diesel Emissions
Over the past few years a number of studies have
investigated the effects of diesel fuel parameters on
diesel emissions, both for heavy-duty and light-duty
vehicles. On the heavy-duty side, some of the most recent
investigation into the effects of fuel quality on diesel
emissions is being performed under contract by Southwest
Research Institute under contract with Coordinating
Research Council.[1] Emission testing is being performed
on various heavy-duty engines (including a Cummins NTCC400,
and a DDA Series 60) on fuels of varying volatility,
aromaticity, and sulfur content. A number of other studies
of this nature have also been performed and published.
Also, many engine manufacturers have supplied EPA with
-------
4-2
emission test data, which, although confidential in nature,
have been helpful in confirming the effect of changing fuel
quality on emissions. The effect of diesel fuel qualities
on light-duty diesel emissions has also been the subject of
investigation.
The collective knowledge gained from these published
studies was used in predicting the effect of fuel
modifications on emissions of both heavy-duty and
light-duty diesel vehicles. The confidential submittals
were only used to improve the choice of a best estimate
from within a range of data where the scatter was such that
a simple averaging of the data was not appropriate.
A. Effect of Fuel Sulfur on Highway Diesel Emissions
l. Heavy-Duty Diesel Engines
The relationship between fuel sulfur and particulate
and sulfurous emissions has been the subject of study and
testing for the past several years. It has been
demonstrated that fuel sulfur contributes to the emissions
of diesel engines, both through gaseous emissions of sulfur
dioxide (SO2), and also through the further oxidation of
SO2 to sulfate particulate. The relative amounts of
sulfur dioxide and sulfate particulate matter emitted
depend on the engine design and operating conditions. Most
test data show that between one and three percent of the
sulfur in the fuel is emitted as particulate sulfate, while
the rest is emitted as SC>2- The determination of a
representative sulfur-to-sulfate conversion fraction from
the available test data is necessary for the purpose of
projecting vehicle emissions, and a discussion of the
methodology and data used in this determination are
presented below.
Some studies have also shown a correlation between
fuel sulfur content and non-sulfate particulates. This
effect appears to be small, and, in fact, has not been
observed in many experiments. If this correlation does
exist, it may be partially explained by the partial
saturation of polycyclic aromatic hydrocarbons in the fuel
during the desulfurization process. A more detailed
discussion of this issue was presented in Section III of
Chapter 2.
Table 4-1 documents the average results of transient
mode testing performed on various heavy-duty diesel
engines.[1-5] The data indicate that the conversion of
sulfur to sulfate varies significantly from engine to
engine. Predicting sulfur to sulfate conversion for a
given engine operating in a given mode is a difficult
matter and would require a more complete understanding of
-------
4-3
Table 4-1
Sulfate Conversion in HDDE Exhaust (Transient Cycle)
Average
Percent
Manufacturer
Enqine
Conversion
Reference
Cat
3406B
1.5
[2]
Curam
L10-270
2.2*
[3]
Curam
NTCC-40 0
2.8*
[3]
DDA
1.1
[4]
Nav
7.3IDI
1. 5
[5]
Nav
DTA466
1.6
[5]
Nav
DTA466
1.9
[5]
Cumm
NTCC-400
2 . 7
11 ]
Sales Weighted Percent Conversion = 1.9
Excluding low sulfur fuel results.
-------
4-4
the mechanism of sulfate formation in the engine than is
available. Thus, a sales fraction weighting of the
conversions seen in the various test engines was
performed. Projected sales figures were taken from engine
manufacturers' confidential estimates submitted to EPA.
The results of this sales weighting indicate an average
conversion of approximately 1.9 percent for heavy-duty
diesels.
However, if the results for one engine, the Navistar
7.3 liter IDI engine, were excluded from the data (due to
the fact that it is an IDI engine), the average conversion
for heavy-duty engines becomes 2.2 percent. Therefore, a
conversion of 2.0 percent was used hereafter. The
sensitivity of results of the study to the conversion
assumed was also investigated in Section IV below.
Studies have reported that a significant increase in
the average conversion of fuel sulfur to sulfate
particulate occurs when operating on a low sulfur
fuel.[1,3] A discussion of the data recording this
phenomenon and possible explanations for it are presented
below.
The Coordinating Research Council VE-1 project
currently underway at Southwest Research Institute had, at
the time this analysis was performed, tested only one
engine, a Cummins NTCC400, on a variety of fuels under both
steady-state and transient conditions.[1] The fuel matrix
for this test program consists of nine fuels, three with
low sulfur content (0.05 percent nominal), one at an
intermediate sulfur level (0.15 percent), and five high
sulfur fuels (0.30 percent). At nominal fuel sulfur levels
of 0.30 wt percent, fuel sulfur conversions ranging from
1.6 percent to 2.7 percent had been reported, with average
levels of 2.0 percent. It appears from the data available
that sulfur to sulfate conversion increases to nearly 5.0
percent on the low sulfur fuels.
Transient emission testing of two Cummins engines, a
1986 L10-270 and a 1986 NTCC-400, was performed by Cummins
Engines Company in cooperation with the Department of
Energy, Mines, and Resources of Canada.[3] Average sulfur
to sulfate conversions using normal engine timing and at
normal sulfur levels (e.g., 0.23 percent and 0.39 percent
by weight) were 2.2 percent for the L10 and 2.8 percent for
the NTCC-400. This test data showed a dramatic increase in
sulfur conversion (up to 18 percent) using a low sulfur
(0.02 percent by weight) fuel.
One possible reason for this observation is the
collection of background sulfate particulate in addition to
that produced by the engine. For instance, for a typical
heavy-duty engine operating on 0.05 weight percent sulfur
-------
4-5
fuel, one would predict sulfate emissions of 0.0057
g/BHP-hr (assuming 0.42 lb/BHP-hr fuel consumption, no
bound water, and 2.0 percent sulfur to sulfate
conversion). Whatever background sulfate emissions are
being measured in addition to the fuel sulfur emissions
would distort the percent conversion calculated from the
data. A study performed by John C. Wall of Cummins Engine
Company on a LTA10-300 engine using sulfur free fuel
(dodecane) showed measured background sulfate particulate
levels of 0.005 g/BHP-hr.[3] In the example above,
assuming this background particulate level was measured in
addition to the fuel related sulfate emissions during
testing, a total of 0.0107 g/BHP-hr of sulfate would be
collected, and the sulfur to sulfate conversion calculated
would be 3.7 percent instead of the true 2.0 percent
conversion. If 0.25 weight percent sulfur fuel were being
used the relative contribution of background sulfate would
be smaller, the apparent sulfur to sulfate conversion would
be calculated to be only 2.4 percent versus the true
conversion of 2.0 percent in the engine. Assuming this
background sulfate measurement exists in the test data
reported, this may account for some of the perceived
increase in conversion with low sulfur fuel apparent in the
literature.
Another explanation for the experimental observations
of a higher indicated conversion rate at low sulfur levels
may be the imprecision in experimental measurement. Since
percentage conversion is proportional to the ratio of
sulfates collected to fuel sulfur level, any error in the
measurement of either value will affect the calculated
conversion. An illustration of this, similar to that
presented by John C. Wall, follows.[3]
Due to the low fuel sulfur level being measured (0.05
weight percent) any error in measurement might produce a
drastic change in calculated conversion. For example, a
fuel with a measured level of 0.05 weight percent may have
an error in sulfur measurement of +0.02 weight percent.
Assuming the actual percent conversion of sulfur to sulfate
is 2.0 percent, if the actual level of fuel sulfur were
0.03 percent (while being measured as 0.05 percent), the
sulfate (dry) collected would be 0.0034 g/BHP-hr. Assuming
a +0.003 g/BHP-hr sulfate measurement error (i.e., sulfate
measured = 0.0004 to 0.0064 g/BHP-hr), the calculated
conversion would range from 0.1 percent to 2.3 percent.
Conversely, if the actual fuel sulfur level were 0.0 7
percent (measured as 0.05 percent), calculated conversion
would range from 1.7 percent to 3.9 percent. For an actual
sulfur to sulfate conversion of 2.5 percent, experimentally
measured conversions could range any where from o.l to 3.9
percent.
-------
4-6
If the effects of background sulfate emissions
described earlier (0.005 g/BHP-hr) are included, the
situation gets worse. For the same situation as above,
calculated conversions could range from 1.9 to 5.6
percent. Of course these errors can be minimized by
repeated experimental measurement (including background
sulfate measurement) in a variety of laboratories.
However, the possibility for this type of error does exist
and, assuming a normal distribution of measurement errors,
it is not surprising that some studies have reported an
apparent increase in calculated sulfate conversion.
Furthermore, data also exist indicating the reverse effect
(i.e., that conversion is lower at low fuel sulfur
levels).[4]
To summarize, at this point in time it has not been
conclusively demonstrated that sulfate percent conversion
increases at low fuel sulfur levels. EPA's conclusion is
that the few reports documenting this effect merely point
to the variability in the data and the error inherent in
measurement. Even if conversion increases with low sulfur
fuel in some engines, it certainly does not appear to do so
in all engines, and therefore this study will assume that
the sulfur to sulfate conversion is 2.0 percent at all fuel
sulfur levels.
2. Light-Duty Diesel Engines
Although the vast majority of highway diesel fuel is
consumed by engines classified as heavy-duty, any
modification to diesel fuel quality will have an effect on
emissions from light-duty diesel trucks (LDDT) and diesel
passenger cars (LDDV) as well. To determine the magnitude
of this effect the existing base of emissions test data
from light-duty vehicles operating on fuels of varying
sulfur, aromatics, and volatility was examined.
As described in Chapter 3, light-duty diesel engines
typically convert approximately 1.5 percent of the sulfur
in the fuel to sulfate particulate. Using the mean level
of 1.5 percent, and assuming that the ratio of associated
water to sulfate particulate is 1.32 (50 percent humidity),
the effect of fuel desulfurization on particulate and SO2
emissions can be readily calculated for LDDV and LDDT from
average fuel economy data. Results of such calculations
will be shown in Section III below.
B. Effect of Fuel Aromatics on Highway Diesel
Emissions
1. Heavy-Duty Diesel Engines
Many of the aforementioned studies were also
instrumental in determining the effects of fuel aromatics
-------
4-7
content on emissions. Fuel aromatics have been correlated
with particulate emissions, particularly the soluble
organic fraction (SOF) and the residual carbonaceous
portion (RCP), as well as with gaseous HC emissions. A
discussion of the data as well as some of the confounding
factors involved in interpreting it is presented below.
First, a number of different methods exist for
measuring fuel aromatics. ASTM test D-1319, which is
commonly used in industry, utilizes what is known as
Fluorescent Indicator Absorption (FIA). This test
procedure measures volume percentages of saturates,
olefins, and aromatics in petroleum samples. Those
compounds measured as aromatics include monocyclic and
polycyclic aromatics, aromatic olefins, some dienes, and
compounds containing sulfur, nitrogen, and oxygen atoms.
Unfortunately, compounds which are entirely aromatic, those
which are aromatic with an olefin sidechain, and any
sulfur, nitrogen, or oxygen compounds are all counted
equally as aromatic species. The contribution of each of
these types of compounds to emissions, however, is not
likely the same. Two fuels with identical FIA aromatics
measurement may thus have radically different aromatic
compound types, and consequently will have different
emission forming characteristics,
Furthermore, the FIA measurement technique is
recommended only for petroleum fractions that distill below
600°F, at which temperature only about 90 percent of diesel
fuel distills. As stated in the ASTM test procedure,
results of the test are erratic on petroleum fractions
boiling near the 600°F limit. Because of these problems,
there is some question over the appropriateness of
correlating particulate emissions with FIA aromatics.
Other test procedures for measuring aromatic species
exist. Mass Spectroscopy, and Proton and Carbon-13 Nuclear
Magnetic Resonance (NMR) all quantify aromatic levels in
fuel. These methods make distinctions between aromatic
species type, whether monocyclic, dicyclic, or tricyclic.
The resulting measurement is therefore more representative
of the amount of aromatic carbon atoms in the fuel.
Unfortunately, these aromatic measurement procedures are
not currently used widely, and most available studies of
emissions versus aromatic level do not contain measurements
of aromatics by Mass Spectroscopy or NMR method, only FIA.
Further, little data are available other than FIA aromatics
describing the speciation of aromatic compounds in
commercial fuel.
Because of the large body of data based on the FIA
measurement procedure, an attempt to correlate HC and
particulate emissions with FIA aromatics levels will be
-------
4-8
made. However, NMR and Mass Spectroscopy have been used to
analyze the aromatics content of the fuels used in a single
study, the CRC's VE-1 heavy-duty Emissions testing
project. [1] The emission data taken from the one engine
tested there (Cummins NTCC400) will be used to correlate
emissions with NMR and/or Mass Spectroscopy aromatics
levels. A description of this correlation will be
presented later in this section.
a. Correlation of Emissions with FIA Aromatics
As described above, SOF, RCP, and HC (gaseous)
emissions have been observed to correlate with fuel
aromatics. Preliminary data from the VE-1 project on the
Cummins NTCC400 show that a correlation exists between
these variables and FIA aromatics.[1] Regressions of
emission data taken from this engine show that a reduction
in fuel aromatics from 34.2 to 20 volume percent would
result in SOF and RCP reductions of 14.4 and 9.6 percent
from baseline, respectively.
The effect of aromatics on hydrocarbon emissions
appears from this data to be similar to the effects of
aromatics on carbonaceous particulate. From inspection of
the data, it appears that control of fuel aromatics would
result in HC emission reductions of 9.6 percent from
baseline, respectively.
The Mobil/Caterpillar cooperative study also
investigated the effects of aromatics on particulate
emissions, albeit on an older Caterpillar 3406B engine.[2]
Regressing SOF emissions against FIA aromatics (transient
test results only), it appears that a one percent reduction
in fuel aromatics corresponds to a 2.5 percent reduction in
SOF emissions. Similarly, it appears that a one percent
reduction in aromatics will result in a 2.4 percent
reduction in RCP. HC emissions were also correlated with
fuel aromatics. On this engine, the reduction of HC
associated with aromatics control was 1.1 percent per
percent aromatics reduction in the fuel. This would
translate into SOF reductions of 35.5 percent, RCP
reductions of 34.1 percent, and HC reductions of 16.1
percent for an aromatics reduction to 20 volume percent
from baseline levels, significantly higher than the
reductions achieved on the VE-1 Cummins engine. Since the
results of this engine represent older technology than the
VE-l data, the VE-1 data will be used preferentially in
this analysis. However, the results of the Mobil
Caterpillar study were used as a sensitivity analysis, a
full discussion of which is reserved for Chapter 7.
-------
4-9
Correlation of emissions with FIA aromatics can also
be performed with the data of Shirish Shimpi.[3] Only
three fuels were used, and the correlation of emissions
with aromatics levels was not very clear. An overall
downward trend in non-sulfate emissions with fuel aromatics
reduction was observed, however. Several other studies
have been performed investigating the effects of fuel
aromatics on diesel emissions as well. However, these
studies were performed predominantly on older technology
engines operating under steady state conditions. Although
somewhat useful in confirming the results of the studies
described above, the data contained in them were not used
directly in this analysis.
This analysis assumes that fuel aromatics control will
result in straight percentage reductions on the various
emission components of heavy-duty diesel exhaust,
regardless of the absolute emission level of the engine.
However, not all of the SOF and RCP particulate, nor all HC
emissions, can be attributed to aromatics in the fuel.
Some of the emissions have been shown to be derived from
other sources (e.g., lubricating oil, non-aromatic fuel
components, etc.). For current engines, the percent
reductions in emissions estimated above should implicitly
take into account these other sources. More uncertainties
arise, however, when this approach is applied to future
engines.
Many manufacturers have pointed to improved engine oil
control in future engines as a means of reducing lube-oil
derived emissions. Were oil-derived emissions reduced
disproportionately, the percentage reduction in emission
corresponding to fuel aromatics control would be greater
since the fractional amount of emissions derived from fuel
aromatics would be greater than in current engines.
Therefore, emission reduction estimates based on current
engines and applied to future engines may underestimate
actual emission reductions. However, the opposite could
also be true if the reduction of fuel-based RCP emissions
reductions outpaced those of oil-related emissions on
future engines. Unfortunately, either situation could
occur. This is another reason for using the range of SOF
levels for future engine emissions described in Chapter 3.
Therefore, it appears reasonable to apply percent
reductions in emissions to both current and future engines.
b. Correlation of Emissions with Aromatic Carbon
(Mass Spectroscopy, NMR Analysis)
As mentioned in Section I-A-l of this chapter, some of
the studies investigating fuel effects on emissions have
observed that there may be a direct correlation between
-------
4-10
fuel sulfur levels and the amount of SOF emissions
collected.[1,6] Unfortunately, the effect has only been
seen with a few engines, while on the majority of engines
tested it has not been demonstrated. It has been
hypothesized that the effect may be attributable to the
"scrubbing" of hydrocarbon compounds out of the gas phase
by reaction with sulfuric acid on the particulate filter,
or that the cause may be improved filtration efficiency due
to the presence of sulfuric acid and associated water.[6]
It may also be the case that the correlation between fuel
sulfur and SOF emissions is attributable to the
desulfurization process rather than the fuel sulfur level
per se. Diesel fuel desulfurization via hydrotreating
often results in the partial saturation of polycyclic
aromatic compounds (see Chapter 2, Section III). The
elimination of some of these polycyclic aromatics in the
fuel may be the cause of lower SOF emission levels. Using
the information presented in Chapter 2 on the effect of
hydrotreating on aromatic species and by determining the
effects of various aromatic species on emissions (as will
be done this section), it is possible to predict reductions
in SOF and other carbonaceous emissions resulting from fuel
desulfurization.
It has been hypothesized that different aromatic
species may affect emissions from diesel vehicles to a
different extent. More specifically, diesel engine
emissions may correlate more closely with aromatic carbon
content of the fuel (as determined by Mass Spectroscopy,
NMR) than with the volume fraction of aromatic containing
compounds (FIA analysis). Unfortunately, little emission
and fuel composition data exist in which fuel aromatics are
measured by methods other than FIA. Further, little is
known about the speciation of aromatics compounds in
commercial fuel. Still, the fuels used in the CRC VE-1
emissions study have been analyzed thoroughly by these
various methods, and by making some assumptions about the
distribution of aromatic compounds in commercial diesel
fuel, it was possible to predict here the effect of fuel
control on SOF, RCP, and HC emissions.[1]
Fuel aromatics analysis by Mass Spectroscopy is
performed by first obtaining the aromatic fraction of the
fuel by chromatographic separation. Mass Spectroscopy (MS)
is then performed to separate the aromatics by class and
aromatics composition data are derived in volume percent.
Determination of aromatics content of the fuels used in the
CRC VE-1 project by MS was performed by Chevron Research
Company and Unocal Corporation. Determination of aromatic
carbon by Proton and Carbon-13 Nuclear Magnetic Resonance
Spectrometry (NMR) was performed on the same fuels by
General Motors Research Laboratories. Results of the NMR
-------
4-11
analysis are reported as the mole percent aromatic carbon
of the fuel.
The emission data from the VE-i project were
correlated with fuel aromatics levels determined using both
MS and NMR. A "normalized MS value" was first determined
for each fuel by combining the volume percent of mono-,
di-, and tricyclic structures, weighted according to the
relative amount of aromatic carbon in each type of
structure. Correlation of Emissions were then correlated
with this "normalized MS value" as they were with the mole
fraction aromatic carbon in the fuel, as determined by the
carbon-13 NMR method.
In order to use these correlations to estimate
emission reductions resulting from fuel control, it was
necessary to make some assumptions concerning the
speciation of aromatics in commercial fuel both currently
and under the fuel control scenarios being considered.
Little data exist on the distribution of aromatics in
commercial fuel. Great variability may exist from refinery
to refinery as well as among streams within a refinery,
depending on the processing each has been subjected to.
However, most distillate fuels on which data were reported
show that the majority of aromatic species are monocyclic
structures. Approximately 10 to 50 percent of the
aromatics are dicyclic structures, while typically 5 or
less percent are polycyclic (tricyclic or greater)
structures.[1,7]
As a basis for this analysis, it was assumed that
baseline fuel (FIA aromatics of 34.2 percent) has an
aromatics distribution of mono-, di-, and tricyclic
aromatics of 26.4, 6.8, and 1.0 volume percent. While
selecting any distribution of aromatic species to represent
commercial fuels is highly subjective based on the limited
data available, a baseline assumption had to be made. This
distribution was selected assuming 20 percent of aromatics
were dicyclic, 3 percent were tricyclic, and the remaining
77 percent were monocyclic. This was perceived to be
fairly representative of the distribution of aromatics in
the mid-range volatility/ mid-range aromatic fuel used in
the VE-1 test program (fuel #5) and of the distillate fuels
tested by Yoes and Asim.[7] While a baseline fuel
containing up to 50 percent dicyclic aromatics could have
reasonably been selected, selecting such a fuel as a
baseline would have changed the ensuing particulate
emission reduction estimates by only a few percent.
Therefore, the sensitivity of emission reduction
projections to the distribution of aromatics in baseline
fuel is low and can be ignored.
-------
4-12
As described in Chapter 2, a 50 percent reduction in
di- and tricyclic structures with a corresponding increase
in monocyclic structures should occur with sulfur control
(i.e., when the fuel undergoes hydrodesulfurization). This
would yield an aromatic distribution of 30.3, 3.4, and 0.5
volume percent. Additional processing necessary to reduce
aromatics to 20 volume percent would further shift the
distribution of mono-, di-, and tricyclic structures to
18.5, 1.4, and 0.1 percent, respectively.
Using these distributions of aromatics, and the
correlations of emissions with the "normalized MS value"
developed from the VE-1 project data, emission reductions
can be estimated. From this analysis, it appears that
small reductions of 2.8, 1.1, and 1.9 percent in emissions
of SOF, RCP, and HC would result from fuel sulfur control.
Reductions of 18.5, 6.8, and 10.8 percent in SOF, RCP, and
HC emissions were predicted to result from control of
sulfur in conjunction with aromatics control to 20 percent.
These emission reduction estimates are fairly
consistent with those projected according to the FIA
analysis for the Cummins engine (reductions of 14.4, 9.6,
and 9.6 percent in SOF, RCP, and HC with aromatics control
to 20 percent), and more accurately reflect the fact that
some degree of partial fuel aromatics saturation takes
place with fuel hydrodesulfurization. Results of the
"normalized MS value" correlation were therefore used to
predict reductions in emissions corresponding to fuel
control in the remainder of this report. However, the data
with which to construct reliable models correlating
emissions with more sophisticated aromatics measurement
procedures such as Mass Spectroscopy are limited and so the
analysis presented here should be considered preliminary.
As more data are generated, more sophisticated emissions
modeling of this nature should become feasible, thus
improving the ability to precisely predict the relationship
between fuel aromatics and emissions,
2 . Light-Duty Diesel Engines
Due to the differences between IDI (indirect
injection) and DI (direct injection) diesel engines and the
different duty cycles of cars and large trucks, it is
generally impossible to use data generated on heavy-duty
engines (mostly DI) to predict fuel effects on emissions
from light-duty IDI engines. Given this limitation, in
order to determine the effect of fuel control on light-duty
diesel emissions, one must turn to the data generated on
light-duty diesel vehicles.
-------
4-13
One of the most recent studies which investigated the
effects of fuel variables on emissions was CRC Project
CAPE-32-80, performed under contract by the Southwest
Research Institute.[8] This test program consisted of the
measurement of FTP HC, CO, NOx, and particulate emissions
from four 1982 MY light-duty diesel vehicles operating on
nine different fuels of varying aromaticity, volatility,
and sulfur content. The fuel matrix was selected such that
the effect of aromatics, Tgo* and T^q on HCf C0' anc^
NOx emissions could be evaluated independently. The
amount of sulfur in the fuel, however, varied in
conjunction with the other three fuel parameters and so the
effect of these three parameters on particulate emission is
somewhat confounded with the effects of fuel sulfur on
particulates. By assuming that 1.5 percent of the fuel
sulfur is converted to particulate sulfate (see Chapter 3),
and that 1.32 grams of water is associated with each gram
of sulfate particulate, one can normalize the emissions
results of this study to remove the effect of sulfur from
the particulate emissions data. This was done by
subtracting the assumed weight of the sulfate and bound
waters from the total particulate levels.
After doing this, these adjusted particulate levels
were then regressed against fuel aromatics (FIA), Tgg,
and T^q . From the regression it was determined that for
the four vehicles tested, a reduction in fuel aromatics
from 34.2 to 20 percent would result in a non-sulfate
particulate reduction of approximately 7.2 percent on
average.
Inspection of the CAPE-32-80 test data also revealed
that a decrease in HC emissions of about 50 percent would
result from dearomatization from 34.2 to 20 volume
percent. Data from this testing program also showed some
possible effect of aromatics on NOx emissions, but the
supporting evidence is not sufficient to make any
quantitative statement.
A recent cooperative study by Mobil Research Dev.
Corp. and Toyota Motor Corp. also investigated the effects
of fuel properties on FTP emissions from a light-duty (2.2
liter) Toyota truck engine.[9] Eight different fuels were
tested; two containing 20 percent aromatics, five with
aromatics levels ranging from 32 to 37 percent, and one
containing 49 percent aromatics. A curvilinear regression
of test results showed an average reduction of about 43
percent in total particulate emissions corresponding to a
reduction in aromatics from 34.2 to 20 volume percent.
This corresponds to a reduction of approximately 50 percent
in RCP and SOF, assuming that the sulfate emissions
remained constant. A reduction of approximately 60 percent
in HC emissions was also observed.
-------
4-14
The results of these two studies were combined
according to the number of vehicles tested in each (i.e., a
four to one vehicle weighting) in order to estimate an
average effect of fuel aromatics control on particulate
emissions for this analysis. Combining the results of the
two studies in this way, one would estimate that a
reduction in aromatics to 20 volume percent would reduce HC
emissions by about 52 percent from baseline, while
non-sulfate particulate would be reduced by roughly 15
percent.
Due to the lack of data and the relatively small
contribution of light-duty diesels to the total diesel
particulate emission inventory, it was assumed that fuel
sulfur control (and the corresponding partial saturation of
polycyclic aromatic structures) would have no effect on LDD
HC or SOF emissions.
II. Effect of Fuel Quality on Off Highway Diesel and No .2
Fuel Oil Emissions
The diesel fuel sulfur and sulfur/aromatics controls
studied in this paper are intended for on-highway diesel
fuel only. However, as explained in Chapter 2, refiners
market on-highway, off-highway and a significant amount of
fuel oil from a common distillate pool. With fuel
controls, they are faced with the ";cision to treat the
whole pool or treat part of the p^ .1 and segregate the
products. It is expected that some refiners will treat the
whole pool, and thus at least some of the fuel oil marketed
will have a low sulfur content, or a low sulfur and
aromatics content. This will result in emission reductions
for those sources that use this fuel oil. The purpose of
this section is to explain the changes in emissions.
Total sulfur oxide emissions are almost entirely
dependent on the sulfur content of the fuel and are
generally not affected by boiler size or burner design. On
the average, more than 95 percent of the fuel sulfur is
emitted as SO2 (like internal combustion engines), while
the rest is emitted as sulfate particulate and sulfur
trioxide, which quickly reacts with moisture to become
additional sulfate particulate. Due to the similarity
between the sulfur conversion for these applications and
diesel engines, identical sulfur to sulfate conversion will
be used for these stationary fuel oil sources, and engines
burning off-highway No. 2 diesel.
Non-sulfate particulate emission from off-highway
sources were assumed to remain constant with fuel aromatics
control. As indicated in the Mobil/Caterpillar study,
emissions from engines operating under steady-state
conditions show little sensitivity to fuel aromatics
-------
4-15
levels.[2] Since engines used in off-highway applications
operate primarily in a steady-state mode, it was assumed
that fuel aromatics control would have no effect on
off-highway particulate emissions.
While it can be argued that such an assumption tends
to underestimate the possible emission benefits of fuel
aromatics control, results of the cost-effectiveness
analysis (Chapter 7) indicate that this assumption is not
critical. Even if aromatics control (under the NPRA
segregation scenario) resulted in the same emission
reductions from off-highway as from on-highway diesels, the
cost-effectiveness ($/ton of urban particulate) of
aromatics control estimated under the NPRA segregation
scenario would be no lower than that derived in the 100
percent segregation case. Since, as will be seen in
Chapter 7, aromatics control does not result in cost
effective emission control under either segregation
scenario, this simplifying assumption is somewhat
perfunctory.
Ill. Diesel Aftertreatment Technology With Sulfur Control
The aftertreatment technology mix for compliance with
1991 and 1994 particulate standards was determined assuming
fuel sulfur was reduced to 0.05 weight percent. The
relationships between fuel sulfur control and diesel engine
emissions established in Section I of this chapter were
used to adjust the engine-out emission distributions
developed in Chapter 3. The adjusted engine-out
distributions were then used in estimating the most cost
effective mix of aftertreatment technologies necessary for
compliance with the particulate standards. Results of the
analysis are presented below.
A. Light-Duty Diesels
Emissions for 1991-1993 LDDs with fuel sulfur control
are shown in Table 4-2. Because vehicle technology is
unaffected, LDDV and LDDT1 sulfate and SO2 emissions were
simply reduced from baseline levels (Chapter 3) to reflect
the degree of fuel sulfur control. The fuel sulfur
reduction also caused a reduction in sulfate particulate
emitted by LDDT2s. This did not show up as a reduction in
total particulate, however, but rather as a reduction in
the number of traps required to meet the 0.13 gpm
particulate standard (86.7 percent). Table 4-2 shows
emission levels and aftertreatment costs for these
vehicles. Emission reductions and aftertreatment cost
savings from baseline levels are shown in the Table 4-3.
-------
4-16
Table 4-2
1991 Light-Duty Diesel Af tertreatment Technologies,
Costs and Emissions Under Various Fuel Controls
Vehicle
Class
Percent
Traps
Aftertreat
Cost
($/veh)
Fuel
Economy
(mpq)
Emissions (q/mi)
SOF
RCP
SO*
TPM
SO:
HC*
Fuel Sulfur Control
1991-93 Model Years
LDDV
LDDT1
LDDT2
0
0
86.7
0
0
394
31.4
25.9
25.5
.024
.108
.055
,079
,072
,027
,006
.007
,007
,109
.187
,089
104
,126
,128
. 29
.43
188
1994 and Later Model Years
LDDV
LDDT1
LDDT2
86
0
0
393
34.2
27.8
27.3
,024
.108
.056
.079
.072
.027
,005
,006
,006
, 108
,186
,089
,095
,118
.120
.29
.43
.19
Subsequent Fuel Aromatics Control
1991-93 Model Years
LDDV
LDDT1
LDDT2
0
0
79.6
0
0
362
31.4
25.9
25.5
,020
,092
,053
.067
.061
.029
.006
.007
.007
.093
. 160
.089
104
118
,128
, 139
,206
.072
1994 and Later Model Years
LDDV
LDDT1
LDDT2
0
0
79.3
0
0
361
34.2
27.8
27.4
,020
,092
,053
.067
,061
,030
,005
.006
,006
,092
.159
,089
.095
. 118
.119
,139
.206
,072
a
Zero mile emissions.
-------
4-17
Table 4-3
Light-Duty Savings and Emission
Reductions Corresponding to Fuel Control
Aftertreat. Total Direct
Vehicle Cost Savings Particulate Emissions
Class ($/veh) Reduction (g/mi)
Fuel Sulfur Control
1991-93 Model Years
LDDV 0 0.022
LDDT1 0 0.026
LDDT2 61 0.000
1994 and Later Model Years
LDDV 0 0.020
LDDT1 0 0.025
LDDT2 57 0.000
Subsequent Fuel Aromatics Control*
1991-93 Model Years
LDDV 0 0.016
LDDT1 0 0.027
LDDT2 32 0.000
1994 and Later Model Years
LDDV 0 0.016
LDDT1 0 0.027
LDDT2 32 0.000
Incremental to sulfur control.
-------
4-18
Table 4-2 also shows aftertreatment requirements and
emissions for 1994 and beyond. Once again, LDDV and LDDTl
emissions of sulfate and SO2 were reduced from baseline
fuel levels and trap requirements for LDDT2 (86.5 percent)
were also reduced. Emission reduction and aftertreatment
cost savings are shown in Table 4-3.
2. Heavy-Duty Diesels
The technology mix, cost, and emissions for 1991-3
HDDE's with sulfur control, based on the average of the two
nominal engine-out emission distributions developed in
Chapter 3, are shown in Table 4-4. As can be seen, with
sulfur control it is projected that no traps will be needed
to meet the 0.25 g/BHP-hr particulate standard. While
traps are projected to be needed to meet the standard
without fuel sulfur control, it is possible that trap
technology will not be available in time. If this were the
case, then the cost savings shown can be used as a very
rough indication of the reduction in non-conformance
penalties which would occur. Urban buses will require a
mixture of catalyzed traps and flow-through catalysts to
meet the 0.10 g/BHP-hr standard. Aftertreatment cost
savings of $156 per urban bus sold, not including the cost
of a fuel economy penalty due to traps would be realized.
Aftertreatment cost savings and particulate emission
reductions for each heavy-duty class are shown in Table 4-5.
Table 4-4 also shows the projected exhaust
aftertreatment requirements and emissions for 1994 and
later heavy-duty diesels under fuel sulfur control. As
shown, aftertreatment costs savings range from $268 to $335
per vehicle as shown in Table 4-5. The utilization of high
activity flow through catalysts was increased from the
baseline fuel case due to the reduction in fuel sulfur,
with utilization rates as high as 55 percent of vehicles
sold in the case of LHDDEs. The use of some traps
(approximately 30 percent of vehicles sold), however, will
still be required to meet the 1994 particulate standard.
IV. Diesel Aftertreatment Technology Fuel Aromatics Control
1. Light-Duty Diesels
Projected emissions and exhaust aftertreatment
technology requirements for light-duty diesels under
subsequent fuel control are shown in Table 4-2. As shown
in Table 4-3, emissions of particulate are slightly reduced
from those with only fuel sulfur control for LDDVs and
LDDTls. For LDDT2s, with fuel aromatics control only 79.6
and 79.3 percent of the vehicles in 1991 and 1994,
respectively, will require traps to comply with the 0.13
gpm particulate standard.
-------
Table 4-4
Heavy-Duty Diesel Aftertreatment Technology, Cost and Emissions Under Various Fuel Controls
Technology Usage (%)
Low Cost
High Cost
Average
Low Cost
Flow
Flow
Af tertreat.
Engine
Non Cat
Catalyzed
Through
Through
Cost
End
of Life
Emissions
(g/BHP-hr)
Class
Traps
Traps
Catalysts
Catalysts
($/vehicle)
SOF
RCP
S04
TPM
so?
HC
CO
Fuel Sulfur Control
1991-93
Model Years
LHDDE
0
0
0
0
0
.108
.101
.016
.220
.231
1.12
3.97
MHDDE
0
0
0
0
0
.085
.124
.013
.221
. 187
1.19
4.35
HHDDE
0
0
0
0
0
.050
.160
.012
.222
. 173
1.30
4.87
BUS
0
67.8
0
20.5
$500
.034
.044
.016
.094
.195
.56
3.88
1994 and
i Later Model
Years
LHD
0
29.5
0
54.8
193.6
.028
.042
.023
.093
.221
.614
2.47
MHD
0
36.0
0
30.3
307.4
.029
.049
.016
.093
.183
.745
3.49
HHD
20.7
15.8
0
29.0
367.9
.017
.062
.014
.093
.165
.950
3.96
BUS
0
34.4
0
50.4
343.9
.024
.050
.019
.093
.191
.645
2.86
Subseguent Fuel Aromatics Control
1991-93
Model Years
LHDDE
0
0
0
0
0
.086
.095
.016
.197
.231
1.02
3.9
MHDDE
0
0
0
0
0
.071
.116
.013
.200
.187
1.08
4.3
HHDDE
0
0
0
0
0
.042
.151
.012
.204
.173
1.18
4.9
BUS
0
61.4
0
19.7
453
.031
.048
.016
.094
.197
.56
3.9
1994 and
Later Model
Years
LHDDE
0
24.7
0
50.1
167
.028
.044
.022
.093
.221
.61
2.6
MHDDE
0
27.0
0
38.4
265
.024
.052
.017
.093
.182
.70
3.2
HHDDE
18.3
13.7
0
26.1
327
.015
.064
.014
.093
.165
.90
4.0
BUS
0
28.6
0
39.7
284
.024
.052
.018
.093
.191
.68
3.2
-------
4-20
Table 4-5
Heavy-Duty Vehicle Savings and Emission
Reductions Corresponding to Fuel Control
Vehicle
Class
Aftertreat.
Cost Savings
($/veh)
Total Direct
Particulate Emissions
Reduction (g/BHP-hr)
Fuel Sulfur Control
1991-93 Model Years
LHDDE
MHDDE
HHDDE
BUS
157
192
205
156
1994 and Later Model Years
LHDDE
MHDDE
HHDDE
BUS
268
335
334
315
Subsequent Fuel Aromatics Control*
1991-93 Model Years
LHDDE
MHDDE
HHDDE
BUS
0
0
0
47
1994 and Later Model Years
LHDDE
MHDDE
HHDDE
BUS
27
42
41
65
0 .004
0 .003
0.002
0.000
0.000
0.000
0.000
0.000
0.027
0.024
0 . 020
0 . 000
0.000
0.000
0.000
0 . 000
Incremental to sulfur control.
-------
4-21
2. Heavy-Duty Diesels
Table 4-4 shows the projected emissions and
aftertreatment requirements for heavy-duty diesels under
subsequent fuel aromatics control. As seen in Table 4-4,
and as summarized in Table 4-5, emission estimates of SOF,
RCP, and HC for 1991 model year HHDDEs are slightly lower
than with only fuel sulfur control, due to the reduction in
fuel aromatics. Urban buses in 1991 and all HDDDEs in 1994
will still require the use of trap oxidizers under this
fuel control scenario, although slightly fewer than with
only fuel sulfur control. Aftertreatment costs savings in
1994 range from $27 to $65 per vehicle, not including the
cost of the added fuel economy penalty due to traps.
V. Sensitivity Analysis
The cost effectiveness of fuel quality control will be
determined in Chapter 7. The emissions and technology
results presented thus far will be used there as the "best
estimate" of the engine manufacturing industry response to
the fuel control scenario under consideration and will be
used primarily in that analysis. However, several of the
assumptions used in this analysis may have a significant
impact on the vehicle savings and emission reductions
associated with fuel control. These include: the percent
conversion of sulfur to sulfate in the engine, baseline
engine-out emission levels, and the cost of the trap
oxidizer systems. To evaluate the impact of these
parameters, two sensitivity cases were run on 1994 HHD
engines, a "greatest savings" and a "lowest savings" cost.
The impact of these parameters on results for the other
vehicle class was assumed to be proportionally similar.
In the "greatest savings" case, the assumptions made
regarding sulfur to sulfate conversion, engine-out
emissions, and trap-oxidizer costs were selected in order
to maximize the per-vehicle savings attributable to fuel
control. A sulfur to sulfate conversion of 2.5 percent in
the engine was assumed, as were the "low emission"
engine-out distributions developed in Chapter 3. It was
assumed that trap costs were the same as in the "best
estimate" scenario (i.e., no low-cost traps were
available). Vehicle savings and emission reductions
corresponding to fuel control for the HHDDE vehicle class
in the "greatest savings" case are shown in Table 4-6.
In the "lowest savings" case, modeling assumptions
were selected in order to minimize the vehicle savings
resulting from fuel control. A sulfur to sulfate
conversion of 1.0 percent was used, as were the "low-cost"
trap cost estimates (Table 3-8). Results for the
"lowest-savings" scenario are also shown in Table 4-6.
-------
4-22
Table 4-6
Sensitivity of Heavy-Duty 1994 and Later Model Year
Diesel Engine Savings and Emission
Reductions Corresponding to Fuel Control
Scenario
Vehicle
Class
Aftertreatment
Cost Savings
($/veh)
Total
Particulate
Emission
Reduction
(g/BHP-hr)
Fuel Sulfur Control
Best Estimate
HHD
$334
0
Highest Savings
HHD
$552
0
Lowest Savings
HHD
$131
0
Subsequent Fuel
Aromatics
Control*
Best Estimate
HHD
$ 41
0
Highest Savings
HHD
$ 31
0
Lowest Savings
HHD
$ 36
0
Incremental to fuel sulfur control.
-------
4-23
Inspecting Table 4-6, one sees that, depending on the
assumptions used, the impact of fuel control on engine
technology can vary greatly. Aftertreatment cost savings
range from $131 to $552 per HHDDE for sulfur control (not
including fuel economy benefits) around the mean of $334
per vehicle. Cost savings for subsequent aromatics control
range from $31 to $36 per vehicle around a mean of $41 per
vehicle. These values will be used in the overall
sensitivity analysis in Chapter 7.
-------
4-24
References (Chapter 4)
1. "Investigation of the Effects of Fuel
Composition and Injection and Combustion System Type on
Heavy-Duty Diesel Exhaust Emissions," Terry L. Ullman,
Southwest Research Institute, Coordinating Research Council
Contract CAPE-32-80, Project VE-1, March 1989.
2. "Heavy-Duty Diesel Engine Fuel Combustion
Performance and Emissions - A Cooperative Research
Program," E.G. Barry and L.J. McCabe, Mobil Research and
Development Corp., D.H. Gerke and J.M. Perez, Caterpillar
Tractor Co., SAE Paper 852078, 1985.
3. "Fuel Sulfur Reduction for Control of Diesel
Particulate Emissions," J.C. Wall, S.A. Shimpi, M.L. Yu,
SAE Paper no. 872139, 1987.
4. "General Motors Comments on Environmental
Protection Agency Notice Federal Register Vol. 51, No. 124,
June 27, 1986. Diesel Fuel Quality Effects on Emissions,
Durability, Performance and Costs; Availability of a Draft
Study," Docket No.A-86-03, December 22, 1987.
5. "Fuel Sulfur Effect on Diesel Particulate
Emission," Navistar International Corporation, Engine
Division Engineering, June 29, 1987.
6. "Fuel Composition Effects on Heavy-Duty Diesel
Particulate Emissions," J.C. Wall, and S. K. Hoekman,
Chevron Research Company, SAE Paper 841364, October, 1984.
7. "Confronting New Challenges in Distillate
Hydrotreating," Jack R. Yoes, Mehmet Y. Asim, Akzo Chemie
America, presented at 1987 NPRA Annual Meeting, San
Antonio, Texas, March 29-31, 1987.
8. "Study of the Effects of Fuel Composition, and
Injection and Combustion System Type and Adjustment, on
Exhaust Emissions from Light-Duty Diesels," Charles T.
Hare, Southwest Research Institute, Prepared for
Coordinating Research Council Inc., Project CAPE-32-80,
April 1985.
9. "Effects of Fuel Properties and Engine Design
Features on the Performance of a Light-Duty Diesel Truck -
A Cooperative Study," Barry E.G., Anelrod, J.C., and
McCabe, L.J., Mobile Res. & Dev. Corp., Inove, T. , and
Tsuboi N., Toyota Motor Corp., SAE Paper no. 861526.
-------
Chapter 5
Effect of Fuel Modification on Engine Wear
I. Introduction
This chapter explores the relationship between diesel fuel
sulfur level and engine wear, and estimates the reductions in
certain operating costs that could be experienced by truck
owners and operators using low sulfur fuel. The results are
used in Chapter 7 to estimate the net cost effectiveness of
sulfur controls.
This chapter is divided into two parts. The first part
estimates the amount of reduced engine wear to be expected with
a low sulfur fuel. Both experimental and in-use oil analysis
data are discussed. The second part discusses the implications
of this reduced wear on lubricating oil composition, oil change
interval, engine rebuild interval and vehicle life. Two likely
engine wear benefit scenarios are selected, and reductions in
certain operating costs are estimated and compared for these
scenarios.
II. Effect of Fuel Sulfur on Engine Wear
A. Background
The ERC Study investigated the effect of diesel fuel
sulfur content on engine wear.[l] ERC concluded that a
reduction of diesel fuel sulfur content from 0.27 to 0.05
weight percent would result in a 30 to 40 percent reduction in
engine wear and therefore a 30 to 40 percent increase in engine
life and oil drain interval. These conclusions were based
primarily on the results of a study by Tennyson and Parker,[2]
which was conducted on a two-cycle locomotive engine. However,
today's engine oils have additive packages which are better
able to handle the corrosive wear from fuel sulfur, and some
uncertainty existed in extrapolating the locomotive engine
results to on-road diesel operations. EPA therefore contracted
with Southwest Research, Inc. to further study the effect of
diesel fuel sulfur level in the range of 0.30 weight percent to
0.05 weight percent on wear in on-road heavy-duty diesel
engines. [3] A two part approach was used in this
investigation. The first part was to conduct a literature
search to uncover more data than that contained in the ERC
analysis. The second part was an empirical check on the first
in which Southwest contacted diesel fleets currently operating
on 0.05 weight percent sulfur fuel to obtain used oil
analyses. This effort was concentrated in the Southern
California area, which has undergone a legislated reduction in
diesel fuel sulfur content to 0.05 weight percent maximum.
Comparison was made of used oil analyses prior to sulfur
-------
5-2
reduction with those obtained after the reduction to determine
the effect of fuel sulfur reduction on engine wear.
In addition to contracting the wear study to Southwest,
EPA received comments on the wear benefits presented in the ERC
report from oil companies and engine manufacturers. Some
companies also submitted in-house data which characterized the
wear rates of current engines with current fuels and
lubricants. Most of these data and comments were also provided
by EPA to Southwest to assist them in their analysis, and are
discussed in this chapter where appropriate.
B. Experimental Results
The literature search done by Southwest was more
comprehensive than any previously undertaken in the area of
sulfur and engine wear. However, much of the data relating
sulfur to engine wear was old (1940s and 50s) and, due to the
difference in oil qualities, past and present, cannot be
applied quantifiably to current engines. Another limitation of
the older literature data is that the impact on engine wear was
addressed under carefully controlled steady state conditions,
and not more representative transient conditions. Most of the
studies however, showed that operating on low sulfur fuels
resulted in less piston ring and liner wear.
There was one recent comprehensive wear study that used
current engine lubricating oils and fuels. This was conducted
by Daimler Benz, and has been recently reported in the
literature.[4] Wear testing was performed on a Daimler Benz
0M422 engine (4 stroke, DI, naturally aspirated, 14.6 liter)
operating on API CC 20 W 20 oil. The oil was changed
frequently to retain resolution of the experimental measurement
technique and therefore there was no significant Total Base
Number (TBN-a measure of the oil's alkalinity value)
depletion. The effects of fuels with sulfur levels of 0.26 and
0.05 weight percent were investigated over a range of engine
operating conditions (load, speed, temperature). Reduced wear
due to low sulfur fuel was observed during conditions typical
of cold start and warm-up conditions. For instance, at coolant
temperatures of 122°F, an 80 percent reduction in bore wear was
observed. At 158°F, the reduction was 28 percent, and at
176°F, there was no difference in the observed rates of wear
for the different fuels. At higher temperatures, the low
sulfur fuel showed a slight increase in wear.
Several experimental studies also demonstrated that, for a
given fuel sulfur level, engine wear is sensitive to oil TBN
value [5,6,7], It has been recognized that, when certain
alkaline compounds are added to lubricating oil, these help to
neutralize the sulfuric acid produced in the engine, thereby
reducing corrosive wear. However, the alkalinity of an oil
drops with use due to the neutralizing process. At low
-------
5-3
alkalinity values, wear can increase rapidly. Because of this,
engine manufacturers generally specify a minimum new oil
alkalinity value and oil change interval which depend on an
expected in-use fuel sulfur level. This is done to ensure that
oil TBN does not drop to such a level that excessive wear would
occur. However, oil changes, as will be subsequently
discussed, are expensive for truck owners and operators. Truck
owners and operators who attempt to stretch oil change
intervals may experience more wear on current sulfur fuel than
they would on a lower sulfur fuel.
Experimental data are useful for identifying the general
causes and mechanisms of engine wear - that: 1) wear increases
with sulfur level, 2) wear generally decreases with increasing
temperature (although the Daimler-Benz data seem to indicate
that at high operating temperature wear may start to increase
somewhat), and 3) wear increases at higher loads. However,
in-use engines are operated under a variety of conditions that
are difficult if not impossible to simulate in a laboratory.
In an attempt to get a better idea of in-use wear levels with
different fuels, Southwest analyzed oil samples from fleets
operating in southern California before and after sulfur levels
were reduced.
C. Used Oil Analyses
On January 1, 1985, the sulfur content of diesel fuel was
reduced by regulation to 0.05 weight percent maximum throughout
the South Coast Air Basin, namely Los Angeles, Orange,
Riverside and San Berandino Counties in California. This
change provided an opportunity to gather fleet data indicating
whether or not lower sulfur diesel fuel may result in a
reduction in engine wear. Data on diesel fuel and used engine
oil analyses were sought from fleets operating in the area
before and after the implementation of 0.05 sulfur
legislation. The oil analyses were performed by the fleets at
the time of oil change. Data collected varied between the
fleets, but all four analyzed for iron content and reported the
mileage since the last oil change of each sample. Southwest
selected the largest fleet, the Southern California Rapid
Transit District (a bus fleet with mostly two-stroke Detroit
Diesel engines), and grouped the data by calendar year, engine
type and mileage interval, looking for a trend in iron content
and sample mileage. No consistent pattern of increasing oil
iron content with mileage interval was noted across all engine
types, and calendar years. However, there did seem to be a
somewhat consistent trend of increasing oil iron content vs.
mileage in 1984 and 1985 for all engine types in the two lower
mileage intervals (1-9,000 miles and 9-15,000 miles). The lack
of correlation at higher mileage intervals could be due to
significantly smaller sample sizes in those intervals, which
were not reported. Another cause could be that those engines
that go longer intervals between oil changes might be those
-------
5-4
that also burn more oil, and require more frequent fresh oil
additions, thereby diluting the concentration of wear products
in the oil. Because of the generally weak correlation between
iron content and mileage. Southwest grouped all samples in each
fleet by calendar year (1984, 1985 and 1986), and estimated
average mileages and iron content. Statistical tests were used
to note any significant differences in iron content or sample
mileage from 1984 to 1985 and 1986.
A summary of the reliable iron content data from the four
fleets is shown in Table 5-1. The data are disaggregated by
engine manufacturer and calendar year. Seven engine samples
are shown from three fleets. The percent reductions in iron
content in 1985 and 1986 were both estimated from the 1984
pre-sulfur control iron content values. The Detroit Diesel
engines from the SCRTD bus fleet are the largest samples in the
group.
There were judged to be seven unreliable engine samples
from the four fleets (not shown in Table 5-1). Two samples,
the Chandler Cummins and Rental GMC, were statistically
significant but had very low numbers of oil samples in one year
as compared to another. The trends in iron content were
dissimilar; the Chandler Cummins experienced a 120 percent
increase in iron content from 1984 to 1986, while the Rental
GMC experienced a 72 percent decrease over the same period.
There were also a number of non-significantly different
engine samples from each fleet. The SCRTD fleet had one
(Cummins engines), the Chandler fleet had 3 (Caterpillar,
Detroit Diesel and Mack) and the Rental fleet had 2 (Cummins
and Duetz). In most of these instances where the difference in
iron content was insignificant, there was either a small sample
size in one year compared to another, or little difference in
oil iron content from year-to-year, or both. Also, there were
no predominant trends in iron content from year-to-year in the
insignificant samples; in some cases iron content dropped in
the years with sulfur control and in other cases it increased.
The iron content data in Table 5-1 show a mixture of
results. The SCRTD DDA-71 two-stroke engines show a 19-23
percent reduction in 1985 and 1986 over 1984. The DDA-92
engines show about a 38 percent reduction. The MAN 866
4-stroke engines show a 67 percent increase in iron content in
1985 over 1984. DDA engines from the Rental fleet show a 26-34
percent reduction, similar to the SCRTD fleet. The Rental IHC
4-stroke engines show a 25-31 percent reduction which is also
similar to the DDA 2-stroke engines in the Rental and SCRTD
fleets. The Cummins and DDA samples in the Laidlaw fleet show
increases in iron content of between 26 and 85 percent in 1986
over 1984.
-------
5-5
Table 5-1
Samples with Significant Differences
in Oil Iron Content
Fleet Engine Type
Rental DDA
Year
SCRTD DDA71 2-stroke 1984
1985
1986
DDA92 2-stroke 1984
1985
1986
MAN 866 4-stroke 1984
1985
2-stroke 1984
1985
1986
4-stroke 1984
1985
IHC
1986
Laidlaw Cummins 4-stroke 1984
1986
No.
2054
5111
4836
458
4387
5551
59
67
126
162
200
90
114
108
25
60
Iron
Content (ppm) Standard
Mean Deviation
82.1
63.6
66.2
82.1
50.1
51.6
78.3
130.9
85.3
62.9
56.4
72.0
53.8
49.3
32.9
41.4
65.7
44.7
46.3
65.7
50.7
40.6
51.3
153.3
47 .7
33.2
24.2
36,
26,
25.8
Percent
Iron
Reduction
( 8x to 84)
23%
19%
20
30
39%
37%
-67%
26%
34%
2 5%
31%
-26%
DDA 2-stroke 1984 33
1986 63
31.0
57.4
20,
74,
-85%
-------
5-6
Overall, with the exception of the small Laidlaw DDA
sample, the two-stroke engines show a consistent decrease in
wear with low sulfur fuel. The four stroke engines show
increases and decreases in wear, but the largest sample (IHC)
shows a wear decrease similar to the 2-stroke engines. The
other two samples showing increases in wear, especially the
Laidlaw Cummins sample, could arguably be considered either too
small or imbalanced (in number of measurements in 1984 versus
1986) to make a valid comparison.
It should be pointed out that not all operating parameters
were constant during this three year period. The new oil TBN
used in the SCRTD fleet increased from 5.4 in 1984 to 7.4 in
1985 and 6.1 in 1986 (oil viscosity remained the same). New
oil TBN values for the other fleets were not available. The
oil iron values for DDA engines in the SCRTD fleet show lower
oil/iron contents in 1985 than 1986, which follows the trend of
oil TBN levels. However, the differences in percent reduction
in oil iron content from 1984 to either 1985 or 1986 are
minimal, and so the 6.1 and 7.4 TBN oils are assumed to result
in equivalent wear on a low sulfur fuel. Also, the Chandler
and Rental Truck fleets experienced a significant increase in
the levels of zinc present in the oil in 1985-86 compared to
1984. Zinc dithiophosphate (ZDTP) is well established as an
inhibitor of abrasive wear, and it is therefore difficult to
determine the relative contributions of the zinc increase and
sulfur decrease to the reduction in wear observed. Southwest
did conclude that the zinc increases did not contribute
significantly to the reduction in wear, but this conclusion was
not well supported and therefore the data from the SCRTD fleet
should be considered preferentially.
Although there is some uncertainty with some of the
samples, and it would be better to have more oil analysis data
on four-stroke engines operating under line-haul conditions on
current and low sulfur fuels, the available oil analysis data
indicate that a significant decrease in oil iron content
occurred with low sulfur fuel for the majority of the samples.
And, in spite of the problems with the oil analysis data, the
primary advantages of this data over the experimental data are
that it was gathered under in-use operating conditions and that
it measures total iron worn in the engine due to all causes.
The size of the wear decrease among samples showing
decrease in wear appears to be in the range of 20 to 40
percent. A value of 30 percent, therefore, is reasonable to
expect from all diesel engines, and will be used in all further
analyses. Adjustment to this value is necessary, however,
because the pre- and post-control sulfur levels in California
were 0.35 weight percent and 0.03 percent, respectively,
whereas 0.25 weight percent and 0.05 weight percent were
defined in chapter 2 to be representative levels to use for
this analysis.
-------
5-7
In a preliminary wear analysis which EPA released to peer
reviewers it was assumed that the difference in above sulfur
levels would result in a proportional reduction in the wear
benefit. In their comments on this analysis, the Engine
Manufacturers Association and the Motor Vehicle Manufactures
Association stated that experimental evidence suggested the
effect was non-linear, and that this should be taken into
account. [8] EMA and NMVA suggested using a wear model
developed by Daimler Benz to do this.[4] However, there are a
number of inputs to this model whose in-use values are not
known. Therefore, this analysis will continue to assume a
proportional reduction in wear benefit. The size of the wear
benefit thus adjusted is 18 percent and this is used in all
further analyses,
III. Effect of Reduced Wear on Operating Costs
In the previous section it was shown by both experimental
and in-use oil analysis data that lower diesel fuel sulfur
levels will likely lead to less diesel engine wear. Exactly
how this will translate into benefits to truck owners and
operators, however, is difficult to predict. There are a
number of possible outcomes. First, lubrication oil producers
and engine manufacturers could determine that since there is
less sulfur in diesel fuel, lube oils could be produced with
lower concentrations of anti-corrosive agents. This could
reduce the cost of lube oil, thereby reducing oil change costs
to truck owners and operators. The second possible outcome is
that engine manufacturers could recommend longer oil change
intervals, or truck owners and operators could decide to
stretch oil change intervals. According to a rebuild and
engine maintenance survey performed by EMA,[9] hereinafter
called the EMA Rebuild Survey, heavy heavy-duty truck operators
schedule oil changes at about 11,000 mile intervals. A truck
lasting 500,000 miles would therefore experience about 45 oil
changes. These are expensive for the truck owner and operator,
since both the labor and materials' cost of the oil change as
well as truck downtime are involved (this latter cost is
probably minimized by scheduling other maintenance along with
oil changes and by performing oil changes during natural truck
downtime). Nonetheless, truck owners and operators have a
significant incentive to increase oil change intervals if they
perceive it will not adversely affect engine wear. The third
possible outcome is that reduced engine wear could lead to
increased mileage to rebuild and possibly increased vehicle
life.
Any of these outcomes seems possible. Furthermore, it is
possible that there could be combinations of these outcomes.
The ERC report assumed that there would be an increase in
engine rebuild interval, engine life, and oil change interval.
The increase in engine life, however, was not assumed to
increase vehicle life, nor were oil costs assumed to be lower.
-------
5-8
None of the work done by Southwest revealed any information on
which combinations of these would be most likely.
Comments received by EPA on the ERC report were varied.
The engine manufacturing industry generally acknowledged that
one or more of these outcomes was likely. Caterpillar Co., m
their comments on the ERC report, stated that "the most
attractive prospect could be that the combination of low sulfur
fuel and reduced particulate emissions would result in extended
oil change periods."[10] Caterpillar indicates here that soot
accumulation in oil also has an effect on oil change interval,
i.e., truck owners and operators schedule oil changes based on
oil appearance or ash content. The "reduced particulates" in
Caterpillar's comments refers to future engines designed to
meet the 1991 and 1994 particulate standards, which are
expected to have engine-out particulate levels that are 60 to
75 percent less than current (pre-1988) levels. The
implication is that an increase in oil change interval would be
possible for future engines, but perhaps not current engines.
Caterpillar, Co. also submitted some information showing the
difference in the cost of lube oil at different TBN levels.
These costs were based on different types of oils ordered by
their test facilities, and not on a complete analysis of
additive components that would be required or desirable with a
low sulfur fuel.
API submitted information about the primary causes of
rebuilds in support of a position that any reduced engine wear
due to low sulfur fuel would not affect rebuild interval or
vehicle life.[11] Their comments were focused in two areas.
The first comment cited the 1981 EMA Rebuild Survey which found
that engines are usually rebuilt for either loss of oil
control, or for mechanical failures such as bearing, camshaft
and fuel injector failures. API concluded that trucks in the
second category would not experience any increase in engine
rebuild interval with low sulfur fuel. EPA agrees with this
conclusion since these failures appear unrelated to
corrosivering and liner wear. The second comment was that loss
of oil control was the result of factors also unrelated to
corrosive wear, such as abrasive wear from top land piston
deposits, piston ring scuffing and broken piston rings. API
cited two experimental test programs which tested engines on
several different fuels (of varying sulfur content) and
lubricants to support this.[12,13] The test programs also
showed, however, that sulfur, while it does not lead to the
formation of top land deposit it does form deposits in the
bottom land and bottom ring groove. Also, the programs
demonstrated that corrosive wear could cause increased oil
consumption if oil alkalinity values were too low.
The best way to resolve this issue would be to examine
in-use oil consumption data on two- and four-stroke engines on
current and low sulfur fuels. Unfortunately, this data was not
-------
5-9
available from the Southern California fleets. The fleet oil
analysis data., did, however, show a difference in wear due to
fuel sulfur level. The iron contents of samples with both high
and low sulfur fuels would have included wear products from all
causes of wear (abrasive or corrosive), and in all parts of the
engine. Southwest did say that 80 percent of the wear products
in the oil would come from the ring and liner. If this is
true, then the difference in wear between high and low sulfur
fuels in the fleet oil analysis data would be expected to lead
to differences in oil consumption.
Overall, both the experimental and in-use oil analysis
data indicate that engine wear is lower with low sulfur fuels.
However, there is still some uncertainty as to how this will
benefit truck owners and operators. Therefore, two scenarios
have been chosen that are believed to set upper and lower
bounds to the size of this benefit. The first scenario assumes
there will be a reduction in oil cost and an increase in oil
change interval, with no increase in engine rebuild interval,
engine life, or vehicle life. The basic hypothesis of this
scenario is that wear testing and/or experience on the part of
both engine manufactures and lubricating oil produces with the
lower sulfur fuels will result in the oil producers reducing
the TBN of lubricating oils and engine manufacturers
recommending longer oil change intervals for their engines.
Since comments from at least one manufacturer indicated oil
change interval to some extent depends on engine-out
particulate levels, this benefit will be assumed to apply only
to cars and trucks with low particulate levels (this will be
discussed in the next section). It is further assumed that oil
change interval is not limited in the range it is extended by
the degradation of other oil parameters such as viscosity.
The second scenario assumes there will be no change in oil
cost or oil change interval, but there will be an increase in
engine rebuild interval for those engines which are rebuilt for
reasons related to high oil consumption. Two cases are
examined in this scenario which are derived from assumptions
about why trucks are scrapped. The first case assumes trucks
are always scrapped due to engine problems. In this case, the
extension in engine rebuild interval leads to an extension in
engine and vehicle life. The second case assumes that there is
no correlation between engine rebuild interval and vehicle
life, i.e., that some or many trucks are scrapped because of
reasons unrelated to the increase in rebuild interval (for
example, transmission or rear axle failure). The benefits for
this case, then, are an extension in rebuild interval and a
decrease in the total number of rebuilds needed for the in-use
fleet of trucks.
Operating cost reductions for these scenarios are
developed below.
-------
5-10
A. Oil Change Cost Reduction
This scenario assumes that lubricating oil producers
reduce oil alkalinity values and engine manufacturers recommend
longer oil change intervals to yield wear that is equivalent to
the wear experienced with current sulfur fuels. The vehicle
model year applicability of these benefits will be discussed
first. Next, a discussion of the potential for oil cost
reduction with a reduced alkalinity content will be presented.
Finally, this section will be concluded with a discussion of
the procedures for estimating the reduction in the number of
oil changes and a presentation of the operating cost reductions
using these methods.
Comments from one engine manufactuer have indicated that
increased oil change intervals might be experienced on vehicles
with low engine out particulate levels. The task here is to
decide for this analysis what level is low enough for each
vehicle type. Engine out particulate levels have been reduced
by the engine manufactures in order to comply with the
particulate emission standards, so these can be used to
estimate the model year applicability of this benefit. The
particulate emission standards for the different vehicle types
are shown in Table 5-2. For LDDVs and LDDTs, the emission
standard in 1988 remarks a 60-70% reduction in particulate
levels. For heavy-duty vehicles, the most stringent standard
is in 1994, but the 1991 standard respresents about a 60-70%
decrease from pre-1988 levels. For buses, the emission
standard in 1991 represents almost a 90% reduction from current
levels. This analysis will assume that the increase in oil
interval will apply to 1988 and later LDDVs and LDDTs, and 1991
and later HDVs and buses.
It may actually (for example, 1996 or 1997) be several
years beyond the implementation of sulfur controls that engine
manufacturer decide they can recommend increased oil change
intervals for their new engines. However, this recommendation,
brought about by lower sulfur fuel which older engines also
would be using, would be expected to influence oil change
intervals on the existing low emitting 1991 and late HDVs and
buses, and 1988 and later LDVs and LDTs) fleet.
Therefore, this benefit will be assured to apply to all
1988 and later LDDVs and LDDTs, and 1991 and later HDVs and
buses in all calander years after the start of sulfur
controls. The oil change intervals, lifetime mileages and oil
change costs for all vehicles are shown in Table 5-3. Oil
change intervals for LDDVs, LDDTs and LHDVs came from
conversations with automobile dealers, and the intervals for
medium heavy and heavy heavy-duty trucks were taken from the
EMA Rebuild Survey.[9] The bus oil change interval came from
the in-use SCRTD bus oil analysis data in the Southwest Wear
Report.[3] Lifetime mileages are from the Diesel Particulate
-------
5-11
Vehicle Type
LDDV S< LDDT
LHDV - HHDVs
Buses
Table 5-2
Particulate Emission
Standards
MYR Group
1982-87
1988+
1988-90
1991-93
1994 +
1988-90
1991 +
Particulate Standard
0.6 gpm
0.2 gpm
0.6 g/BHP-hr
0.25 g/BHP-hr
0.10 g/BHP-hr
0.6 g/BHP-hr
0.1 g/BHP-hr
-------
5-12
Table 5-3
Input Parameters for
Estimating Reductions in Oil Change Costs
Parameter IPPVS LOOTS IHPVS MHPVS
Oil Change Interval (mi) 5.000 5,000 5,000 11,000
Lifetime mileage (mi) 100,000 124,000 128,000 268,000
Lifetime (yrs) 7.2 10.5 8.0 8.1
Oil Change Cost ($) 30.00 30.00 30.00 100.00
Oil Changes Per Year 2.78 2.36 3.20 3.0
Annual Cost 83.40 70.80 96.00 300.00
HHPVs
Buses
12,000 6,000
529,000 540,000
8.2 12.0
100.00 100.00
5.4 7.5
540 00
750.00
-------
5-13
Study.[14] Oil change costs were determined by contacting
several service shops in the Detroit area.
When lifetime mileage is divided by oil change interval
and vehicle life, the result is the number of oil changes per
year, which is also shown in Table 5-3. The number of oil
changes ranges from 2.36 for LDDTs to 7.5 for buses.
Multiplying annual oil changes by oil change cost gives the
annual oil change cost, shown in the bottom row.
Information supplied by Caterpillar Co. indicated that
there is about a 1.3 percent decrease in the wholesale price of
lubricating oil for every unit of reduction in TBN value. This
was based on the wholesale cost of oil at two TBN levels - a
TBN of 13 ($2.01 per gallon), and a TBN of 7 ($1.86 per
gallon). This analysis will assume an equivalent percentage
change in the retail price of oil over this TBN range.
Additional items needed to estimate the reduction in
lubrication oil cost are the current TBN level of oils used in
trucks and the expected oil TBN level when using low sulfur
fuel. The ERC report stated that new oil TBN values typically
are in the range of 6-10. Engine manufacturers commenting on
the report concurred, and data on new oil used in the 1984
SCRTD fleet in the wear report indicate new oil TBN values of
5.4 to 7.4. Therefore, this analysis will use a TBN value of
7.0 for current sulfur fuel.
It is difficult to predict how low oil producers would
reduce TBN with low sulfur fuel. The engine manufacturers have
widely varying specifications for new oil TBN.[3] However,
Caterpillar Co. recommends a new oil TBN level of about 20
times the sulfur level. A 0.05 weight percent sulfur fuel
would thus yield an oil TBN value of 1.0. It is doubtful that
an engine manufacturer would recommend such a low TBN level,
even with low sulfur fuel, and so for this analysis a TBN value
of 3.0 will be used. As will be be presently demonstrated,
overall oil change costs are not at all sensitive to this TBN
level given the TBN price relationship described above.
The final piece of information needed is a typical oil
crankcase capacity. These can also vary considerably from
manufacturer to manufacturer, and from engine to engine. The
1980 edition of Motor's Truck and Diesel Repair manual lists
the crankcase capacities of Cummins engines in the range of
28-36 quarts.[15] For Mack engines, the capacity ranges from
15-20 quarts and for Detroit Diesel engines ranges between 20
and 30 quarts. This analysis will use 24 quarts (6 gallons) as
a typical volume of oil needed for an oil change.
The savings in new oil cost obtained by using this input
data for a 3.0 TBN oil is about $1.50 per oil change. At TBN
levels of 2.0 and 4.0 the savings are $1.00 and $2.00,
respectively. When these savings are compared against the
-------
5-14
total oil and filter change cost of $100 (Table 5-2) it is
clear that oil TBN reduction appears to have a small impact on
oil change price, and that overall oil change cost is not
sensitive to TBN level. Therefore, oil cost savings of $1.50
(at a TBN level of 3.0) will be used for all further
calculations.
To conclude the oil change cost analysis, the extension in
oil change interval with low sulfur fuel must be estimated.
The hypothesis used in this analysis is that oil change
intervals can be increased and oil cost decreased to yield
equivalent wear to a current sulfur fuel. The conclusion of
the Southwest oil sample analysis section was that diesel
vehicles would experience 18 percent less wear with the low
sulfur fuel (.05 weight percent) than with the current sulfur
fuel (.25 weight percent). It is therefore assumed that trucks
and buses could travel 18 percent further between oil changes.
The adjusted oil change intervals and numbers of oil
changes per year are shown in Table 5-4. Also shown are the
adjusted annual costs, the net savings per year and discounted
lifetime cost reductions. The net savings per year are
estimated by subtracting adjusted annual cost as presented in
Table 5-4 from the annual oil change cost as presented in Table
5-3. These are further discounted over the vehicle life in the
last row. The results show discounted lifetime cost reductions
of $35 to $45 for LDDVs, LDDTs and LHDVs, and $300 to $900 for
the heavy duty vehicles and buses. On a fuel consumption basis
(taking undiscounted lifetime costs and dividing by
undiscounted lifetime miles, then multiplying by model year
2000 fuel economies as listed in Table 5-4 [16,17]) this works
out to savings of between 0.92 cents per gallon (HHDVs) and
1.75 cents per gallon (buses). These benefits are assumed to
apply only to 1988 and later LDDVs and LDDTs and 1991 and later
heavy-duty duty trucks and buses.
In this example the most uncertain input parameters are
the oil cost reduction and increased oil change interval.
However, only 6 to 8 percent of the benefit (depending on truck
type) is due to the oil cost reduction. The remaining 92-94
percent is due to the reduction in oil change frequency.
Therefore, the size of this benefit is directly related to the
extension in oil change interval—if oil change intervals can
be extended further than 18 percent, the benefit will be
larger. If oil change intervals are extended less than 18
percent, the benefit will be proportionally smaller. At this
juncture there is no technical basis for either a higher or
lower percentage oil change interval extension.
-------
5-15
Table 5-4
Bus
7.080
6.4
630.40
119.60
895.00
1.75
Fuel
economies
LODV
33 mpg
LDDT
26 mpg
LHDV
16 mpg
MHDV
8.1 mpg
HHDV
6.8 mpg
Bus:
6.6 mpg
Discounted Lifetime Cost for
Oil Change Interval Scenario
Parameter LPPVs IPPTS LHOTs HHP HHPT
Adjusted Oil Change Interval (mi) 5,410 5,410 5,410 12,980 14,160
Adjusted Oil Changes Per Year 2 56 2 18 2.95 2.5 4.6
Adjusted Annual Cost ($) 76.80 65.40 88.50 246.25 453.10
Net Savings Per Year ($) 6.60 5.40 7.50 53.75 86.90
(undiscounted)
Discounted Lifetime Cost Reduction ($) 37.02 36.78 45.46 326.00 526.00
Cost-Reduction Per Gallon* (?/gal) 1.56 1.18 0.75 1.32 0.92
-------
5-16
B. Extension in Engine Rebuild Interval/Vehicle Life
This scenario assumes there is no change in oil
composition (or cost) or oil change interval, but there is
reduced engine wear which leads to increased engine rebuild
intervals, engine life and vehicle life in those engines that
fail or are rebuilt for reasons related to high oil
consumption. In this scenario, it is necessary to analyze all
diesel vehicles, not just those with engines that are rebuilt,
since reduced engine wear could affect vehicle life in vehicles
whose engines are not rebuilt also.
Although LDDV, LDDT and LHDV engines are occasionally
rebuilt where the condition of the rest of the vehicle is
excellent, these situations are believed to be rare and this
analysis assumes these engines are never rebuilt. Also, some
MHDVs with nonsleeved engines are not rebuilt.[18] These are
referred to in this paper as MHDVls, and are analyzed with the
LDDVs, LDDTs, and LHDVs. (The remainder of the MHDV class will
be discussed in a later section.)
The question then becomes how increased engine life could
affect vehicle life for these lighter vehicle classes that are
not rebuilt. Several sources were contacted to determine why
vehicles are scrapped. Little information was obtained which
could be used to quantify the effect of engine life on vehicle
life. Therefore, two sensitivity assumptions are possible.
The first one is that vehicle life is not influenced by engine
life (i.e., all vehicles are scrapped for non-engine related
reasons such as transmission, body, etc.), and in this case
there is no economic benefit. The second one is that an
extension in engine life always leads to increased vehicle
life. In this case an economic benefit exists in deferring the
purchase of a replacement vehicle. The actual situation
probably lies somewhere between these two extremes, and since
the engine is the single most expensive piece of equipment in a
vehicle, it probably lies closer to the second one, since
vehicle ownersare likely to replace or repair cheaper car
components (except for perhaps the transmission) until the
engine fails. However, there is currently no technical basis
for estimating the "actual" situation for these vehicle types
(LDDVs, LDDTs, LHDTs, and MHDVls), and so only the second case
is estimated (i.e., extension in engine life always leads to an
extension in vehicle life).
The reduction in engine wear is also expected to result in
longer rebuild intervals for rebuilt MHD, HHD and bus engines
which are rebuilt for reasons related to high oil consumption.
There is an economic benefit, then, in rebuild costs which are
delayed. Like the lighter diesel vehicles, there is
uncertainty in how reduced wear will affect vehicle life.
Therefore, two cases are examined for these vehicles also.
-------
5-17
The first assumes that an extension in rebuild interval(s)
always extends vehicle life. In this case an economic benefit
occurs not only due to delayed rebuilds, but also to a delayed
replacement purchase. The second case assumes that no
correlation exists between extended engine rebuild interval and
vehicle life. In this second case an economic benefit occurs
from delayed rebuilds, and a further economic benefit occurs
from some rebuilds which are not needed. These occur in
engines which experience first and second rebuilds at high
odometer values with a current sulfur fuel, but only one
rebuild (at an even higher mileage) with low sulfur fuel. In
this case, the cost of a second rebuild is avoided. Under this
approach, vehicle life is not extended.
As stated earlier, some MHDVs have nonsleeved engines that
are not rebuilt by truck owners and operators (MHDVls).
However, many of the non-sleeved MHD engines are rebuilt.
Also, many manufacturers make sleeved diesel engines for MHDV
use, and these have characteristics (mileage accumulation vs.
age,) like HHDV engines. Therefore, all three classes of MHDV
engines will be analyzed separately and weighted together at
the conclusion of the analysis to yield a single MHDV wear
benefit. For convenience, unsleeved nonrebuilt engines are
referred to in this analysis as MHDVls and unsleeved rebuilt
engines will be referred to as MHDV2s. The rebuilt, sleeved
MHDVs will be assumed to have the same benefits as HHDVs, and
thus will not receive a separate label.
To summarize, the types of benefits for which reductions
in operating costs need to be estimated depend on the
correlation between engine and vehicle life. Where an
extension in engine life also extends vehicle life, the
benefits are deferred replacement vehicle cost for all diesel
vehicles, and deferred rebuild costs for rebuilt MHDVs, HHDVs
and buses. Where an extension in engine life does not lead to
an extension in vehicle life, the benefits are reductions in
the total number of rebuilds, and a delay in when they occur.
These benefits apply only to MHDVs, HHDVs and buses; there are
no benefits for LDDVs, LDDTs and LHDVs for this situation.
Operating costs for the situation where an extension in
engine life leads to an extension in vehicle life will be
examined first. Base case (current sulfur fuel) operating
costs are examined first, followed by low sulfur operating
costs.
1. Reduction in Operating Costs - Increased Engine
Rebuild Interval and Vehicle Life
Base Case - Base case vehicle replacement costs are
estimated for all vehicles first. It should be mentioned that
this analysis does not examine all operating costs such as
minor maintenance, fuel, etc., but only those that are
-------
5-18
significantly affected by an increase in engine rebuild
interval and vehicle life. All other operating costs are
assumed to be the same on a per mile basis with current and low
sulfur fuel.
To estimate the base vehicle replacement costs, the
analysis requires estimates of new vehicle cost and vehicle
lifetime (in years and mileage). Vehicle cost and mileage can
then be discounted and divided to yield a net replacement cost
per mile for the base case.
New vehicle cost, lifetime mileage, and lifetime in years
for all vehicles used in this analysis are shown in Table 5-5.
New vehicle costs were obtained by contacting several
Detroit-area car dealerships. Very few manufacturers are
offering diesel LDVs for sale in the U.S. at this time (Ford
and Daimler-Benz), so this cost is difficult to project.
Rather than rely on the price of the one or two models
available, a price was selected that may be closer to an
average if several manufacturers were marketing diesel
light-duty vehicles. Costs for LDDTs, LHDTs, MHDVs and HHDVs
were determined from conversations with manufacturers and
dealerships. These costs can vary greatly with the type of
optional equipment chosen by the buyer, and are not meant to be
any kind of industry weighted retail average. The bus price
was determined from conversations with transit managers.
Lifetime mileages were taken from the RIA for the 1988 and
later heavy-duty engine NOx and particulate standards.[19]
Vehicle lifetimes were estimated from the above lifetime
mileages and the M0BILE4 VMT vs. age distributions for these
vehicle types.[20] The ages for LDDVs, LHDVs, and MHDVs appear
somewhat low in this analysis, but should not have much of an
effect on the results of this analysis since the analysis will
be focusing on the difference in lifetime and lifetime mileage
between fuel scenarios, rather than on the absolute values of
these estimates.
When vehicle costs are discounted over the ages listed in
Table 5-4 with a factor of 10 percent, and divided by
discounted lifetime mileage, the result is discounted base case
operating costs, also shown in Table 5-5. These costs range
from 6.57 cents per mile for LDDTs to 18.66 cents/mile for
MHDVls.
Next, rebuild costs must be developed for MHDV2s, HHDVs
and buses. The key input parameters are rebuild mileages and
ages, the fraction of vehicles receiving rebuilds, and rebuild
costs. Rebuild mileages and ages can be used in discounting
mileage and rebuild costs to determine their net present
values. The fractions of vehicles that receive rebuilds are
used to estimate a net fleet rebuild cost, since some vehicles
may be scrapped before a particular rebuild occurs.
-------
5-19
Table 5-5
Base Case Vehicle Replacement Cost
Input Parameters for All Vehicles
Parameter
Vehicle Type
Value
Vehicle Costs (1987 $)
Lifetime Mileage (mi)
Lifetime (yrs.)
LDDV
LDDT
LHDT
MHDV1
MHDV2
HHDV
Buses
LDDV
LDDT
LHDT
MHDV1
MHDV2
HHDV
Buses
LDDV
LDDT
LHDT
MHDV1
MHDV2
HHDV
buses
12,000
14 , 000
20,000
45,000
45,000
80,000
150,000
100,000
124,000
128,000
200,000
321,000
529,000
540,000
7.2
10.5
8.0
5.4
10.8
8.2
12. 0
Discounted Vehicle
Replacement Cost (#/mile)
LDDV
LDDT
LHDT
MHDV1
MHDV2
HHDV
Buses
8.89
6. 57
10.83
18.66
8.10
10 . 50
14 .90
-------
5-20
Rebuild mileages for MHDVs, HHDVs and buses were developed
from survey data reported by a publication called Fleet
Equipment (formerly Fleet Maintenance and Specifying).[18] A
portion of the data on rebuild mileages for sleeved diesel
engines is shown in Table 5-6. Fleets were requested by the
publication to report engine mileage before and after the first
major overhaul. Data were disaggregated by fleet type. The
"Miles Before Overhaul" data were used for the mileage to
first overhaul, and the "Miles After Overhaul" data were used
for the mileage between the first and second overhaul. The
survey did not presume a second overhaul, so data were not
gathered on the mileage after a second overhaul. For buses the
bus fleet data from Table 5-6 were used in this analysis; for
HHDVs the non-bus weighted averages were used. The Fleet
survey listed a mileage to first rebuild for MHDV2s of 175,000,
and a mileage after rebuild of 146,000 miles.
The next item that needs to be estimated to determine the
base case rebuild costs is the fraction of vehicles that
receive rebuilds. At this point, estimates have been made of
average vehicle lifetime mileage and rebuild mileage. However,
few vehicles are actually rebuilt (or scrapped) at these
average mileages, but are rebuilt (or scrapped) over a
distribution of mileages centered about these average
mileages. Therefore, it is possible and probable that some
HHDVs and buses are scrapped before the second rebuild, or that
some MHDVs are scrapped before the first rebuild. For these
vehicles, there is no rebuild cost to estimate.
A standard Monte Carlo analysis was used to predict the
fraction of trucks receiving rebuilds.[21] In this technique,
distributions are created which are centered about the average
rebuild and scrappage mileage. A finite truck sample is then
created, and for each truck in the sample, the analysis picks a
scrappage mileage and rebuild mileages. This is accomplished
with the average rebuild and scrappage mileage, the estimated
sample coefficient of variation (COV), and a table of random
normal deviates. Scrappage and rebuild mileages are calculated
for each truck by multiplying the random normal deviates, the
COV and the sample average. A new random deviate is selected
each time a new rebuild mileage or scrappage mileage is
estimated. The scrappage and rebuild mileage for each truck
are then compared to see if the scrappage mileage is less than
the rebuild mileage. If so, then there is no rebuild for that
truck. If the scrappage mileage is greater than the rebuild
mileage, then a rebuild is assumed to have taken place.
The use of this technique requires estimates of the COV
for the scrappage and rebuild mileages. The only data found on
how scrappage mileages are distributed, were some data on
rebuild interval distributions in the EMA Rebuild Survey.
Truck owners were asked to recall the mileage at which the
first rebuild was performed. A portion of the data for Class
-------
Table 5-6
Rebuild Mileages for Sleeved Engines
Category
N
Non-Bus
Fraction
Miles Before
Overhaul
Miles After
Overhaul
Private Fleet
176
.51
273,000
224,000
Contract/
Common Carrier
120
.35
308,000
268,000
Bus Fleet
18
-
304,000
236,000
Utility Fleet
32
.09
227,000
163,000
Lease/
Rental Fleet
4
. 02
123,000
224,000
Government Fleet
10
.03
123,000
97,000
Non Bus Total
342
Non Bus Weighted Avg.
277,000
230,000
-------
5-22
6, 7 and 8 trucks is shown in Table 5-7. The data for Classes
7 and 8 appear to be approximately normally distributed, while
Class 6 is bimodal. The Class 6 bimodal distribution is
probably the result of non-sleeved diesels being rebuilt at a
lower mileage and sleeved engines being rebuilt at a higher
mileage. Ideally it would be best to develop separate COVs for
all of the classes for which rebuild mileages are being
estimated, but there does not appear to be sufficient data to
do that. Therefore, a COV will be estimated from all of the
data in Table 5-7 and applied to all vehicles. About 60
percent of the rebuilds for these vehicles are performed
between 200,000 and 350,000 miles. The coefficient of
variation (COV), then, is about 27 percent. Therefore, this
analysis will use a value of 25 percent for the COV of all
average rebuild mileages. Furthermore, the analysis will also
use this value to establish the distribution of scrappage
mileages. The percentages of trucks receiving rebuilds using
these values and the Monte Carlo analysis are shown in Table
5-8.
From the data in Table 5-8 it is evident that nearly all
trucks receive a first rebuild, about one-half of the trucks
receive second rebuilds, and a few trucks receive third
rebuilds. At this point this portion of the analysis rests on
how well the rebuild and scrappage distributions have been
estimated. No survey data have yet been found to compare to the
values in Table 5-8. It is also now evident that to estimate
rebuild costs, estimates of vehicle mileage between second and
third rebuilds, and third rebuilds and scrappage are needed.
(Mileages before and after first rebuild were presented in
Table 5-6). The mileage between the first and second rebuild
as presented in the previous section will be used for these
mileages. The final item that needs to be determined is the
cost of the various rebuilds.
Rebuild costs are a function of the cause of engine
rebuild, whether the rebuild can be performed in the frame or
must be removed, the complexity of the engine, and a host of
other factors. Generally, if only the power cylinder
components are being replaced, the rebuild is done in-frame.
If the rebuild requires crankshaft removal, however, the
rebuild is performed with the engine removed from the frame.
According to the EMA rebuild survey the components most often
replaced or rebuilt during the first rebuild are piston rings,
connecting rod bearings, main bearings, cylinder liners,
injectors and nozzles, and the cylinder head assembly and
pistons. Camshafts, crankshafts, push rods/tubes, roots
blowers, aftercoolers and vibration dampers are seldom rebuilt
on the first rebuild, but may need attention in the second
rebuild.
Several repair shops were contacted for estimates of
rebuild costs and most would only give ranges of costs. From
-------
5-23
Table 5
-7
Distribution of
Mileages
to
First
Overhaul by
GVW Class
from EMA
Rebuild
Survey
GVW
First
Overhaul
Total
8
7
6
Under 100,000 Miles
5%
5%
2%
6%
100,000
- 149,999
7
6
4
16
150,000
- 199,999
9
8
8
24
200,000
- 249,999
19
19
22
24
250,000
- 299,999
18
19
21
3
300,000
- 349,999
23
24
18
16
350,000
- 399,999
9
9
5
9
400,000
- 449,999
5
5
14
1
450,000
- 499,999
2
2
—
—
500,000
- 549,999
1
1
3
—
550,000
- 599,999
—
—
—
—
600,000
- Over
2
2
3
1
100%
100%
100%
100%
Average
(xl,000)
259
262
280
203
Median
(xl,000)
278
283
281
208
Sample Size:
423
271
100
52
-------
5-24
Table 5-8
Percentage of Trucks Receiving Rebuilds
from Monte Carlo Analysis (COV = 25%)
MHDV2S
HHDVs
Buses
1st Rebuild
92%
95%
93%
2nd Rebuild
54%
62%
53%
3rd Rebuild
8%
14%
9%
-------
5-25
these conversations, a first rebuild repair cost for MHDV2s of
$4,000 is estimated, and second rebuild cost for MHDV2s are
estimated at $6,000. First rebuild costs for HHDVs and buses
are estimated at $5,000. Second rebuild costs for heavy
heavy-duty trucks and buses are estimated at $8,000. Third
rebuild costs are assumed to be the same as the first rebuild
cost for all vehicle types.
An analysis of base case rebuild costs for HHDVs using the
data developed in previous sections is shown in of Table 5-9.
Lifetime mileages, vehicle life, and discounted lifetime
mileage were the same values as those used in the oil cost
analysis. Mileages to rebuild were determined from the Monte
Carlo analysis of the fractions of trucks receiving each
rebuild. Total present rebuild cost is estimated by weighting
the present value of each rebuild cost by the fraction of
vehicles receiving that rebuild. Lifetime rebuild costs are
obtained by dividing the total present cost by discounted
lifetime mileage. Lifetime base case rebuild costs range from
1.29 cents/mile for buses to 1.93 cents per mile for MHDV2s.
Total base case costs for heavy-duty trucks and buses can now
be estimated by adding the vehicle replacement costs (Table
5-5) to the rebuild costs. Total base case costs for all
vehicle types are summarized in Table 5-10. Replacement and
rebuild costs for MHDVs have been weighted together using
weighting factors from the RIA for the NOx and particulate
Final Rule.[19] Total costs range from 6.57 0/mi (LDDTs) to
16.19 0/mi (bases). Rebuild costs range between 8 and 15
percent of the total of rebuild and replacement costs.
Sulfur Control Case - To estimate the sulfur control
vehicle replacement and rebuild costs estimates of the number
of vehicles that will benefit from sulfur control and size of
the benefit are needed. These estimates can then be used to
extend rebuild interval and vehicle life for those that
benefit, resulting in lower fleet average operating costs. The
first item that is estimated is the fraction of vehicles
benefiting from sulfur control. It was established earlier in
this chapter that the vehicles that would benefit from sulfur
control are those that are rebuilt or scrapped due to the
problems related to high oil-consumption.
Information on this item was available from the EMA
Rebuild Survey. This survey included both fleets and
owner-operators. Participants were asked to indicate the
criteria they use to determine when an engine is rebuilt.
Several criteria were provided on the data form (such as engine
hours, years in use, oil consumption), and participants could
also write-in alternate criteria.
Some of the results of this survey are shown in Table 5-11
in the column labeled "Percent of Fleets Citing." Eighty
percent of the fleets and 71 percent of the owner operators
-------
5-26
Table 5-9
Base Case Engine Rebuild Costs
MHDV2s HHDVs Buses
Lifetime Mileage (mi) 321,000 529,000 540,000
Vehicle Life (yrs) 10.8 8.2 12.0
Discounted Lifetime Mileage (mi) 196,927 348,689 320,290
First Rebuild
Mileage to First Rebuild 170,960 266,114 293,473
Age at First Overhaul (yrs) 4,5 3.4 6.5
Cost of Overhaul ($) 4,000 5,000 5,000
Present Value Cost of
First Rebuild ($) 2,605 3,616 2,691
Second Rebuild
Mileage to Second Rebuild (mi) 297,665 480,123 490,157
Age at Second Rebuild (yrs) 9.4 7.1 10.9
Cost of Second Rebuild ($) 6,000 8,000 8,000
Present Value Cost of
Second Rebuild ($) 2,449 4,0 66 2,831
Third Rebuild
Mileage at Third Rebuild (mi) 368,966 631,751 596,910
Age at Third Rebuild (yrs) 13.5 10.8 13.3
Cost of Third Rebuild ($) 4,000 5,000 5,000
Present Value Cost of
Third Rebuild ($) 1,104 1,786 1,407
Total Present Cost ($) 3,807 6,206 4,130
Lifetime Rebuild Cost (0/mi)
1.93^/mi
1.78^/mi
1.29^/mi
-------
5-27
Table 5-10
Base Case Replacement
and Rebuild Costs (cents/mile)
Vehicle Type Replacement Costs Rebuild Costs Total Cost
LDDV
LDDT
LHDT
MHDV
HHDV
Buses
8.89^/mi
6. 57
10 .83
11.58
10 . 50
14 . 90
none
none
none
1.46^/rni
1.78
1.29
8.89^/mi
6.57
10.83
13 . 04
12.28
16 .19
-------
5-28
Table 5-11
EMA Rebuild Criteria Survey
Fleet Survey
Criteria
Oil Consumption
Maj Engine Failure
Reduction in Oil Press
Blowby
Loss of Performance/
Reduced Power
Miles
Oil Analysis
Compression Pressure
Engine Hrs.
Years in use
Total Responses
Percent of
Fleets Citing
80
65
62
57
46
32
22
18
9
6
397
Frequency of
Reasons Cited %
20.2
14 .4
5.8*
5.5
2.2
48. 1%
Owner Operator Survey
Percent of Owner Frequency of
Criteria Operators Citing Reasons Cited (%)
Oil Consumption
71
16.3
Maj Engine Failure
69
Reduction in Oil Press
65
12.9
Blowby
56
Loss Of Performance/
Reduced Power
53
6.1*
Miles
34
Oil Analysis
33
7.6
Compression Pressure
29
Engine Hrs.
15
3.3
Years in use
10
Total Responses
435
46.2%
Value includes only half of responses
-------
5-29
cited oil consumption as the major cause for engine rebuild.
However, participants in many cases listed more than one
criteria for engine rebuild. Therefore, it cannot necessarily
be inferred that 80 percent of engines are rebuilt for high oil
consumption, since there may be other failures in conjunction
with high oil consumption that actually trigger a rebuild. To
determine the likely percentages of those engines which were
rebuilt for oil consumption alone (and that would experience an
increase in rebuild interval with lower oil consumption and a
low sulfur fuel), criteria were selected in Table 5-LO that
relate to oil consumption, the total responses were summed, and
the oil consumption criteria were renormalized to the total
responses. The results are shown in the column labeled
"Frequency of Reasons Cited Percent." The criteria selected
were oil consumption, blowby and oil analysis. Excessive
blowby can only be caused by worn rings and/or liners. A
rebuild undertaken because of oil analysis would only be
performed if excessive wear products were observed in the oil,
again primarily related to ring and liner wear. Reduction in
oil pressure was not selected because this problem is usually
caused by a faulty oil pump or looseness in engine parts where
oil is applied. (Low oil pressure due to excessive oil
consumption can usually be solved by adding oil, or topping
off.) Loss of performance/reduced power and compression
pressure could be caused by valves or lack of fit between rings
and liners, or both. For this analysis, it will be assumed
that one-half of the responses are due to valves and one-half
due to ring/liner wear for each of these two criteria. The
resulting percentage of engines rebuilt due to ring and liner
wear for fleets is 48.1 percent, and for owner operators is
46.2 percent. The remainder of this analysis will use 47
percent as the percent of trucks benefiting from sulfur control.
In their comments on the draft EPA wear analysis EMA and
MVMA questioned the use of this survey data to predict the
percentage of engines benefiting from sulfur control.[22] It
is expected that maintenance supervisors for fleets and
owner/operators responding in the survey have identified some
reasonable, experience-based criteria for rebuilding engines
that seeks to minimize total maintenance cost. That the survey
participants identified oil consumption is not surprising,
since excessive oil consumption can lead to excessive wear and
a catastrophic engine failure, which could substantially
increase the cost of an engine rebuild or result in that engine
being scrapped. Therefore, EPA views this survey data as
representative for actual engines.
There are no data on the causes of engine failure in
LDDVs, LDDTs, LHDTs, and MHDVls. Therefore, this analysis will
use the truck fraction developed above (47%) for the fraction
of these vehicles benefiting from reduced sulfur levels.
-------
5-30
Data from the in-use oil analysis indicated that diesel
vehicles would experience 18 percent less wear on a low sulfur
fuel. The assumption used in this analysis is that vehicles
benefiting from sulfur control travel 18 percent further with
low sulfur fuel before engine rebuild or scrappage. However,
some of these vehicles will fail for non oil-consumption
reasons, or crash in the extension period. These vehicles
receive some, but not the full, benefit.
The effect of these two phenomena can be accounted for by
reducing the percent of vehicles benefiting (i.e., 47 percent)
from low sulfur fuel to a lower level. To do this, estimates
of the length in miles of the extension period must be
evaluated. Next, the rate of crashes and non-oil control
failures with mileage must be calculated. Combining these
estimates will give the percentage of vehicles receiving
partial benefits. This analysis will assume that those
vehicles that crash or fail for non-oil control reasons in the
extension period receive one-half of the full 18 percent
benefit, or 9 percent. However, this can be accounted for by
assuming that one-half of the vehicles which fail in the
extension period receive the full benefit. Therefore, all that
is needed is to estimate the percent of vehicles crashing or
failing for non-oil control reasons in the extension period.
Using HHDVs as an example, the mileage to first rebuild is
266,114 miles (Table 5-9). Some trucks will travel 18 percent
farther until rebuild, or to 314,014 miles, which is 47,900
miles further and 0.8 years older. Data from MVMA show that
0.3 percent of trucks crash every year, so 0.24 percent of the
trucks crash in the extension period.[23] If 47 percent of
trucks are rebuilt for oil control reasons at 266,114 miles,
then the remaining 53 percent of trucks are rebuilt for non-oil
control reasons at 266,114 miles. Assuming the rate of non-oil
control failures in the extension period is the same as in the
first 266,114 miles, the percent of trucks failing for non-oil
control reasons in the extension period of 47,900 miles is 9.5
percent. Therefore, the total percent of trucks receiving a
half benefit is 9.74 percent, and if this is rounded to 10
percent and halved, the crashes and non-oil control failures
can be accounted for by subracting 5 percent from 47 percent,
so that 42 percent of all trucks receive a full 18 percent
benefit. This is the same as saying that 37 percent receive a
full benefit, and 10 percent receive a one-half benefit, and is
easier from a calculational perspective. Although this was
estimated for HHDVs, the same percentage can be applied to the
rebuild and scrappage mileages of the other vehicle classes,
since the extension benefit and the percentages of vehicles
crashing and failing for non-oil control reasons are both
mileage-based (i.e., a higher rate of non-oil control failures
in a lighter vehicle class will be balanced by a lower absolute
mileage extension).
-------
5-31
The results of the above analyses show that 42 percent of
vehicles will travel 18 percent further on low sulfur fuel
before rebuild and scrappage. These factors need to be applied
to the vehicle replacement and rebuild costs as developed in
the base case.
New fleet vehicle replacement costs are estimated by first
estimating replacement costs for the 42 percent of vehicles
traveling 18 percent further, then weighting the two vehicle
groups back together. These costs are shown in Table 5-12 in
the "Vehicles Benefiting," and "Fleet" costs. The last column
shows the difference in cost from the base case, which range
from 0.91 cents/mile for LDDVs to 1.78 cents/mile for MHDVls.
Next, rebuild costs for vehicles benefiting from sulfur
control must be estimated. These are shown for MHDV2s, HHDVs
and buses in Table 5-13. Rebuild costs and the fraction of
vehicles receiving each rebuild are assumed to be the same as
the base case. The sulfur control rebuild costs range from
0.91 cents per mile for buses to 1.52 cents per mile for
MHDVs. These are about 20 percent lower than the base case
costs. The total costs for the base and sulfur control cases
for all the vehicle types are summarized in Table 5-14. The
differences are shown in both cents/mile and cents/gallon of
fuel, the latter using model year 2000 current fuel
economies.[16,17] The benefits in cents/gallon range from 8.4
cents/gallon for HHDVs to 30.7 cents/gallon for LDDVs.
2. Reduction in Operating Costs - Increase in Engine
Rebuild Interval and Reduction in Number of Rebuilds
The prior methodology assumed that an increase in engine
rebuild interval also increased vehicle life. In this method,
an increase in engine rebuild interval leads to a reduction in
numbers of rebuilds, particularly second rebuilds for HHDVs.
The base case costs are exactly the same as in the
previous example. To predict the reduction in the number of
rebuilds, the Monte Carlo analysis was rerun without adjusting
vehicle lifetime mileages. The effect on the fraction of
second and third rebuilds is shown in Table 5-15 (the base case
fractions are also shown for comparison). Second rebuilds have
dropped about 10 percent for each vehicle type. The impact on
control case operating costs is shown in Table 5-16. Benefits
are much lower than the previous example because there is no
deferral of vehicle replacement costs. The benefits range from
1.98 cents/gallon for buses to 2.52 cents for HHDVs. Since
there is no extension of vehicle life, there are no benefits
for LDDVs, LDDTs and LHDTs in this method.
-------
5-32
Table 5-12
Vehicle Replacement Cost Comparison
Costs,
(cents/mile)
Sulfur
Control
Vehicles
Base
Difference
Vehicle Type
Benefitinq
Fleet
Case
Base-Control
LDDV
6.730/mi
7.98^/mi
8.89^/mi
0.91^/mi
LDDT
4.16
5.56
6.57
1.01
LHDT
7.38
9 .38
10.83
1.45
MHDV1
14.41
16.88
18. 66
1.78
MHDV2
5. 45
6.99
8.10
1.11
HHDV
7.58
9.27
10 . 50
1. 23
Buses
10 . 76
13. 16
14.90
1. 74
-------
5-33
Table 5-13
Sulfur Control Engine
Rebuild Costs for Vehicles Benefitting
Parameter MHDV2s HHDVs Buses
Lifetime Mileage (mi) 378,780 624,220 637,200
Vehicle Life (yrs) 14.2 10.6 14.2
Discounted Lifetime Mileage (mi) 213,280 384,326 392,317
First Rebuild
Mileage to First Rebuild (mi) 201,732 314,014 346,298
Age at Rebuild (yrs) 5.5 4.2 7.7
Second Rebuild
Mileage to Second Rebuild (mi) 351,244 566,545 578,385
Age at Rebuild (yrs) 12.3 9.1 12.9
Third Rebuild
Mileage to Third Rebuild (mi) 435,380 745,466 704,354
Age at Rebuild (yrs) 20 14.6 15.7
Lifetime Rebuild Cost (#/mi) 1.51 1.42 0.91
-------
5-34
Table 5-14
Operating Costs and Reductions
Vehicle Type Base
LDDV 8.89
LDDT 6.57
LHDT 10.83
MHDV 13.04
HHDV 12.28
Buses 16.19
Reduction
Control
tf/mile
<£/qal
7.98
. 91
30.0
5.56
1 .01
26.3
9 .38
1. 45
23 .2
11. 60
1.44
11.7
10 .90
1.23
8.4
14.29
1. 74
11.5
Fuel economies used:
LDDV - 33 mpg
LDDT - 26 mpg
LHDT - 16 mpg
MHDV - 8.1 mpg
HHDV - 6.8 mpg
Buses - 6.6 mpg
-------
5-35
Table 5-15
Percentage of Trucks Receiving
Rebuilds - Monte Carlo Analysis (COV = 25%)
Case
Vehicle
1st Rebuild
2nd Rebuild
3rd Rebuild
Base
MHDVs
92%
54%
8%
HHDVs
95%
62%
14%
Buses
93%
53%
9%
Control
MHDVs
91%
43%
7%
HHDVs
93%
51%
7%
Buses
90%
41%
6%
-------
5-36
Table 5-16
Operating Costs and Reductions
for Changes in Numbers of Rebuilds
Vehicle Type
Base
Control
tf/mile
£/qallon
MHDVs
13.04
12. 73
0 .313
2.51
HHDVs
12.28
11.91
0.37
2.52
Buses
16 .19
15 .89
0 .30
1 . 98
Fuel economics used:
MHDV - 8.1 mpg
HHDV - 6.8 mpg
Buses - 6.6 mpg
-------
5-31
References (Chapter 5)
1. "Control of Air Pollution From New Motor Vehicles
and New Motor Vehicle Engines; Gaseous Emission and Particulate
Emission Regulations", Federal Register, 10606, Friday, March
15, 1985.
2. "Locomotive Radioactive Ring Studies of Fuel
Lubricant, and Operating Variables", Tennyson, T.A. and C.K.
Parker (1970), SAE Paper H700892, Society of Automotive
Engineers, Warrendale, PA.
3. "Study of the Effects of Reduced Diesel Fuel Sulfur
Content on Engine Wear", EPA 460/3-87-002, June, 1987.
4. "Diesel Fuel Sulfur and Cylinder Liner Wear of a
Heavy Duty Diesel Engine", Weiss, E.K.J. , et. al., SAE Paper
872148, Society of Automotive Engineers, Warrendale, PA.
5. "Interrelation of Diesel Engine Lubricant Quality
and Sulfur Content of Diesel Fuel," W.C. Gergel, presented at
NPRA 1980 Fuels and Lubricants Meeting.
6. "Effect of Surface Temperature on Wear of
Disel-engine Cylinders and Piston Rings," H.V. Nutt, E.W.
Landen, and J.A. Edgar, SAE Transactions, Vol.63, 1955.
7. "Wear Prevention by Alkaline Lubricating Oils," J.C.
Ellis and J.A. Edgar, SAE Tranactions, Vol. 61, 1953.
8. Letter from Thomas Young, Executive Director, Engine
Manufacturers Association, to Charles Gray, Director, Emission
Control Technology Division, U.S. EPA, September 24, 1987.
9. "A Study to Determine Engine Rebuild Critieria among
Owners of Diesel Powered Vehicles", conducted by Survey Data
Research, Inc. for the Engine Manufacturers Association, May
10, 1981.
10. Caterpillar comments on ERC Report, Docket #A-86-03,
II-D-19.
11. API comments on ERC Report, Docket KA-86-03, II-D-11.
12. "The Effect of Piston Temperature and Fuel Sulfur on
Diesel Engine Piston Deposits", J.A. McGreehan, et.al., SAE
Paper 821216, October, 1982.
13. "Effect of Piston Deposits, Fuel Sulfur, and
Lubricant Viscosity on Diesel Engine Oil Consumption and
Cylinder Bore Polishing ", J.A. McGreehan, SAE Paper 831721,
November, 1983.
14. "Diesel Particulate Study," SDSB, ECTD, OMS, OAR,
EPA. November 1983. Available in Docket #A-80-18.
-------
5-32
15. "Motor Truck and Diesel Repair Manual", New York,
Motor, 1980.
16. "Heavy-Duty Vehicle Emission Conversion Factors II,
1962-2000", Paul A. Machiele, EPA-AA-SDSB-89-01, October 1988.
17. "MOBILE3 Fuel Consumption Model", Mark A. Wolcott,
U.S. EPA, OAR, OMS, and Dennis F. Rahlbaum, Computer Sciences
Corp., EPA-AA-TEB-EF-85-2, February 1985.
18. Fleet Equipment
19. Regulatory Impact Analysis for Heavy-Duty Engine
NOx/Particulate Final Rule, Available in Docket #A-80-18.
20. MOBILE4 Travel Characterization Data Handout,
MOBILE4 Workshop, November 1987. U.S. EPA, OAR, OMS, ECTD.
21. Lipson, C. and Shelth, N. , Statistical Design and
Analysis of Engineering Experiments, McGraw-Hill, New York,
1983.
22. EMA/MVMA Comments on EPA Analyses, addressed to
Charles L. Gray, Jr., September 24, 1987.
23. MVMA Motor Vehicle Facts and Figures 1987, Public
Affairs Division, Motor Vehicle Manufacturers Association.
-------
Chapter 6
Effect of Fuel Modifications on Air
Quality, Public Health, and Welfare
This chapter develops the emissions impacts for the fuel
control options, and assesses the resulting changes in air
quality (particulate concentrations), health effects (cancer
incidences) and welfare effects (visibility and S02
reductions). The emission reductions developed here are
further used in Chapter 7 to develop particulate cost
effectiveness estimates.
The chapter is divided into four sections. The first
section explains how emission inventories are estimated for
on-highway mobile, off-highway mobile and stationary sources,
and presents the urban emission reductions for the fuel control
scenarios. The second section presents the methodology for
predicting improvements in particulate and SO2 concentrations
in urban and rural areas. Included in this section is the
development of a methodology for estimating SO2 conversion to
sulfate particulate, and the resulting impact on air quality.
The third section assesses the impact of fuel controls on
cancer incidences. The fourth and final section estimates
changes in urban and rural visibility, and discusses some of
the implications of the SO2 reductions.
The methodologies used to develop particulate
concentrations, cancer incidences and visibility impacts draw
extensively on methods developed in the Diesel Particulate
Study and the Draft and Final RIAs for the heavy-duty
particulate standards.[1,2,3] Some modifications and
improvements have been made, however, and these are noted where
applicable.
I. Emissions Inventories
This section develops gaseous (SO2 and HC) and
particulate (carbon, SOF and sulfate) emissions inventories for
the base and control cases outlined in the introduction.
Emissions are estimated on an annual basis for on-highway
mobile, off-highway mobile, and stationary diesel and No. 2
fuel oil emission sources. An overview of these sources and a
description of which sources are affected by on-highway diesel
fuel controls for both maximum and minimum segregation is
presented first. This is followed by a general description of
the methodology used to estimate the emission inventories.
Next, since emission inventories depend on how much fuel is
consumed each year, a discussion of how future fuel consumption
was estimated is presented. Next, the emission factors of each
source are discussed. Finally, the emission inventories for
the base and fuel control cases are presented and discussed.
-------
6-2
A. Emission Sources
The controls discussed in this RIA would be applicable
only to on-highway diesel vehicles. However, as discussed in
Chapter 2, refiners may choose to treat a substantial portion
of the off-highway diesel fuel and No. 2 fuel oil also.
Sources potentially affected by diesel fuel controls other than
on-highway sources fall into three general categories: 1)
off-highway mobile sources, which includes construction and
agricultural equipment, 2) stationary diesel engines used as
generators and in other power equipment, and 3) No. 2 fuel oil
applications, which include commercial and industrial boilers
and residential furnaces. Rail sources, vessels, and military
applications, while they are off-highway mobile sources which
use No.2 diesel fuel, would be segregated under these controls
as discussed in Chapter 2, and are therefore excluded from this
emissions analysis. For the remainder of this report,
off-highway mobile will include only agricultural and
construction. Stationary diesel will include only those
sources burning No. 2 diesel.
B. General Inventory Estimation Methodology
An overview of the methods used to estimate emissions
inventories is as follows. First, for highway mobile sources,
emission rates in grams per mile are estimated for all diesel
vehicle types and ages. These have been developed in Chapters 3
and 4. (The emission rates for heavy duty trucks and buses
were developed in units of g/BHP-hr, but are converted to units
of g/mi in this chapter using BHP-hr/mi conversion factors.)
The emission rates by vehicle type and age are then combined
with registration and travel data to obtain fleet emission
rates by vehicle type in a particular calendar year. These are
then multiplied by vehicle miles traveled (VMT) data in each
calendar year to obtain total annual emissions for each vehicle
class. These VMT estimates are readily developed from
estimates of total annual fuel consumption by vehicle class,
and class specific fuel economy values. The emissions by class
are then summed to obtain total inventories for all diesel
mobile sources.
The emission rates for off-highway mobile, stationary
diesel and fuel oil sources are available from the literature
in units of grams of pollutant per gallon of fuel consumed.
All that is needed to estimate pollutant inventories for these
sources are estimates of fuel consumption by source category.
The emissions analysis in this chapter estimates emission
inventories on an annual basis for a variety of future calendar
years. The purpose in doing the analysis in this manner is to
evaluate the changes in emission reductions as old trucks are
replaced with new one with lower emission levels.
-------
6-3
The year of the most stringent emission standards for most
trucks is 1994, and generally, most if not all trucks are
scrapped before they are 20 years of age. Therefore, the 2015
calendar year emissions analysis for on-highway trucks includes
only the lowest emitting trucks (all higher emitting pre-1994
trucks are assumed to be scrapped), and so this year represents
a steady state condition in terms of emission standards. The
cost effectiveness methodology presented in the next chapter
however, utilizes a 33-year discounted analysis that requires
the estimate of emission reductions beyond 2015. This 33 year
analysis will start with the 1994 calendar year. Therefore, to
accommodate this analysis, emission reductions will be
estimated through the calendar year 2027.
Although the 33 year analysis will extend from 1994 to
2027, diesel fuel controls could be implemented before 1994, as
early as 1992. EPA is conducting a leadtime analysis on this
issue, however the results are not yet available. Therefore,,
emission inventories will be developed for 1992 as well as for
the years 1994-2027.
C. Fuel Consumption
Fuel consumption estimates are needed for the four sources
mentioned in the previous section: on-highway mobile sources,
off highway mobile sources, stationary diesel sources and fuel
oil sources. On-highway sources are discussed first, followed
by the other sources.
Estimates for diesel fuel consumption for on-highway
diesel vehicles were taken from the Motor Fuel Consumption
Model Thirteenth Periodical Report, hereinafter referred to as
the MFCM.[4] Estimates of diesel fuel consumption by vehicle
type for all classes except buses are available from the report
for calendar years 1980-2000 (only the data in years 1992-2000
were used in this analysis). Total non-bus on-highway diesel
fuel consumption ranges from 18.8 billion gallons per year in
1992 to 21.9 billion gallons per year in 2000.
One adjustment was made to the MFCM fuel consumption for
1992-2000. These estimates do not include fuel consumed by
buses. The techniques used to estimate bus fuel consumption
are explained in Appendix 6-A.
A number of techniques and sources of data were used to
obtain total annual VMT by vehicle type for the calendar years
199 0-2000, and these are described in Appendix 6-A. The
purpose of these calculations was to ensure consistency between
five commonly estimated and used parameters - fuel consumption,
VMT, fuel economy, the total numbers of vehicles in each class,
and annual VMT per vehicle. Generally, VMT for each class for
1992-2000 was determined by dividing class specific fuel
consumption by class specific fuel economy estimates. This
-------
6-4
resulted in a 2.75 percent compound annual increase in VMT for
all diesels between 1992 and 2000. This was rounded to 3
percent and applied in a compound fashion for the 2000-2027
calendar years, since the rate of growth was rising between
1992 and 2000. Fuel consumption beyond 2000 could then be
estimated for calendar years 2000-2027 by dividing VMT by fuel
economy. The 3 percent compound VMT growth rate resulted in
about a 2.5 percent compound annual increase in total fuel
consumption over the entire period of 1992-2027. The specific
procedures used in this analysis are discussed in detail in
Appendix 6-A.
The foregoing discussions developed on-highway diesel fuel
consumption and diesel VMT for 1992 through 2027. Fuel
consumption estimates for off-highway mobile sources,
stationary diesel and fuel oil sources for the same time period
are also needed.
As discussed in Chapter 2, the minimum (NPRA) segregation
volume of fuel controlled includes on-highway, off-highway
mobile, stationary diesel and fuel oil sources. With the
on-highway fuel consumption developed above, the sum total of
off-highway mobile, stationary diesel and fuel oil consumption
can be estimated by subtracting the on-highway fuel consumption
from the total volume of fuel controlled under the minimum
segregation scenario. For 1990 the volume of fuel controlled
with minimum segregation is 37.2 billion gallons. The
on-highway demand (from Appendix 6-A) is 20.2 billion
gallons. Thus, for 1990, the volume of fuel consumed by
off-highway mobile, stationary diesel and fuel oil sources is
about 17 billion gallons. This consumption must be properly
allocated between the different sources, and projections of
fuel consumption must be made for the years 1992-2027.
Table 2-2 in Chapter 2 lists the percentages of middle
distillate fuel used by the emission sources and these
percentages can be used (after removing military, railroad and
vessel bunkering, and renormalizing) to allocate the non
on-highway fuel consumption into the different categories
(off-highway mobile, stationary diesel and fuel oil). When
this is done, about 24 percent of the off-highway mobile plus
stationary diesel plus fuel oil is off-highway mobile, 20
percent is stationary diesel, and the remainder (56 percent) is
fuel oil. However, as will be shown shortly in the emission
rate analysis, off-highway mobile and stationary diesel engines
have similar emission rates that can be combined to form one
source. Thus, this analysis will combine the estimates from
these sources, and refer to it as "other diesel" fuel
consumption. The resulting percentage for this category of non
on-highway diesel is 44 percent (24 + 20). The 1990 projected
volume of fuel consumed by "other diesel" and fuel oil sources,
estimated by the above methodology is 7.5 and 9.5 billion
gallons, respectively. The only item remaining is to
-------
6-5
estimate fuel consumption for these categories in other
calendar years. The VMT growth rates developed for on-highway
mobile sources resulted in about a 2.5 percent compound annual
increase in on-highway fuel consumption. This percentage was
applied to the other diesel and fuel oil categories as well.
This is consistent with growth rates used for off-highway
emissions in the past. [2,3] A summary of all fuel consumption
values for calendar year 1990-2027 thus obtained for the NPRA
segregation case obtained is shown in Table 6-1.
D. Emission Factors
This section describes in more detail the emission rates
used for the different sources and the methodologies for
estimating these emission rates. The situation is most
complicated for on-highway sources since the emission rates
change by model year, and thus, vehicle scrappage and travel
must be accounted for to properly estimate emission rates by
vehicle type. For the other sources, the emission rates do not
change from year-to-year, other than with fuel composition, so
estimating inventories for these sources is easier.
The emission rates and methodologies for on-highway mobile
sources are discussed first, followed by the other diesel
emissions (off highway mobile and stationary diesel) and
finally, fuel oil sources.
1. Qn-Hiqhway Mobile Sources
Emission rates from the six vehicle types (LDDVs, LDDTs,
LHDTs, MHDVs, HHDVs, and buses) for base and control fuels were
developed in Chapters 3 and 4, with the exception of emission
rates for vehicles with trap failures which will be discussed
presently. The LDDV and LDDT emissions are in units of grams
per mile, and can be multiplied by class-specific VMTs to
obtain inventories. However, the other classes are presented
in units of g/BHP-hr, and must be converted to units of g/mi.
The conversion factors used to do this are those developed for
MOBILE4, and are listed in Table 6-2.[5] The GVWR
class-specific conversion factors in the referenced report were
weighted by nationwide VMT to obtain similar factors for the
somewhat broader vehicle class categories used here. For
example, this analysis groups classes 6-8a in the above report
into the single class of MHDVs.
The emission rates developed in Chapters 3 and 4 were
end-of-life emission rates for fleets of vehicles, some
including traps, with no trap failures. As developed in the
supporting analyses for the diesel particulate standards
[1,2,3], trap failures may occur because of failure of
electronic controls or due to unforeseen operating conditions.
When trap failure occurs, the vehicle is assumed to be emitting
at its engine-out emission levels. (This applies to HC and
-------
6-6
Table 6-1
Middle Distillate Fuel Consumption
by Source (10-LQ gal per year)
Calendar Year
Hiqhway
Other
Fuel Oil
Total
1990
1. 97
0.768
0 .978
3.716
1991
1.99
0.784
1.007
3.781
1992
2.01
0 .799
1.037
3.847
1993
2. 03
0.815
1.068
3.914
1994
2. 05
0.832
1.101
3 .983
1995
2.08
0.848
1. 134
4.062
1996
2. 12
0.865
1.168
4 .153
1997
2. 16
0.882
1.203
4 .246
1998
2.2
0.900
1.239
4 .339
1999
2.25
0.918
1.276
4 .445
2000
2.31
0.936
1.314
4.561
2001
2.35
0.955
1.354
4.659
2002
2. 41
0.974
1.394
4 . 779
2003
2.46
0.994
1.436
4 .890
2004
2. 53
1.014
1.479
5. 023
2005
2. 59
1. 034
1.524
5.148
2006
2 . 66
1. 055
1.569
5.285
2007
2 . 74
1.076
1.616
5.466
20 08
2.81
1. 097
1. 665
5.573
2009
2.89
1.119
1. 715
5.725
2010
2.97
1. 142
1. 766
5.879
2011
3 . 06
1. 165
1 .819
6.044
2012
3. 15
1. 188
1 .874
6.212
2013
3. 24
1.212
1.930
6 .382
2014
3.34
1.236
1.988
6. 564
2015
3.43
1. 261
2.048
6. 739
2016
3 . 54
1.286
2. 109
6 .935
2017
3 . 64
1.311
2. 173
7 . 125
2018
3 . 75
1.338
2.238
7 .326
2019
3.86
1.364
2 .305
7.530
2020
3.98
1.392
2.374
7.746
2021
4 . 09
1.420
2.445
7.955
2022
4.22
1.448
2.519
8 . 187
2023
4.34
1. 477
2.594
8.412
2024
4 . 47
1. 507
2.672
8.649
2025
4 . 6
1. 537
2 .752
8 .889
2026
4 . 74
1. 567
2.835
9 . 143
2027
4 . 88
1. 599
2 . 920
9 .399
-------
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
6-7
Table 6-2
Heavy-Duty Diesel Conversion Factors (BHP-hr/mi)[5]
LHDVS
MHDVs
—
2.72
—
2.77
—
2.81
—
2.85
—
2.87
—
2.83
—
2.82
—
2 . 82
—
2.72
—
2.57
—
2.41
. 942
2.39
. 923
2.31
. 922
2.44
. 921
2.40
.919
2.44
. 919
2.32
. 919
2.32
.919
2 .32
. 919
2.31
. 919
2.31
.919
2.31
.919
2.31
.919
2.30
.919
2.30
HHDVs
Buses
3 .20
3.07
3 .27
3 . 00
3 .27
3.19
3 .35
2.99
3 .35
2.82
3.30
2.93
3 .35
2.80
3 . 40
2.84
3.36
2.84
3.31
2.83
3 .33
2.83
3.26
2.88
3 . 15
2.81
3. 15
2.86
3. 14
2.62
3 . 13
2.56
3. 13
2.56
3.13
2.55
3.13
2.55
3 .13
2.53
3.13
2.52
3 .13
2.51
3.13
2.50
3.13
2.49
3 .13
2 . 48
-------
6-8
carbon and SOF particulates - there is little or no effect on
SO2 or sulfate particulate,) To estimate fleet emission
rates, the fraction of trap failures must be estimated, and
emissions of vehicles with operating traps and failed traps
weighted together. This analysis uses a two percent annual
trap failure rate estimate, which is the same as the analysis
for the final diesel particulate standards. Engine-out
emission levels used for trap-failed vehicles were developed in
Chapter 3 and 4.
The registration and travel data used in this analysis are
shown in Tables 6-3, 6-4, and 6-5. The registration by age and
mileage accumulation rates by age data comes from MOBILE4.[6]
The diesel sales fractions were developed from data available
in the MFCM.[4] Registration fractions, VMT, and diesel sales
fractions in each model year are multiplied, and the product is
summed over all the model years within a calendar year and
normalized to obtain model year travel fractions. The model
year emission factors are then weighted by these travel
fractions to obtain class-specific emission rates by vehicle
types for each calendar year. These emission rates can then be
multiplied by estimates of class specific nationwide annual VMT
to yield annual emission inventories.
Lastly, to obtain estimates of urban, as opposed to
nationwide, emissions, the VMT estimates by class developed in
previous sections are multiplied by the fractions of urban
travel by class. These are obtained from the conversion factor
analysis for MOBILE4 and are shown in Table 6-6.[4]
2. Off-Highway Mobile and Stationary Diesel Sources
Other sources which use diesel engines are agricultural
equipment, construction equipment, and stationary diesel
engines used in generating electricity and operating machinery
(such as oil drilling rigs).
The emissions of interest in this analysis are
particulates, hydrocarbons, and sulfur dioxide. The analysis
of HC and particulate emissions in Chapter 4 showed that HC and
particulate emissions vary with fuel aromatics content under
transient conditions, but do not vary under steady-state
conditions. Most of the diesel engines in these off-highway
categories are operated under steady-state conditions,
therefore, HC emission from these sources were not estimated at
all. SO2 emission rates are a function of sulfur content and
sulfate conversion in the engine and exhaust system. This
analysis uses a two percent S02~to~sulfate conversion rate,
which is the same rate used for on-highway diesel engines in
Chapter 4. Using a diesel fuel density of 7.1 lbs per gallon,
and current and low sulfur percentages of .25 and .05 wt
percent, the SO2 emission rate for non-highway diesel engines
on current sulfur fuel is 15.8 g/gal, and for low sulfur fuel
is 3.15 g/gal.
-------
6-9
Table 6-3
Vehicle Registration Distributions by Age
Vehicle
Age
LDDVs
LDDTs
All
Heavy-Duty
Trucks*
Buses
l
0.062
0 . 070
0.082
. 077
2
0 . 082
0 . 092
0. 165
. 077
3
0 .079
0. 088
0. 135
. 076
4
0 . 075
0.083
0. Ill
. 074
5
0 . 071
0 . 077
0.091
. 072
6
0 . 067
0. 072
0.075
. 070
7
0 . 063
0.067
0.061
.068
8
0.060
0.063
0.050
.065
9
0 . 056
0 .057
0.041
. 062
10
0.052
0 .051
0.034
.059
11
0.048
0 . 047
0 . 028
. 055
12
0.045
0 . 041
0.023
. 050
13
0 . 041
0 . 036
0 .019
. 043
14
0 . 037
0 .031
0.015
.034
15
0 . 033
0 . 026
0.013
.026
16
0 . 029
0 . 021
0 . 010
.020
17
0.026
0. 016
0 . 009
.016
18
0 . 022
0 .011
0 , 007
.013
19
0 . 018
0 . 007
0 . 006
.011
20+
0 . 034
0 . 044
0 . 024
. 032
Classes 2b-8.
-------
6-10
Table 6-4
Mileage Accumulation Rates by Age (miles per year)
Vehicle
Age
LDDV
LDDT
LHDV
MHDV
HHDV
Buses
1
17,825
20,140
23,611
43,946
86,375
45,000
2
16,475
17,572
20,947
40,504
79,434
45,000
3
15,233
15,434
18,583
37,332
73,051
45,000
4
14,081
13,639
16,486
34,408
67,408
45,000
5
13,017
12,133
14,625
31,713
61,782
45,000
6
12,033
10,863
12,975
29,229
56,817
45,000
7
11,124
9,788
11,511
26,939
52,252
45,000
8
10,283
8,877
10,212
24,829
48,053
45,000
9
9,506
8,103
9,059
22,885
44,191
45,000
10
8,788
7,444
8,037
21,092
40,640
45,000
11
8, 123
6,883
7,130
19,440
37,374
45,000
12
7,509
6,405
6,325
17,918
34,371
45,000
13
6,942
5,999
5,612
16,514
31,609
45,000
14
6,417
5,655
4,978
15,221
29,069
45,000
15
5,932
5,365
4,416
14,029
26,733
45,000
16
5,484
5,123
3,918
12,930
24,585
45,000
17
5, 069
4,924
3,476
11,917
22,609
45,000
18
4,686
4,763
3,084
10,984
20,792
45,000
19
4,332
4,637
2,736
10,123
19,121
45,000
20 +
4,005
4,543
2,427
9,331
17,585
45,000
-------
6-11
Table 6-5
i
Annual Diesel Vehicle Sales Fractions
Model
Year
LDV
LDT
LHDV
MHDV
HHDV
Buses
1970
0.0
0.0
0.0
. 160
.925
. 082
1971
. 001
0.0
0.0
. 160
.923
. 122
1972
.002
0.0
0.0
. 145
.923
. 139
1973
.002
0 , 0
0.0
. 130
.921
. 1145
1974
.003
0.0
0.0
. 150
.920
. 196
1975
.003
0.0
0.0
. 170
.920
.209
1976
. 003
. 001
0.0
. 191
. 960
. 212
1977
.003
.011
0.0
.249
1. 00
. 129
1978
.009
. Oil
0.0
.307
1.00
. 200
1979
.026
. 013
0.0
.365
1. 00
.230
1980
. 045
. 056
.041
.409
1. 00
.308
1981
.060
. 051
.081
. 454
1. 00
.473
1982
.039
.071
. 122
.500
1.00
. 451
1983
.019
. 074
. 183
.543
1. 00
. 509
1984
.014
. 037
. 198
.578
1 .00
.473
1985
.008
. 015
.215
. 633
1 .00
. 662
1986
.003
. 019
.232
.575
1. 00
.743
1987
.004
.019
.250
.582
1. 00
.791
1988
.004
.021
.260
. 593
1.00
.832
1989
.004
.021
.270
.605
1.00
.861
1990
.005
.023
.280
.613
1.00
.886
1991
. 006
.023
.290
.629
1.00
.907
1992
.006
.023
.300
.639
1.00
.926
1993
. 007
.023
.300
.648
1.00
.942
1994
. 007
. 023
.300
.662
1.00
. 956
1995
. 008
. 023
.300
. 670
1.00
. 968
1996
. 009
.023
.300
.680
1 . 00
. 977
1997
. 009
.023
.300
.689
1. 00
. 986
1998
. 010
.023
.300
.690
1. 00
. 999
1999
. 010
.023
.300
.690
1.00
1 . 00
2000 +
, Oil
.023
.300
. 691
1 . 00
1. 00
-------
6-12
Table 6-6
Urban Travel Fractions
All Years
LDDV LDDT LHDDT MHDDT HHDDT Buses
0.6
0.45
0.26
0 . 55
-------
6-13
Particulate emission rates for these off-highway diesel
engine sources are shown in Table 6-7. The source of the
particulate data is the Fourth Edition of AP-42, Volume l.[7]
Particulate emissions range from 11.8 g/gal for construction
equipment to 23.3 g/gal for non-tractor agricultural
equipment. The emission rates for agricultural equipment and
stationary diesel engines are very similar. Due to the
similarity in emission rates, an arithmetic mean emission rate
was estimated for all of the sources, which is shown under the
dashed line in Table 6-7. The non-highway diesel particulate
average (agricultural, construction and stationary) is 17.6
g/gal, and this is used for all off-highway mobile and
stationary diesel engines. However, these are total
particulate emissions, and must be split into the various
particulate types. Using a two percent sulfur to sulfate
conversion rate, sulfate emissions for a 0.25 wt percent sulfur
fuel are l.l g/gal fuel consumed. Data from pre-1988 engines
in Chapter 3 shows that of carbon plus SOF emissions, about 75
percent is carbon and 25 percent is SOF. Applying these
fractions to the difference between total and sulfate
emissions, the result is 12.3 g/gal for carbon emissions and
4.1 g/gal for SOF emissions. On a low sulfur fuel, only the
sulfate particulate are affected. A drop of 80 percent (from
0.25 weight percent to 0.05 weight percent) in sulfur content
leads to an 80 percent drop in sulfate emissions. Therefore,
the sulfate emission rate of non-highway diesel engines on low
sulfur fuel is 0.22 g/gal.
To obtain urban emissions for these sources, an urban
fraction for this category is needed. Data on urban fuel
consumption by these sources is very hard to find. However, it
is likely that agricultural fuel consumption is almost entirely
rural, and that construction and stationary diesel fuel
consumption are almost entirely urban. There may be some
overlap in these categories, however. Assuming that only
construction and stationary diesel consumption is urban and
using the renormalized consumption percentages from Table 2-2
(that is, factoring out railroad, vessel bunkering and military
use), the urban fraction for these categories is 0.63. Thus,
urban "other diesel" emissions are estimated to be 63 percent
of total nationwide other diesel emissions.
3. Fuel Oil Sources
Sources which use fuel oil are commercial and industrial
boilers, and residential furnaces. Some commercial and
industrial furnaces use residual oil, but the emissions of
these sources are not affected by these middle distillate fuel
controls. According to AP-42, the fuel oil sources emit only
about 2 lbs of non-sulfate particulate matter per 1000 gallons
of diesel fuel, and this is probably the result of the very
high air to fuel ratios used in burners. [7] This is only about
one-tenth of the particulate emission rates of diesel engines.
-------
6-14
Table 6-7
Particulate Emission Factors for
Off Highway Mobile and Stationary Diesel Sources
Category (g/gal)
Agricultural
Tractor 20.8
Non-tractor 23.3
Average 22.0
Construction
Average of 10 types 11.8
Stationary Diesel
Industrial 15.2
Large Bore 22.7
Average 18.9
Non-Highway Diesel (All 3 above combined) Averages
Total 17.6
Carbon 12.3
SOF 4.1
Sulfate 1.1
-------
6-15
Also, these sources are operated under steady-state conditions,
and so the carbon or SOF emission emitted by these sources are
not expected to be affected by changes in fuel quality. With
respect to sulfate emissions, AP-42 reports that sulfur to
sulfate conversion is on the order of 1-5 percent, so a value
of two percent will be used to be consistent with the other
sources. At two percent, sulfate emissions on 0.25 weight
percent sulfur fuel are 1.13 g/gal, and 0.226 g/gal on a low
sulfur fuel. SO2 emission rates are 16.1 g/gal on high
sulfur fuel, and 3.2 g/gal on low sulfur fuel.
Urban fuel oil consumption is also difficult to estimate
accurately. In the midwest, much of the fuel oil for
residential furnaces is consumed in rural areas since the urban
areas are served by natural gas pipelines. In the northeast,
however, fuel oil is burned even in the urban areas.
Therefore, this analysis will assume that urban fuel oil
consumption is proportional to urban population. According to
1980 census data, 61 percent of the U.S. population lives in
urban areas of 50,000 and greater.[8] Thus, the urban to
nationwide fraction of fuel oil consumed is estimated to be
61 percent.
E. Emission Inventory Results
The changes in total urban particulate and gaseous
emission inventories due to sulfur and aromatics control are
shown in Tables 6-8 and 6-9, respectively. Nationwide
reductions are shown in Tables 6-10 and 6-11. Rural emission
reductions can be obtained by subtracting the urban from the
nationwide estimates. The rural estimates are used in later
analyses to estimate the effects of fuel controls on rural
visibility. However, the following discussions will center
only on urban emission reductions (Tables 6-8 and 6-9). The
emission changes in Table 6-8 and 6-9 due to aromatics control
are incremental to sulfur control. Total emission inventories
for the sources addressed here are contained in Appendix 6-B.
In Table 6-8, SOF, carbon, directly emitted sulfates, and
indirect sulfates are shown separately, as well as the sum,
total PM. Direct sulfates are those emitted at the tailpipe of
a diesel vehicle; indirect sulfates are those formed from the
conversion of middle distillate-derived SO2 to sulfate
compounds. The extent of conversion of SO2 to sulfates is
developed in the next section (Indirect Sulfate Analysis), but
the results are used here. To summarize the next section,
about 12 percent of urban SO2 is converted to sulfates within
the urban atmosphere (mostly ammonium sulfate). Put another
way, for every ton of SO2 emitted, conversion of SO2 in
urban areas to sulfates results in about 0.288 tons of
sulfates. The quantity of these "indirect" sulfates far
outweighs the directly emitted sulfates.
-------
6-16
Table 6-8
Annual Urban Particulate Emission
Reductions (tons per year)
Particulate
Calendar
Year
Sulfur
Control
Aroraatics
Control
100 %
NPRA
100 %
NPRA
Sulfates
1992
7, 500
18,344
-2
-2
(direct)
1995
7,630
19,336
-10
-10
2000
8,230
21,544
-32
-32
2005
9,322
24,461
-42
-42
2010
10,734
27,951
-50
-50
2015
12,440
32,028
-59
-59
Sulfates
1992
30,203
74,189
0
0
(indirect)
1995
31,856
79,324
0
0
2000
36,506
90,425
0
0
2005
41,920
103,203
0
0
2010
48,439
118,132
0
0
2015
56,172
135,473
0
0
SOF
1992
-126
-126
3,212
3,212
1995
-1,033
-1,032
2,172
2,172
2000
-907
-907
1,023
1,023
2005
-979
-979
722
722
2010
-1, 144
-1,144
681
680
2015
-1,289
-1,289
761
761
Carbon
1992
-711
-711
2,391
2,391
1995
-3,101
-3,101
1,508
1,508
2000
-5,689
-5,688
320
320
2005
-7,208
-7,208
-42
-68
2010
-8,570
-8,569
-214
-214
2015
-9,969
-9,969
-281
-281
Total Direct
1992
6,663
17,507
5,601
5,601
PM
1995
3,496
15,203
3,670
3,670
2000
1,634
14,949
1,311
1,311
2005
1, 135
16,274
612
612
2010
1,020
18,238
417
417
2015
1,182
20,770
421
421
Total PM
1992
36,866
91,696
5,601
5,601
1995
35,352
94,526
3,670
3,670
2000
38,140
105,373
1,311
1,311
2005
43,055
119,479
612
612
2010
49,459
136,369
417
417
2015
57,354
156,244
421
421
-------
6-17
Table 6-9
Type
Urban
SO?, HC and
CO Emission
Reductions
Control
Calendar
Year
Sulfur
Control
Aromatics
100 %
NPRA
100 %
NPRA
so2
1992
104,873
257,599
0
0
1995
110,612
275,430
0
0
2000
126,756
313,976
0
0
2005
145,556
358,346
0
0
2010
168,190
410, 180
0
0
2015
195,042
470,396
0
0
HC
1992
-3,189
-3,189
18,125
18,125
1995
-8,996
-8,996
14,946
14 ,946
2000
9,509
9,509
8,217
8,217
2005
18,829
18,829
6,684
6,684
2010
23,692
23,692
7,171
7,171
2015
29,530
29,530
8,133
8,133
CO
1992
26,215
26,215
24,077
24,077
1995
67,698
67,698
23,998
23,998
2000
176,810
176,810
26,424
26,424
2005
234,110
234,110
28,878
28,878
2010
279,883
279,883
32,696
32,696
2015
328,132
328,132
37,697
37,697
-------
6-18
Table 6-10
Nationwide Particulate Emission Reductions
(tons)
Particulate
Calendar
Year
Sulfur
Control
Aromatics
Control
100 %
NPRA
100 %
NPRA
Sulfates
1992
21,202
38,980
-2
-2
(direct)
1995
21,648
40,839
-23
-23
2000
23,276
45,105
-79
-79
2005
26,417
51,242
-103
-103
2010
30,422
58,656
-122
-122
2015
35,267
67,390
-143
-143
Sulfates
1992
84,928
157,034
0
0
(indirect)
1995
89,352
167,168
0
0
2000
101,291
189,694
0
0
2005
116,380
216,844
0
0
2010
134,472
248,723
0
0
2015
155,995
285,998
0
0
SOF
1992
-355
-355
8,425
8,425
1995
-2,738
-2,738
5, 650
5,650
2000
-2,268
-2,268
2,552
2,552
2005
-2,395
-2,375
1, 756
1,756
2010
-2,789
-2,789
1, 628
1,628
2015
-3,127
-3,127
1,810
1,810
Carbon
1992
-2,112
-2,112
6,865
6,865
1995
-9,284
-9,284
4,258
4,258
2000
-16,784
-16,784
728
728
2005
-21,217
-21,217
-403
-403
2010
-25,213
-25,213
-860
-860
2015
-29,377
-29,377
-1,097
-1,097
Total Direct
1992
18,735
36,513
15,288
15,288
PM
1995
9,626
28,817
9,885
9,885
2000
4,224
26,053
3,201
3,201
2005
2,805
27,630
1, 274
1,274
2010
2,320
30,654
646
646
2015
2,763
35,426
570
570
Total PM
1992
103,662
193 ,547
15,288
15,288
1995
98,978
195,984
9,885
9,885
2000
105,516
215,738
3,201
3,201
2005
119,186
244,476
1,250
1,250
2010
136,891
279,376
646
646
2015
158,798
320,924
570
570
-------
6-19
Table 6-11
Type
Nationwide
SO?, HC and CO Emission
Reductions
Control
Calendar
Year
Sulfur
Control
Aromatics
100 %
NPRA
100 %
NPRA
so2
1992
294,888
545,260
0
0
1995
310,251
580,443
0
0
2000
351,708
658,626
0
0
2005
404,096
752,932
0
0
2010
466,916
863,622
0
0
2015
541,649
993,047
0
0
HC
1992
-9,422
-9,422
46,421
46,421
1995
-29,856
-29,856
37,908
37,908
2000
25,382
25,382
18,082
18,082
2005
54,827
54,527
13,173
13,173
2010
69,674
69,674
13,462
13,462
2015
87,979
87,979
14 ,956
14,956
CO
1992
76,080
76,080
71,077
71,077
1995
182,257
182,257
69,488
69,488
2000
495,757
495,757
69,339
69,339
2005
654,388
654,388
74,999
74,999
2010
781,347
781,347
84,854
84,854
2015
916,293
916,293
97,777
97,777
-------
6-20
The two segregation cases are shown in both tables - 100
percent segregation, which represents refinery control of only
on-highway vehicle fuel, and NPRA segregation, which represents
refinery control of nearly all middle distillate. Emission
reductions are represented by positive values. Increases in
emissions are represented by negative values.
Examination of the data in Table 6-8 reveal two general
items. First, changes in SOF and carbon particulate emissions
are identical for both 100 percent segregation and NPRA
segregation because only on-highway engines are affected by
changes in fuel aromatic content. Changes in sulfate emissions
(both direct and indirect) with aromatics control only arise
from changing control technology of on-highway diesels. Thus,
such sulfate emission changes only occur with on-highway
diesels.
1. Sulfur Control
Emission changes due to sulfur control are shown in the
first two columns. The reductions in directly emitted sulfates
for the 100 percent segregation case range from 7,500 tons in
1992 to 12,440 tons in 2015. The increase in reductions is due
to an increase in fuel consumption. Reductions in sulfate
emissions for NPRA segregation range from 18,000 tons in 1992
to 32,000 tons in 2015. The differences between the 100
percent and NPRA cases represent reductions from other diesel
and fuel oil sources.
Reductions in indirect sulfates for the maximum
segregation case range from about 30,000 tons in 1992 to about
56,000 tons in 2015. These reductions are the result of
eliminating much of the urban SO2 emitted by on-highway
sources. For NPRA segregation, the reductions range from
74,000 tons in 1992 to 135,000 tons in 2015.
SOF emissions from on-highway vehicles increase with
sulfur control. The general upward trend in SOF emissions is
due to newer vehicles (1991-93 and 1994+ heavy-duty vehicles),
which, because of the reduction in direct sulfate particulate,
can emit more SOF and carbon emissions in meeting the same
particulate emissions standards. There is a larger increase in
SOF emissions for 1991-93 trucks than 1994+ trucks. This is
what causes SOF to temporarily peak in 1995 and drop somewhat
in 2000. For example, the increase in SOF emissions for
1991-93 HHDVs is about 0.0119 b/BHP-hr, which is a factor of
4.5 times the 1994+ HHDV SOF increase due to sulfur control of
0.0027 g/BHP-hr. After calendar year 2000, SOF emissions rise
again due to the increase in fuel consumption.
Carbon particulate emissions show distinct increases with
sulfur control, similar in direction but greater in magnitude
than SOF emissions. The reason for the increase in carbon
-------
6-21
emissions is the same as for SOF emissions: with sulfur
control and emissions averaging (as discussed in Chapter 3),
some HDVs do not need traps and others with traps can meet the
emission standard with higher carbon emissions.
In using these results, particularly the projected
increases in SOF and carbon PM emissions, it should be
remembered that the analyses of Chapter 3 and 4 assumed that
the 0.1 g/BHP-hr PM standard could be met with high sulfur
fuel. If this were not the case, SOF and carbon PM emissions
could be much higher without fuel sulfur control. Then, sulfur
control would show significant reductions in these emissions,
Direct PM includes direct sulfates, carbon and SOF.
Reductions in total direct PM for 100 percent segregation range
from 6,663 tons in 1992 to 1,182 tons in 2015. They drop
because the sulfate reductions are balanced by increases in
carbon and SOF. They do not go to zero, however, because there
are always some net PM reductions (all sulfates) from LDDVs and
LDDTs. Reductions in total direct PM emissions for NPRA
segregation range from 17,000 tons in 1992 to almost 21,000
tons in 2015.
Reductions in total PM emissions for 100 percent
segregation range from nearly 37,000 tons in 1992 to 57,000
tons in 2015, and for NPRA segregation range from 92,000 tons
in 1992 to 156,000 tons in 2015. These reductions are
dominated by the indirect sulfate emission reductions.
Turning momentarily to Table 6-9 to complete the
discussion on sulfur control, urban emission reductions of
SO2 for 100 percent segregation range from 105,000 tons in
1992 to 195,000 tons in 2015. These reductions are based on
SO2 inventories for the base and control cases before
conversion of some of the SO2 in the atmosphere to sulfate
compounds. The growth in reductions is due to the growth in
fuel consumption over the same period. For NPRA segregation,
SO2 reductions range from 258,000 tons in 1992 to about
470,000 tons in 2015. These reflect additional reductions in
SO2 from other diesel and fuel oil sources.
Urban HC emissions increase in 1992 and 1995 with sulfur
control, but decrease later. The changes range from an
increase of 8,996 tons in 1995 to a reduction of almost 30,000
tons in 2015. The near term increases are due to the absence
of aftertreatment on 1991-93 trucks and buses with sulfur
control. If trap technology is not available until 1994,
sulfur control will not affect the HC emissions of 1991-93
engines and this temporary increase will not occur. The
ability to use oxidation catalysts on some 1994+ trucks and
buses brings significant HC reductions in later years.
-------
6-22
Urban CO emission reductions range from 26,000 tons in
1992 to 328,000 tons in 2015. In 1992 the emission reduction
is entirely due to a 5.7 percent decrease in engine-out CO
emissions from the in-use fleet. In the later years, sulfur
control allows the use of oxidation catalysts, providing higher
percentage reductions in CO emissions.
2. Aromatics Control
Table 6-8 indicates that directly emitted sulfate
emissions increase slightly for the 100 percent segregation
case (aromatics control has no effect on either SO2 or
indirect sulfate emissions). The increases are negligible when
compared to the sulfate reductions due to sulfur control. The
cause of the increases can be traced to slight increases in
sulfate emissions of some of the heavy-duty vehicle types due
to shifts from trap technology to oxidation catalysts (which
oxidizes slightly more exhaust SO2 to sulfates). The
increases in sulfate emissions are identical for the NPRA
segregation case, since they occur only in on-highway vehicles
in either case.
SOF emissions are clearly lower with aromatics control,
especially in the earlier years, where the entire in-use fleet
experiences a reduction in SOF emissions. The reductions
decline because not as many 1994+ engines will need traps with
aromatics control, thus resulting in the same average tailpipe
emissions. The reductions do not go to zero, though, because
there is always some benefit of aromatics control on LDDVs,
LDDTs, and HDVs with failed traps.
Carbon particulate emissions show reductions in the
earlier years and increases in later years. The reasons are
the same as for SOF emissions. The net increases in carbon PM
emissions are due to greater absolute growth in fuel
consumption in the HDV classes as opposed to the LDDV and LDDT
classes.
Turning again to Table 6-9, there are no SO2 reductions
due to aromatics control, but there are significant HC
reductions. They range from about 18,000 tons in 1992 to about
8,000 tons in 2015. The decrease in HC reductions is brought
about by newer trucks ( 1994+), in which the differences in HC
emissions due to aromatics control is less than for in-use cars
and trucks. CO emission reductions due to aromatics control
range from 24,000 tons in 1992 to about 38,000 tons in 2015.
II. Effects of Emission Changes on Pollutant Ambient
Concentr at ions
The previous section developed urban and nationwide
emission reductions in tons due to diesel fuel sulfur and
aromatics control. The purpose of this third section is to
-------
6-23
determine the effect those emission reductions will have on the
ambient concentrations of these pollutants. The improvements
or changes in air quality are important results in and of
themselves, but they are further used in a subsequent section
of this chapter to predict changes in visibility, and are also
used in a later chapter to estimate other welfare benefits
(reductions in materials and crops damage) due to diesel fuel
controls.
Urban ambient concentrations are developed for five
pollutants - carbon particulate, sulfate particulate, total
particulate (SOF+carbon+sulfates), SO2 and ozone. Rural
ambient concentrations are developed for the same pollutants
with the exception of ozone. Changes in concentrations of
carbon and sulfate particulates due to fuel controls are used
in Section IV of this of this chapter to develop urban and
rural visibility impacts of fuel controls. Changes in
concentrations of total PM, SO2 and ozone due to fuel
controls are used in Chapter 8 to develop welfare benefits.
One of the most significant factors in this analysis that
affects the ambient concentrations of both SO2 and sulfate
particulate is the extent of the reaction of SO2 to form
sulfate particulate in the atmosphere. It is known that SO2
reacts with other species in the atmosphere to form various
sulfate species like ammonium sulfate and sulfuric acid. The
reaction rate is affected by many factors, such as meteorology
(sunlight, temperature) and the presence of other reactant
species. About 98 percent of the sulfur burned in diesel
engines reacts to form SO2/ only two percent react somewhere
in the engine or exhaust system to form directly emitted
sulfates. Therefore, if much of the SO2 emitted by diesel
engines reacts in the urban atmosphere to form sulfate
particulate, urban particulate reductions due to fuel controls
can be large. For example, if only 10 percent of the SO2
emitted by diesel engines reacts to form sulfate particulate in
urban areas, the amount of sulfate particulate formed in the
atmosphere is five times the amount of particulate directly
emitted by diesel engines (at a two percent conversion rate).
The first part of this section discusses the approach used
in this study to determine the extent of the reaction of SO2
to indirect sulfates in urban areas. (The results of this
analysis have already been used in the previous section to
quantify total sulfate and total particulate reductions.) The
second part discusses the methodologies used to develop ambient
concentration changes of the various pollutants. The final
part presents and discusses the air quality results.
A. Indirect Sulfate Analysis
There are two basic questions to be addressed in this
analysis - how much of the SO2 reacts in urban areas to form
-------
6-24
sulfate particulate, and what types of sulfate particulate are
formed? The answers to these questions help to quantify the
amount of ambient particulate reductions due to diesel fuel
sulfur controls.
This section is divided into three parts. The first part
summarizes estimates that have been made by the California Air
Resources Board (CARB) on SO2 conversion, and discusses some
of the theory of SO2 transformation. The second section
discusses the modeling approach used to estimate SO2
conversion in several (10) urban areas based on ambient SO4
and SO2 measurements. The third and final section presents
the results of this analysis, and estimates the mass of
indirect sulfate that would be expected to be reduced by each
unit of mass of reduction of SO2 emissions. This has already
been used in the previous section to characterize indirect
sulfate reductions due to diesel fuel sulfur controls.
1. Background
SO2 reaction rates have been studied in the Los Angeles
Basin by Cass. [9] The rates range from about 6% per hour in
the spring, summer, and fall to about 0.5% per hour in the fall
and winter. At 3 percent per hour, it only takes a little over
16 hours for the rest of the SO2 to react, so it has been
estimated that the other 50 percent of SO2 that has not been
lost to surface deposition reacts entirely to form sulfates.
CARB estimates that most of this ends up as ammonium sulfate,
with one associated water molecule.[10] On a mass basis, each
pound reduction in SO2 is expected to lead to a 1.16 pound
reduction in ambient (ammonium) sulfate particulate.
It has been widely recognized, however, that the
meteorological conditions affecting SO2 conversion (such as
sunlight, temperature, humidity, wind speed, and other
pollutants present) in the LA basin are quite different from
other parts of the U.S. For example, the SO2 conversion
rates used by EPA in several air quality models in the draft
RIA for SO2 National Ambient Air Quality Standards ranged
from 0 to 2 percent per hour. [11] These are much lower than
the California conversion rates. Thus, the rates or level of
SO2 conversion used in this analysis should be more
representative of the average meteorological conditions in the
U.S.
2. SO? Reactions and Products
SO2 follows a basic reaction tree that starts with
oxidation and branches off into a number of reactions, one of
which is ammoniation. For the purpose of assessing sulfate
particulate, the oxidation and ammoniation steps are most
important.
-------
6-25
SO2 initially reacts with oxygen to form sulfur
trioxide, or SO3. SO3 is very unstable and quickly reacts
with water vapor (H2O) to form sulfuric acid (H2SO4).
Sulfuric acid is very hygroscopic, and so depending on the
relative humidity, significant quantities of water vapor can
become attached to the sulfuric acid molecule and agglomerate
into a particle that would not have existed if the H2SO4
had not been present. Next, if there is ammonia present
(significant sources of ambient ammonia come from decaying
plant matter and animal waste), the sulfuric acid can react
with ammonia to form ammonium bisulfate (NH4HSO4) or
ammonium sulfate (^4)2804). The ammonium sulfates are
less hygroscopic than sulfuric acid, but the amount of
associated water is again related to relative humidity. These
reactions are summarized in Table 6-12.
Methods used to sample these species in the atmosphere are
limited to measuring the concentrations of the various ions -
S04=, H+, and NH4. Unfortunately, when H+ is measured, it
cannot be determined whether the H+ ion being measured was
associated with H2SO4 or NH4HSO4, so estimates of the
quantity of species present are always subject to some
uncertainty. A few recent studies have been conducted,
however, on sulfate species in urban areas, and most show a
predominance of the (NH4)2S04 species over sulfuric acid
and ammonium bisulfate. For example, one study measured the
sulfate species in Philadelphia in the daytime and found it to
be about 80 percent (NH4)2S04 and 20 percent
H2S04-[12] A nighttime study of the air around Houston
found the percentages to be 93 and 7 percent, respectively.[13]
In addition to the uncertainty in sulfate species, there
is also uncertainty in the amount of associated water with each
molecule. In the Regulatory Impact Analysis on the National
Ambient Air Quality Standards for Sulfur Dioxide, a range was
developed for the ratio of fine particulate mass (sulfate
anions, cations, and water) to sulfate mass.[11] The amount of
water associated with ammonnia sulfate and bisulfate was
estimated in the 0 to 50 percent range. The ratio ranged from
1.2 (100 percent ammonium bisulfate and no water) to 1.7 (100
percent ammonia sulfate and 60 percent water). A value of 1.6
was used in the RIA as a high estimate (1.5 was the middle and
1.4 was the low estimate). The previous measurements of
Philadelphia and Houston indicate, however, the presence of
some sulfuric acid in urban areas. At 50 percent relative
humidity, it has been estimated that 7 water molecules
associate with each sulfuric acid molecule.[ 14] The fine PM
mass to sulfate mass for this species is about 2.3. Therefore,
this analysis will use the fine mass to sulfate ratio of 1.6 in
estimating indirect sulfate inventories from the reaction of
S02.
-------
6-26
Table 6-12
Simplified SO? Reactions
Oxidation
S02 + 1/2 02 + S03
so3 + h2o h2so4
Ammoniation
H2S04 + NH3 ^ NH4HSO4
or
H2S04 + 2 NH3 (HN4)2 S04
-------
6-27
The above analysis has established the kinds of sulfate
species that are to be expected from SO2 reactions and the
mass of sulfate species per mass of sulfate anion. In order to
estimate sulfate inventories, however, the extent of SO2
conversion to sulfate species must still be estimated.
3. Methodology for Estimating SO? Conversion
In order to assess the overall conversion of SO2 to
sulfates in the urban atmosphere, an approach to model this
transformation is necessary. Two general methods exist to
determine the values of SO2 conversion-one by dispersion
modeling and the other by modeling based on measured sulfate
and SO2 concentrations,
Several photochemical dispersion models have been
developed to predict short-term ground level concentrations and
deposition fluxes of one or two gaseous or particulate
pollutants. The Pollution Episodic Model Version 2 (PEM-2),
developed by K. Shankar Rao, is one such model.[15] Because
this program accounts for the effects of dry deposition,
sedimentation, and first-order chemical transformation, it
possesses many capabilities, including a consideration of up to
300 isolated point sources and 50 distributed area sources. As
a result, PEM-2 is intended for studies of the atmospheric
transport, transformation, and deposition of acidic, toxic, and
other pollutants in urban areas to assess the impact of
existing or new sources or source modifications on air quality
for regulatory purposes and urban planning. However, a
necessary input to the model is the chemical transformation
rate of SO2 to sulfate, and this is not known to any degree
of precision for various cities under typical conditions. As a
result, such models could not be used here; and a different,
simplified method of conversion modeling is used in this study.
A simpler method, and the one used here is to calculate
SO2 conversion from urban ambient sulfate and sulfur dioxide
concentration data. This data in Pg/m3 can be converted
into molar concentrations by dividing the mass concentrations
by the molecular weights of each substance, and estimating the
ratio of sulfate moles to total sulfur moles (SO2 + sulfate
moles). Estimating SO2 moles is straightforward; however, to
estimate sulfate moles requires determining what sulfate
compound is present in the atmosphere, since only the SO4
anion concentration is measured. This step, however, was
accomplished in the previous section. The molecular weight of
a compound with a fine PM mass to sulfate ratio of 1.6 is 1.6 x
98, or 156.8 grams per gram-mole.
This simplified procedure for SO2 conversion requires
several assumptions which could make the conversion rates thus
developed somewhat uncertain. First, all of the sulfate is
assumed to come from conversion of S02> and none as direct
-------
6-28
sulfate emissions. As such, the conversions might be somewhat
high, since indirect sulfate concentrations would be lower if
direct sulfate emissions were subtracted. However, it was
shown in the last section that indirect sulfates probably far
outweigh direct sulfates, even at low conversions of SO2/
therefore, the error due to this assumption is probably not
large. Secondly, this technique does not attempt to account
for the influence of sulfate and/or SO2 from other areas, nor
does it explicitly account for deposition of either sulfate or
S02- In the "simple" urban area where there is no influx of
sulfate or SO2 from surrounding urban areas, one need only
consider the relative differences in advection and deposition
of sulfate and SO2 • If the deposition of sulfate is much
higher than SO2, then measured sulfate will be low and SO2
conversions based on ambient data will be low. The converse is
also true. Indirect sulfate particles are extremely small, on
the order of 2 microns or less, and there is evidence that they
disperse like a gas. Therefore the rates of deposition and
advection between SO2 and sulfate particles are probably not
materially different, so estimates of SO2 conversion based on
ambient SO2 and sulfate in urban areas should be reasonable.
In the more complex situations, sulfates and SO2 are
coming into the urban area from an adjacent urban area.
Assuming equal deposition, there is a good chance the sulfate
to SO2 concentration is higher than the ratio of the area
from which it came due to the fact that there has been more
time for the various sulfate reactions to take place. However,
both the sulfate and sulfur dioxide concentrations of the
subject urban area are increased by the emissions from the
adjacent area. The apparent SO2 conversion in the subject
urban area therefore may be higher than the actual SO2
conversion.
The above discussions illustrate some of the assumptions
involved in this modeling approach. While there may be some
uncertainty in the conversion rates developed, by selecting a
variety of urban areas and looking at the range of conversion
rates, the analysis will yield results which can be compared to
the CARB analysis.
4. Results
Sulfate and sulfur dioxide concentrations were available
from ten cities.[16,17 ] These were average concentration
measurements over the period of 1982-1984. The data are shown
in the first two columns of Table 6-13. Sulfate concentrations
range from 4.4 Mg/m3 in Tucson to 13 Mg/m3 in
Cleveland. SO2 concentrations range from 14 Rg/m3
(again, Tucson) to 64 Mg/m3 in New York City. Molar
conversion ratios estimated from these concentrations are shown
in the third column. They range from a low of 10 percent of
SO2 reacted in New York City to 21.4 percent in Toledo. The
-------
6-29
Table 6-13
City and National SO? Conversion Ratios
Molar
Ambient Cone. (Mg/m^)* Conversion
City SOa SO2 Ratio
New York 11.00 63.69 .103
Denver-
Boulder 6.90 30.54 .132
Kansas City 6.03 34.03 .106
Nashville-
Davidson 8.23 30.54 .152
Allentown 10.17 40.03 .156
Toledo 11.47 27.92 .214
Tucson 4.43 14.83 .166
Columbus 12.03 42.75 .158
Cleveland 13.00 52.35 .170
Buffalo 9.70 42.75 .131
National
Average - - .120
Averages of concentration measurements from 1982-1984.
-------
6-30
national population - weighted average using 1980 census
populations for these areas is 12 percent.
These molar conversions are all much lower than the 50
percent estimated by CARB for the LA Basin, but this in not
surprising, since they are based on yearly average
concentrations which reflect much colder weather than typically
occurs in the LA basin. An SO2 conversion rate of 12 percent
will be used in the remainder of the air quality analysis. At
a 12 percent conversion rate and 1.6 lbs of fine particulate
mass per pound of sulfate, about 0.288 lbs of sulfate
particulate mass are produced for every pound of SO2 emitted.
B. Air Quality Impacts
Two basic methods are used in this section to estimate
changes in urban concentrations of particulates, SO2, and
ozone. The methods used to estimate concentrations of
particulates and SO2 are very similar to the methods that
were used in establishing the heavy-duty engine NOx and
particulate standards and rely on atmospheric lead monitoring
data as a surrogate for diesel particulate and SC^-[1,2,3]
The ozone analysis uses changes in urban HC inventories and
results from the Empirical Kinetic Modeling Approach (EKMA)
model to estimate ozone concentration changes due to diesel
fuel controls. The lead surrogate methodology, and resulting
diesel particulate and SO2 concentrations, are discussed
first.
1. Particulate and SO? Concentrations
Since particulate and SO2 emissions derived from
on-highway mobile, off-highway mobile, stationary diesel and
fuel oil sources are not easily distinguishable from the same
emissions from other sources, air quality monitoring data are
not presently available specifically for these pollutants from
these sources, especially under the conditions expected to
exist in the future. Thus, any method for estimating air
quality impacts due to diesel fuel controls must use some
measurable surrogate in the ambient air that is directly
related to a source of emissions which is similar spatially and
temporally to the sources studied in the analysis. Various
studies in the past have used such substances as lead or CO to
provide a link between vehicle emissions and air quality. Once
this link is established, then vehicle emissions of the
surrogate substance are related to diesel particulate
emissions, resulting in an estimate of diesel particulate air
quality impacts.
The methodology used in this analysis uses lead as a
surrogate for diesel particulate and SO2. This type of
analysis uses historical data from urban sites in the national
urban lead monitoring network as an index of mobile source
-------
6-31
pollutant levels. An estimate is made of the fleet's
automotive lead emissions which caused the observed ambient
lead levels, and is compared to the composite particulate or
SO2 emissions developed from the sources covered in this
analysis. These composite emissions are estimated by summing
the annual emission inventories of the different sources and
dividing by total highway VMT. The result is an emission
factor in g/mi that includes stationary source and off-highway
emissions, but is directly comparable to the lead emission
factor developed only from on-highway sources. Very generally
speaking, if the composite particulate or SO2 emissions in
2015 are expected to be three times automobile lead emissions
in 1975, for example, then ambient composite particulate or
SO2 concentrations in 2015 can be expected to be three times
the 1975 ambient lead concentrations. In this case, 1975
monitoring data was chosen over more recent data to avoid, for
the most part, the errors associated with estimating the
leaded/unleaded vehicle mix.
The basic mathematical expression of this methodology is:
C(P)cy = (E(P)cy/E(Pb)75)*(S
-------
6-32
the pollutant in the particular calendar year, and these
estimates are already available from the previous section. The
work done for the final heavy-duty particulate standards and
previous work done by EPA developed estimates of the lead
emission factor in 1975, and the dispersion factors of lead and
diesel particulate (diesel particulate was assumed to disperse
like a gas because of its small size, so SO2 will use the
same dispersion factor). The lead emission factor used was
0.11 g/mi. The lead dispersion factor was determined to be
0.43, and the diesel particulate dispersion factor was
determined to be 1.00. The 1975 urban VMT is 758 billion
miles.[l ]
The use of the factors mentioned above results in the
following general equation:
C(P)cy = P(tons,urban,cy) * 1.00 * C(Pb,1975) * 1
0.11 g/mi 0.43 7.58xl0"miles 1/908,000
or
C(P)cy = P(tons,urban,cy) * C(Pb)75 ur£,an
39,500
For rural areas, the lead emission factor E(Pb) has been
estimated to be the same as for urban areas (0.11 g/mi). The
1975 rural VMT can be estimated by multiplying urban VMT by the
ratio of rural to urban VMT for 1975 from the MOBILE3 Fuel
Consumption Model.[18] In this model, rural VMT is about 7 6
percent of urban VMT. Rural VMT is estimated therefore at 572
billion miles. The equation for rural areas, then, is
C(P)cy = P(tons,rural,cy) * C(Pb)75 rurai
52,340
At this point it is necessary to consider one additional
factor, which was also used in previous analyses, to make the
lead ambient concentrations fully compatible with mobile source
modeling. This factor is used to correct the lead air quality
data for the fact that only an estimated 89 percent of the
total nationwide lead emissions are due to mobile sources.
Therefore, the ambient lead concentrations were multiplied by
0.89 to account for the possible contribution of non-automotive
sources. This adjustment is probably conservative, though to
an unknown degree, because only lead monitors in areas of no
known large stationary source of lead were chosen for this
analysis and the majority of the non-automotive sources of lead
emissions reside in a few identifiable areas which have been
excluded from this analysis.
-------
6-33
Therefore, the final equations become:
Urban: C(P) = P(tons,urban) * 0,89 *C(Pb)75 urban
39,500
Rural: C(P) = P(tons,rural) * 0.89 * C(Pb>75 rural
52,340
Tables 6-14 through 6-17 present the lead-based
concentrations for carbon, total sulfates, total particulates
and SO2 for each fuel control scenario and four groups of
cities. Two calendar years are shown, 1990 and 2015. The
first two columns show concentrations for 1990 and 2015 without
fuel controls. The rest of the columns show concentrations
with fuel controls for both 100 percent and NPRA segregation.
The cities used and their corresponding lead values are shown
in Table 6-18 (other data are also shown for these cities,
which will be explained in the section on visibility) . The
rural lead value comes from the rural New York Bight area. [19]
(Distributions of concentrations by city size category are not
presented here, but can be estimated from the coefficient of
variation of lead concentrations, which is 41 percent for
cities over 1,000,000, 35 percent for cities between 400,000
and 1,000,000, 31 percent for cities between 200,000 and
400,000, and 37 percent for cities under 200,000.) Generally,
pollutant concentrations are lower in the smaller cities, with
pollutant concentrations being the lowest in rural areas, where
they are only about one-tenth of the values in cities over one
million in population.
Carbon concentrations in Table 6-14 for 1990 range from
4.31 Mg/m^ to 0.43 Mg/m3. They are only slightly
higher in 2015 with no fuel controls. This is the result of
two counterbalancing effects. First, the 1991 and 1994 diesel
particulate standards cause a decline in carbon particulate
from on-highway sources. However, this is balanced by an
increase in fuel consumption in both on-highway, off highway
mobile, and stationary diesel sources of carbon particulate,
which act to cancel out the reductions in on-highway carbon
particulate. Sulfur control raises carbon particulate because
on-highway engines are able to increase carbon particulate as
direct sulfate particulate decreases with sulfur controls in
meeting the same particulate standards (1991 and 1994).
Subsequent aromatics control has virtually no effect on carbon
particulate concentrations, due to only a small change in
emissions.
Sulfate concentrations are shown in Table 6-15. The
levels in 2015 with no fuel controls are almost double the
concentrations in 1990, owing to the increase in fuel
consumption. For sulfur control under 100 percent segregation,
2015 sulfate concentrations are significantly lower than the
levels in 2015 without fuel controls, but still higher than
-------
Table 6-14
Average Ambient Carbon
Concentrations by Population
City Size
Grouping
Greater than
1,000,000
400,000 - 1,000,000
200,000 - 4,000,000
< 200,000
Rural
No Fuel Control
1990 1994 2015
4.31 3.86 4.36
2.96 2.66 3.00
2.49 2.23 2.52
2.61 2.34 2.64
0.43 0.33 0.25
Sulfur Control
100% —NPRA
1994 2015 1994 2015
3.96 4.76 3.96 4.76
2.72 3.27 2.72 3.27
2.29 2.75 2.29 2.75
2.40 2.88 2.40 2.88
0.35 0.32 0.35 0.32
Subsequent
—Aromatics Control
100% —NPRA
1994 2015 1994 2015
3.87 4.77 3.88 4.77
2.67 3.27 2.67 3.28
2.24 2.75 2.24 2.76
2.35 2.88 2.35 2.89
0.33 0.32 0.33 0.32
-------
le
Average Ambient Total Sulfate
Concentrations by Population
City Size
Grouping
Greater than
1,000,000
400,000 - 1,000,000
200,000 - 4,000,000
< 200,000
Rural
No Fuel Control
1990 1994 2015
4.49
3.09
2.59
2.72
0.42
4.89 8.51
3.37 5.86
2.83 4.92
2.96 5.15
0.45 0.78
Sulfur Control
100% —NPRA
1994 2015 1994 2015
3.31 5.74 0.98 1.74
2.28 3.94 0.67 1.19
1.91 3.31 0.57 1.00
2.01 3.47 0.59 1.05
0.21 0.37 0.90 0.16
Subsequent
—Aromatics Control
100% —NPRA
1994 2015 1994 2015
3.31 5.74 0.98 1.74
2.28 3.94 0.67 1.19
1.91 3.31 0.57 1.00
2.01 3.47 0.59 1.05
0.21 0.37 0.09 0.16
-------
Table 6-16
Average Ambient Total Particulate
Concentrations by Population
City Size
Grouping
Greater than
1,000,000
400,000 - 1,000,000
200,000 - 4,000,000
< 200,000
Rural
No Fuel Control
1990 1994 2015
10.60 10.3 14.48
7.29 7.07 9.96
6.13 5.94 8.37
6.42 6.22 8.77
1.02 0.91 1.13
Sulfur Control
100% —NPRA
1994 2015 1994 2015
8.83 12.16 6.50 8.16
6.08 8.37 4.47 5.62
5.10 4.03 3.76 4.72
5.35 7.37 3.94 4.95
0.69 0.79 0.57 0.58
Subsequent
—Aromatics Control
100% —NPRA
1994 2015 1994 2015
8.65 12.14 6.31 8.15
5.95 8.35 4.34 5.61
4.99 7.02 3.65 4.71
5.24 7.36 3.82 4.94
0.67 0.79 0.54 0.58
-------
T 6
Average Ambient SO2
Concentrations by Population
City Size
Grouping
Greater than
1,000,000
400,000 - 1,000,000
200,000 - 4,000,000
< 200,000
Rural
No Fuel Control
1990 1994 2015
10.96 11.99 20.92
7.54 8.24 14.39
6.34 6.93 12.09
6.65 7.27 12.67
1.01 1.10 1.92
Sulfur Control
100% —NPRA
1994 2015 1994 2015
8.12 13.98 2.40 4.19
5.59 9.62 1.66 2.88
4.68 8.08 1.38 2.42
4.82 8.47 1.46 2.53
0.52 0.90 0.22 0.38
Subsequent
—Aromatics Control
100% —NPRA
1994 2015 1994 2015
8.12 13.98 2.40 4.19
5.59 9.62 1.66 2.88
4.62 8.08 1.38 2.42
4.92 8.47 1.46 2.53
0.90 0.90 0.22 0.38
-------
6-38
Table 6-18
City Lead Values, Baseline Visibilities, Radii
City Category (by pop)
Ambient
Lead
Cone.(Mg/m3)
Baseline
Visibility
(mi)
City
Radius(mi)
> 1,000,000
Houston
Los Angeles
New York
Philadelphia
2.09
2. 68
1.05
1.34
18
11
15
15
13.3
12.2
9.8
6 . 6
400,000 to 1,000,000
Atlanta
Boston
Denver
Kansas City, MO
New Orleans
Phoenix
Pittsburgh
San Diego
St. Louis
1
0,
1
0.
1,
2
0
1
1
05
92
59
80
06
10
85
13
58
12
21
75
17
9
58
10
25
13
6.
3,
6,
10,
8,
10,
4.
10
3
200,000 to 400,000
Birmingham
Cincinnati
Jersey City
Louisville
Oklahoma City
Portland
Tucson
1
0
1
0
1
0
0
22
81
03
96
66
81
75
10
11
15
10
17
24
60
5,
5 ,
2,
4,
13.
5
5
Less than 200,000
Mobile
New Haven
Salt Lake City
Spokane
Trenton
Waterbury
Yonkers
0
1
0
0
0
1
1
96
15
98
58
88
88
16
10
17
80
36
17
17
15
6
2
4
4
1
3
2
Rural
0 .13
30
N/A
-------
6-39
1990 levels. Sulfate concentrations are lowest for sulfur
controls with NPRA segregation. There is no effect of
aromatics control on sulfate concentrations.
Total particulate concentrations are shown in Table 6-16.
Significant reductions in particulate concentrations are
obtained with sulfur control for both maximum and minimum
segregation cases. The impacts of subsequent aromatics control
on total particulate reductions are extremely small, however.
SO2 concentrations are shown in Table 6-17. The SO2
concentrations essentially follow the same trends as sulfate PM.
2. Ozone Concentrations
Section I of this chapter developed urban HC reductions
due to diesel fuel controls. The purpose of this section is to
estimate the impacts of those reductions on urban ozone levels.
The model used in this analysis to estimate changes in
ozone concentrations from changes in HC inventories is the
Emperical Kinetic Model, or EKMA. Instead of running EKMA,
however, the urban HC reductions will be compared to urban HC
reductions from the proposed regulations on gasoline
volatility, which used EKMA directly in estimating the effects
such changes would have on urban ozone levels.
The RIA for volatility controls included analysis of
nationwide inventories for a variety of scenarios such as with
and without onboard refueling requirements, analyses that were
based on design-value versus July average temperatures, and a
few others.[20] The choice of scenario is not important; what
is important is the change in the ozone levels that corresponds
to a specific change in inventories. This analysis will use
the HC emission inventories for 61 non-attainment areas
{equivalent to "urban" here) volatility controls at 9.0 and 9.5
psi RVP, assuming a prior onboard refueling control
requirement, and based on design-value temperatures (Table 3-25
of the RIA). In that analysis, the inventory with 9.5 RVP
control in 2000 is 6.144 million tons, and the 9.0 psi RVP
inventory is 6.113 million tons, for a net difference of RVP
31,000 tons. The reduction from 1983 ozone levels for 9.5 RVP
control was 10 percent, and for 9 RVP control was 11 percent.
However, there clearly could be round-off error involved, since
no change in average ozone levels was seen between other
similar control levels.
The HC reductions due to diesel fuel controls in 2010 are
at most 30,000 urban tons, which is about the same as the HC
reduction due to the 9 RVP versus 9.5 RVP volatility controls.
The HC reductions due to diesel fuel controls, therefore, would
be expected to bring about at most a one percent decrease in
ozone levels.
-------
6-40
III. Cancer Risk Assessment
This section will develop estimates of individual lung
cancer risk for the fuel control scenarios. There are two
parts to a risk assessment of this type. The first part is to
assess the carcinogenic potency of diesel particulate. The
second part is to combine that potency with estimates of
exposure to diesel particulate to estimate lung cancer risk.
This was the procedure that was followed in setting the final
NOx and particulate standards for heavy-duty trucks.
An investigation was made into new research performed on
assessing the carcinogenic potency of diesel particulate since
the heavy-duty diesel particulate standards were established
(March of 1985). Although some basic research has been done,
efforts to incorporate this work into previous estimates of
diesel potency have not been completed. Also, such work is not
expected to significantly change previous potency estimates.
Thus, it appears that the potency estimates developed for the
final particulate standards can also be used in this analysis.
There is one minor difference in the methods used to
develop estimates of lung cancer risk in this section versus
those used in setting the particulate emission standards. This
analysis develops exposure estimates and potencies for the
soluble organic fraction (SOF) of particulate, since emission
estimates are now available for the different particulate
types. The analysis for the emission standards used exposures
and potencies for all diesel particulate (SOF, carbon and
directly emitted sulfates combines). The final results from
both analyses (lung cancer risks per person-year), however, are
directly comparable.
This section is divided into three parts. The first part
explains how the SOF exposure estimates are developed, and
presents the exposure results for the fuel control scenarios.
The second part summarizes the approach used to develop
carcinogenic potency estimates for diesel particulate in
establishing the heavy-duty particulate standards, which was
also used here. The final part presents the lung cancer risks
for the fuel control scenarios, combining exposures with
carcinogenic potencies.
A. Population Exposure Analysis
The population exposure estimates used in this report are
based on a general air pollutant exposure model, called the
NAAQS Exposure Model (NEM), developed by OAQPS for the
evaluation of relative population exposures under alternative
NAAQS.[21] The NEM is an activity pattern model that simulates
a set of population groups called cohorts as they go about
their day-to-day activities. Each of these cohorts are
-------
6-41
assigned to a specific location type during each hour of the
day. Each of several specific locations types in the urban
area are assigned a particular air quality value based on fixed
site monitor data. The model computes the hourly exposures for
each cohort and then sums these values over the desired
averaging time to arrive at average population exposures and
exposure distributions. Thus, the model simulates pollutant
concentrations in urban areas by relating pollutant
concentrations in urban locations to fixed-site monitor levels
and simulates the activities of people by relating the
population to a fixed set of cohorts with defined activity
patterns.
For example, a certain fraction of the total urban
population might be assigned to an office worker cohort with a
home-work-home activity pattern. This cohort would experience
a consistent set of microenvironments, such as, home,
transportation, and office in a normal day's activity. Each of
these microenvironments would have an associated pollutant
concentration related to the fixed-site monitor level for the
specific time of day and date. The fixed-site monitor levels
are adjusted to correct for the differences that typically
exist between the monitored locations and the microenvironment
location. These adjustments are general enough to account for
multiplicative and additive types of correlations between the
monitors and locations.
A unique feature of the model is that it separates
concentrations, people, places, and time (all of the important
ingredients of exposure) into discrete elements. The
concentrations are broken up into values determined by the
precision of the fixed site monitors (e.g., for CO into whole
integers of ppm). The urban population of an area is separated
into cohorts that are assumed to have specific activity
patterns and hence exposures. The places are separated into a
discrete number of areas which are assumed to have identical
pollutant concentrations over a given period of time. Thus,
the definition of places may be influenced by the type of
pollutant studied and its emission sources. Time is separated
into the smallest unit of measure which is desire. Since this
methodology was designed to be used with the NAAQS, which are
based, at a minimum, on a 1-hour time period, one hour is the
shortest time period considered by the general model. Longer
averaging times, such as 24 hours or a year, can be obtained by
calculating the appropriate averages from the 1-hour exposures.
The general NEM approach has been used by OAQPS to develop
specific models for a number of criteria pollutants, such as
CO, SO2» NO2 and particulate.[ 21] These pollutants have
been studied by applying the specific pollutant data base from
the appropriate EPA monitoring program and by designing the
place designations to those most appropriate to the pollutant.
Place or location designations in the NEM are determined as
-------
6-42
exposure districts or exposure neighborhoods, depending on
whether the pollutant of interest is a point or dispersed
source of emissions, respectively. The exposure district
approach is more geographical in nature and relies on the fact
that pollutants which are primarily emitted by large point
sources can be adequately characterized by exposure districts
of fairly large areas. In contrast, pollutants with emission
sources which are dispersed throughout an urban area (including
mobile sources) are best characterized by considering exposure
neighborhoods with common exposure patterns. These
neighborhoods may be spread out in a random fashion throughout
the urban area.
In the analysis used in setting the heavy-duty NOx and
particulate standards, the NEM modeling approach as applied to
CO levels using monitoring data in four cities (Chicago, Los
Angeles, Philadelphia, and St. Louis) was applied to the
criteria pollutants mentioned above resulting in four average
exposures.[22] A limited set of 24 total monitors in four
cities was used because of the extensive computer time required
to run the NEM. A rough nationwide exposure extrapolation was
developed by OAQPS, which involved relating each of the large
urban areas in the country to the most similar of the four
modeled NEM cities.
Direct measurements of diesel SOF particulate levels in
overall urban areas have not been made because of the
difficulty in distinguishing diesel particulate from other
carbonaceous particulate. The use of a surrogate approach to
relate diesel SOF particulate to some other pollutant which can
be readily measured in urban areas appears to afford the best
chance of obtaining reasonable estimates of diesel particulate,
as was done for urban air quality estimates in the previous
section of this report. Of the criteria pollutants which have
been assessed by OAQPS using the NEM methodology, CO appears to
have the most desirable characteristics as a surrogate. The
dispersion characteristics of CO are also very similar to those
of diesel particulate, since diesel particulate are very fine
and disperse essentially like a gas.
The CO NEM is designed to provide an overall estimate of
CO population exposure related to different values of the CO
NAAQS. While the selection of CO and the NEM methodology is
the best basis for a mobile source assessment of population
exposure, a number of modifications to the NEM CO assessment
are necessary and/or desirable in order to provide the best
estimate of diesel particulate exposure.
One modification to the standard NEM methodology involves
the removal of all indoor sources of CO, such as gas stoves,
from the air quality inventories. This change allows for a
more precise estimate of automotive sources, particularly
indoors, where the contribution of outdoor emissions is still
-------
6-43
present but to a lesser degree than indoor sources. This is
the most important modification to the model in order to allow
a reasonable estimate of automotive exposure to CO and, via an
appropriate conversion, to diesel SOF particulate.
Table 6-19 presents the NEM based average CO
concentrations for the four cities used in the NEM program.
The average CO exposure concentrations for each city in Table
6-19 are combined in order to provide the desired nationwide
average exposure. A simple method to use and the one used by
the CO NEM and in this report groups each of the large urban
areas in the country (populations greater than 200,000) with
one of the four modeled cities according to overall urban
characteristics including populations, vehicle use patterns,
etc.[22] Under this nationwide extrapolation a large portion
of the population (43 percent) is grouped under Chicago.
The national population in Table 6-19 as mentioned
previously (Column 4) counts only persons in urban areas with
populations greater than 200,000. While it is desirable to
extend this analysis to the population in all urban areas
including those with populations less than 200,000, the
likelihood that the exposures in smaller urban areas would be
lower than any of the four NEM cities prevents this from being
precisely done. Therefore, the analysis has been limited to
the populations in the large urban areas without considering
the exposures of rural or small urban areas.
The aforementioned nationwide extrapolation of the NEM-CO
average output results in the calculation of an overall average
nationwide concentration (based on CO) of 2.12 ppm. This total
national average can then be manipulated into a national
average diesel SOF exposure by ratioing SOF and CO inventories,
and multiplying the result by 2.12 ppm. (A correction factor
of 1149 is needed to convert 2.12 ppm at standard temperature
to Mg/m^ at 75°F.) The SOF inventories are found in
section I of this chapter, and the 1978 CO inventory is found
by multiplying an average 1978 CO emission factor of 67,3 g/mi
by 1978 total VMT, 2.0396 trillion miles.[1]
Table 6-20 presents the national average SOF exposures for
the base case (no fuel controls), sulfur and aromatics control
for calendar years 1992, 2000, and 2015. Only exposure to
on-highway diesel emissions are shown since potency estimates
are not available for the particulate emitted by the other
sources. Results for the two segregation cases are exactly the
same because the SOF emissions of on-highway diesel sources are
affected in the same way in both cases.
On-hicrhway SOF concentrations for the base case drop from
0.816 Ug/nw in 1992 to 0.247 in 2015. This drop is due to
the influence of the 1991-93 and 1994+ particulate standards on
SOF emissions. For sulfur control, on-highway SOF
-------
6-44
Table 6-19
Average Total CO Exposure In Four Cities
City
Chicago
Los Angeles
Philadelphia
St. Louis
CO ppm
(Annual Avg.)
1.8 ppm
3.0
1.3
2.0
Population
of City As
Used In CO NEM
2,363,014
7,716,895
2,933,790
1,221,461
Associated Total
Urban Population
In Cities Greater
Than 200,000-1970
38,894,395
26,339,249
10,553,523
17,350,712
Overall
2.12
93 ,137,849
-------
6-45
Calendar
Year
1992
Table 6-20
On-Hiqhway Diesel SOF Particulate Exposure
Annual Average SOF Exposure (Hq/m3)
Subsequent
Base Sulfur Aromatics
0.816 0.821 0.688
2000
2015
0 .277
0 .247
0 .314
0 .300
0.272
0.269
-------
6-46
concentrations are very slightly higher due to higher SOF
emissions from trucks designed to meet particular standards
with low sulfur fuel. For aromatics control, significant
reductions in SOF exposures from on-highway vehicles are
evident, especially in the earlier calendar years when
reductions are obtained from in-use trucks developed to run on
a higher aromatics fuel.
3. Cancer Assessment
The ideal means to determine the risk due to cancer from a
given exposure of diesel SOF particulate would be to conduct a
long-term epidemiological study. Such a study would trace the
health of several well defined groups of people who were
exposed to precisely known concentrations of diesel SOF
particulate emissions for known periods of time.
Unfortunately, these kind of data are not available, and so
other techniques must be used to estimate cancer risks for
diesel SOF particulate emissions.
The basic technique used in assessing cancer risks for the
heavy-duty diesel particulate standards was to use estimates of
risk from epidermiological studies on other similar pollutants
such as coke oven emissions; and adjust these estimates by
measures of the relative potencies of diesel particulate
emissions to these other emissions. The relative potencies are
developed by comparing the mutagenicity and carcinogenicity of
diesel particulate to the other emissions at equal
concentrations.
Due to the uncertainty in mutagenicity and carcinogenicity
estimates that were (and are) available, a range was developed
in the analysis for the heavy-duty diesel particulate standards
for the normal risk due to diesel particulate (lung cancer risk
per person per Hg/m3 particulate).[1,2,3] That range was
from 0.26xl0-6 to 1.4xl0-6 annual incidences/Mg/m3 of
diesel particulate. This range was then applied to estimates
of diesel particulate exposure to give comparative estimates of
lung cancer risk for the different control scenarios.
This same range can be applied to the SOF exposures
developed in the previous part in this study, however, one
adjustment is necessary. The risk range from the above
analysis was for whole diesel particulate
(SOF+carbon+sulfate). This needs to be adjusted for the SOF
only.
The low end of the risk range came from the low end of a
range developed by EPA's carcinogen assessment group and
represented the potency of particulate from an Oldsmobile
diesel engine. The potency of this particulate's SOF was
1.6xl0~6 annual incidence/Hg/m3.[23] This latter number
is a factor of six times the figure for total diesel
-------
6-47
particulate, due to the simple fact that the mass of the
organic extract was one-sixth the total particulate mass. It
will be used for the low end of the risk range for SOF
particulate.
The high end of the risk range (1.4xl0-6) was developed
from bioassays of diesel particulate by Harris.[24] The
corresponding potency of the SOF in this study was 16.7 times
higher, or 23.3xl0~6 annual incidence/Ug/m3 of SOF. The
full range, then is from 1.6 x 10~° to 23.3 x 10~6
risk/Pg/nw SOF particulate.
When this risk range is combined with the SOF exposure
estimates given in Table 6-20, the results are individual lung
cancer risk estimates, which are given in Table 6-21. These
lung cancer risks follow the same trends as the SOF exposures.
Risk estimates for off-highway construction and stationary
diesel are not shown, since their operational characteristics
differ dramatically from on-highway diesels and could cause a
dramatic difference in the potency of their particulate. For
sulfur control, the risk appears equal to or slightly higher
than the base case.
The cancer impacts of just reducing aromatics levels in
diesel fuel to 20 weight percent can be found by comparing the
estimated individual risks in Table 6-21 for the reduced
aromatics and sulfur scenario to the risks of just reducing
sulfur levels. It can be seen that the annual risk reduction
due to aromatics control would be approximately 0-0.11
incidences per million persons in 2000. Applying this to the
urban population listed in Table 6-19 for cities over 200,000
(93 million persons) would give an estimated reduction of up to
10 incidences per year.
IV, Visibility Assessment
Section II developed estimates of the changes in the
concentrations of different types of particulate (carbon,
sulfates) for the fuel control scenarios. Because particles
scatter and absorb light, changes are expected in visibility
with the changes in particulate concentrations. Furthermore,
different types of particulate have different effects on light
scattering and absorption. In section II, there was an
increase in carbon concentrations, but a large reduction in
sulfate concentrations. The purpose of this section is to
estimate the net visibility changes due to diesel fuel controls.
This section is divided into two parts. The first part
summarizes the visibility methodology used in setting the
heavy-duty particulate standards, which is also used here, and
the modifications made for this analysis. The second part
presents the changes in visibility due to fuel controls using
this methodology.
-------
6-48
A. Methodology
The methodology used in setting the heavy-duty particulate
standards is based on an analysis of the visual properties of
air and the Beer-Lambert law, in which the reduction in light
intensity is a function of distance and the extinction
coefficient (bext^°f media:
-------
6-49
Table 6-21
On-Highway Diesel Cancer Risk
Projections Per Million Population
Estimated Individual Risk Range
(1Q-6 lung cancer risk/person year)
Calendar
Year
Base
Case
Sulfur Aromatics (20%)
Control and Sulfur (0.05%)
(0 . 05%) Control
1992
13-1.9
13-1.9
.05-1.6
2000
.05-.64
.05-.74
.05-.63
2015
.04-.59
.05-.70
.04-.63
-------
6-50
I = l0 e~bextL
Where
I0 = light intensity at object being observed, and
I = light intensity at distance L from the object.[1]
This method also used Koschmieder's Law, which states that
objects become invisible to the human eye when the ratio of I
to I0 becomes 0.02.[1] The above equation then becomes
Ly = 3.91
bext
where
Ly = visual range, and
bext = total extinction coefficient of the media
The most important parameter of these equations is the
total extinction coefficient (bex-t-). This coefficient is the
sum of four components:
1. scattering by gas molecules
2. absorption by gas molecules
3. scattering by particles
4. absorption by particles
The impact of fuel control on items 1 and 2 is negligible,
therefore, this analysis is primarily concerned with scattering
and absorption by particles, the particles of interest being
sulfates and carbon particulates. SOF emissions, even though
absorbed onto the carbon particulate are assumed to be
transparent (as they are when in the gaseous phase) and
therefore do not affect visibility. Also, as will be discussed
subsequently, the extinction coefficient used in this analysis
will be for fine elemental carbon, which does not include SOF.
The equations given above assume that the particles which
absorb and scatter light are evenly distributed in the visual
range. This is not necessarily the case in urban areas where
the visual range may extend beyond the area affected by diesel
particulate. The methodology used in the heavy-duty
particulate standards analysis assumed that the PM
concentrations estimated in Section II applied within but not
outside the city radius, and derived two new equations from the
above equations to account for this. The first equation below
was used for urban areas or cities where the baseline
visibility extended beyond the radius of the city, and the
second equation was used where the city radius was equal to or
larger than the baseline visibility.
-------
6-51
Lv = (18.6xl0_4[miles/m] - bext,dp Mc La)/bo
Lv = 18.6xl0~4[miles/m]/[18.6xlO~4/Lvo)+bext,dp
Where
(1)
(2)
bext,dp = extinction coefficient for whole diesel
particulate
Mc = mass concentration of diesel particulate
In the previous analyses, bext and Mc were derived for
total diesel particulate. This analysis will account for the
effects of sulfates and carbon separately. To do this, the
term bext;dp Mc wiH be replaced by the sum of the product
of the same two parameters for sulfates and carbon. Equations
(1) and (2) then become:
Lv = LVoC.00186-(bext>c*[C]+bext/S[S])]/.00186 (3)
Lv = .00186/[(.00186/LVo)+bext,c*[cl+bext,s*[s^ <4>
The concentrations of carbon and sulfate are available
from section II. The extinction coefficient of elemental
carbon is 11.5 m2/q and the extinction coefficient used for
sulfates is 3.5 rn^/g (assuming a mean particle radius of
0.1-0.6 urn for the latter).[1] Baseline visibilities and city
radii are the same as these used in the previous analyses, and
are shown in Table 6-18.
For rural areas, baseline visibility was not available, so
this analysis used 30 miles, which falls in the range of the
cities less than 200,000. In most rural areas of the west, the
baseline visibility is probably significantly greater than 30
miles, but in other areas of the midwest, east, south and
northwest where relative humidities are much higher the
baseline visibility may be significantly less than 30 miles.
The absolute baseline visibility is not that significant,
however, as the analysis will be focusing on percentage changes
in visibility, rather than on their absolute values.
Also concerning the analysis for rural areas, the affected
radius was assumed to extend beyond the baseline visibility (30
miles), as there is probably a fairly even distribution of a
low concentration of particulate emissions in these areas.
Where
[S]
baseline visibility of each city in miles
extinction coefficient of carbon
extinction coefficient of sulfate
ambient concentration of carbon, Pg/m3
ambient concentration of sulfate, Ug/m3
-------
6-52
The results of the visibility analysis for fuel controls
are shown in Table 6-22. The estimates shown are percentage
changes in visibility from the base case in 1990 to each
control case in 2015. Positive values are improvements in
visibility, negative values are reductions. Results are
disaggregated by city size category and by level of fuel
segregation. The base case shows percentage changes in
visibility due to on-highway, off-highway and other diesel
sources from 1990 to 2015 with no fuel controls. All other
fuel control cases measure percentage changes in visibility
from 2015 with controls to 1990 without controls.
In the base case, visibility is projected to decrease
between 1.8 and 11.2 percent from 1990 to 2015. Rural areas
show a small improvement (1.1 percent) in visibility. There is
a reduction in particulates from on-highway sources (because of
the particulate standards) which improves visibility, but this
is outweighed by the negative impact of the growth of SO2
emissions (and thus, indirect particulates)in all sources in
the absence of fuel controls. Rural areas show an improvement
of visibility due to improvement of on-highway emissions, which
are mostly (67 percent) rural.
An improvement in visibility is noted with sulfur controls
with 100 percent segregation over the base case. This is due
to lower direct and indirect sulfate emissions from on-highway
vehicles. There is no change in visibility for aromatics
control.
A significant further improvement in visibility is noted
if diesel fuel sulfur level is lower for off-highway and other
diesel sources as well as on highway vehicles, as shown in the
percentage changes in visibility for minimum segregation. The
positive values indicate an improvement in visibility in 2015
over 1990 levels.
-------
6-53
Table 6-22
Average Percentage Changes in Visibility
From 1990 to 2015 by City Size Category
-100% Segregation- -NPRA Segregation-
City Size
Base
Sulfur
Aromatics
Sulfur
Aromatics
>1 million
1
»-»
h-»
to
-7.4
-7.5
3.4
3.3
400,000 1 million
-5.1
-3.3
-3.4
1. 6
1.5
200,000-400,000
-4 . 4
-2.9
-2.9
1.4
1.3
<200,000
-1.8
-1.2
-1.2
0 . 5
0.5
Rural
1. 1
2.1
2.1
3.2
3.2
-------
6-54
References (Chapter 6)
1. "Diesel Particulate Study," SDSB, ECTD, OMS, OAR,
EPA, October 1983. Docket A-80-18.
2. "Draft Regulatory Impact Analysis and Oxides of
Nitrogen Pollutant Specific Study," SDSB, ECTD, OMS, OAR, EPA,
October 1984. Docket A-80-18.
3. "Regulatory Impact Analysis, Oxides of Nitrogen
Pollutant Specific Study and Summary and Analysis of Comments,"
ECTD, OMS, OAR, EPA, March 1985. Docket A-80-18.
4. "The Motor Fuel Consumption Model, Thirteenth
Periodical Report," prepared for DOE by Martin Marietta Energy
Systems, Inc., and Energy and Environmental Analysis, Inc., May
26, 1987.
5. "Heavy-Duty Vehicle Emission Conversion Factors II,
1962-2000," Paul Machiele, EPA-AA-SDSB-89-01, October 1988.
6. MOBILE4 Travel Characterization Data Handout,
MOBILE4 Workshop, November 1987. EPA, OAR, OMS, ECTD.
7. "Compilation of Air Pollutant Emission Factors,
Volume I: Stationary Point and Area Sources," 4th Edition,
AP-42, September 1985.
8. "State and Metropolitan Area Data Book - 1982,"
Bureau of Census, Dept. of Commerce, August 1982.
9. "Sulfate Air Quality Control Strategy Design," by
Glen R. Cass, Atmospheric Environment, Vol. 15, No. 7, pp.
1227-1249, 1981.
10. "Review of Motor Vehicle Diesel Fuel Modification
Emission Reduction Strategies," SSD, CARB, October 1986.
11. "Regulatory Impact Analysis on the National Ambient
Air Quality Standards for Sulfur Oxides (Sulfur Dioxide)
(Draft)," SASD, OAR, EPA, May 1987.
12. "A Composite Receptor Method Applied to Philadelphia
Aerosol," by Dzubay, T.G.,; Stevens, R.K.; Gordon, G.E.; Olnez,
I.; Sheffield, A.E.; and Courtney, W.J., submitted to Environ.
Sci. Techno1., Vol. 22, No. 1, pp. 46-52, 1988.
13. "Visibility and Aerosol Composition in Houston,
Texas," by Dzubay, Thomas G. ; Stevens, Robert K.; and Lewis,
Charles W., Environ. Sci. Techno1¦, Vol. 16, No. 8, 514-524,
1982.
-------
6-55
14. "An Acid Aerosols Issue Paper: Health Effects and
Aerometrics," EPA, OHEA, ECAO, May 1987.
15. "User's Guide for PEM-2: Pollution Episodic Model
(Version 2)," prepared for the U.S. EPA by Rao, K. Shankar,
NOAA, Contract No. 1AG-DW13930021-01-1, 1984.
16. EPA Memorandum from Jake Summers, RIS, NADB, to
Angela Lindner, January 29, 1987.
17. "National Air Quality and Emission Trends Report,
1986," OAQPS, EPA, EPA-450/4-88-001, February 1988.
18. "MOBILE3 Fuel Consumption Model," Mark A. Wolcott,
U.S. EPA, OAR, OMS, and Dennis F. Kahlbaum, Computer Sciences
Corporation, February 1985, EPA-AA-TEB-EF-85-2.
19. "Air Quality Criteria for Lead", EPA 600/8-77-017,
Health Effects Research Lab, RTP, NC, December 1977.
20. "Draft Regulatory Impact Analysis, Control of
Gasoline Volatility and Evaporative Hydrocarbon Emissions from
New Motor Vehicles," OMS, OAR, EPA, July 1987.
21. "A General Model for Estimating Exposure Associated
With Alternative NAAQS," Biller, W., et. al., June, 1981.
22. "The NAAQS Exposure Model (NEM) Applied to Carbon
Monoxide," Johnson, T. , et. al. , Draft EPA-OAQPS Report,
December, 1982.
23. "A Comparative Potency Method for Cancer Risk
Assessment: Application to Diesel Particulate Emissions,"
Albert, R. E., J. Lewtas, S. Nesnow, T. W. Thorslund and E.
Anderson, Risk Analysis, Vol. 3, No.2, p.101-117 1983.
24. "Potential Risk of Lung Cancer from Diesel Engine
Emissions," Harris, J., National Academy Press, Washington,
D.C., 1981.
-------
Appendix 6-A
Diesel Fleet VMT, Fuel Consumption, Registrations
Fuel Economy, and Annual VMT/Veh Calculations
The calculations involved in this report required diesel
fleet fuel consumption data, diesel fleet annual VMT data,
diesel annual VMT per vehicle data, diesel fleet fuel economy
data, and diesel fleet registration data which were accurate,
and consistent for all five categories. Since the DOE 13th
Periodical Report appeared to provide a fairly accurate
estimate of total fuel consumption, it was used as much as
possible as a basis for the following calculations.[4]
Fuel Economy
Fuel economy values for the years through 2020 were taken
from the MOBILE3 fuel consumption model with the exception of
those for HHDDTs.[18] For the years 2021 through 2023, the
fuel economies were assumed to remain constant at the 2020
value. Since the fuel economy for the HHDDTs was available
from the DOE 13th report through the year 2000, it was taken
from there to remain as consistent as possible with the total
fuel consumption data. In order to project it out to 2023, it
was assumed to increase in proportion to that for the HHDDVs in
the MOBILE3 fuel consumption model. The fuel economy values
for all classes are shown in Table 6-A-l. The average for all
buses was drawn from the MOBILE3 information by weighting the
bus categories by their fleet registrations (which is discussed
later).
Fleet VMT
Fleet VMT through the year 2000 was determined by
multiplying the class specific fuel economy estimates
determined above, by the diesel fuel consumption values for
these classes found in the DOE 13th report. [4] The 13th
report, however, contained no information on buses. As a
result diesel bus fleet VMT was determined by multiplying bus
fleet registrations by the average annual VMT per bus. The
derivation of these factors is discussed below. For the years
2001 through 2023 fleet VMT values were calculated by assuming
an annual increase of three percent. The diesel fuel
consumption values for all the classes are shown in Table 6-A-2.
Fleet Registrations
Once again, to remain consistent with the assumptions of
total fuel consumption, the diesel fleet registration values
through the year 2000 for all vehicle classes except buses were
taken from the 13th periodical report. The projections for the
2001 to 2023 time frame were derived by dividing the fleet VMT
values by the estimates of average annual VMT per vehicle.
-------
6-A-2
Since buses were not included in the 13th periodical
report, fleet registration values had to be derived
independently. The values in the M0BILE3 fuel consumption
model for public buses were assumed to be representative of
transit and commercial buses. However, the estimates for
diesel school buses appeared to grossly underestimate the
recent trends in sales of diesel school buses. Diesel sales
data for school buses from the M0BILE4 conversion factor
analysis was used to estimate the fraction of school buses
which were diesel for the years through 1990. The assumption
was then made that 20 years after all school bus sales were
diesel that all school bus registrations were diesel (The year
2019). To estimate the fraction of school buses which were
diesel for the years 1991 to 2023, a linear interpolation
between the 1990 and 2019 values was performed. Diesel school
bus registrations for 1991 to 2023 were then determined by
multiplying these diesel registration fractions by the total
school bus registrations available from the MOBILE3 fuel
consumption model. The school bus registrations plus the
public bus registrations were then combined to form the total
bus fleet registrations. The diesel fleet registration values
for all classes are shown in Table 6-A-3.
Average Annual VMT per Vehicle
For all vehicle classes except for buses, the average
annual VMT per vehicle values through the year 2000 were
calculated by dividing the fleet VMT estimates by the fleet
registrations (which were discussed above). In order to take
into account the effects of changes in diesel penetration into
a vehicle class, the estimates for the years 2001 through 2023
were obtained by changing the 2000 values proportionately to an
independent estimate of average annual VMT per vehicle which
was based on MOBILE4 and MOBIL.E4 conversion factors
information. This estimate was determined by multiplying the
MOBILE4 fleet registration distributions by vehicle age by
estimates of model year specific diesel sales fractions, and
renormalizing these products to get an age specific diesel
registration distribution. These in turn were multiplied by
estimates of the age specific annual VMT per vehicle. When
summed together across all vehicle ages, this results in a
calendar year specific average annual VMT per vehicle for each
vehicle class.
For buses, average annual VMT per vehicle estimates for
public buses, and school buses were taken from the Federal
Highway Administration's Highway Statistics Book, and weighted
together based on projected diesel registrations in each
class. This resulted in a weighted average annual VMT per bus
for each calendar year. Average annual VMT per vehicle values
for all classes of diesel vehicles are shown in Table 6-A-4.
-------
6-A-3
Fleet Fuel Consumption
Fleet fuel consumption for all calendar years could now be
determined by dividing the fleet VMT values by the fleet fuel
economy values. (Since for 1975 and 1990 through 2000 these
were based on the estimates of fuel consumption from the 13th
periodical report, with the exception of buses, the fuel
consumption for these years is identical to that from the 13th
report.) These fuel consumption values are shown in Table
6-A-5.
-------
6-A-4
Table 6-A-l
Diesel Fleet Fuel Economy (mpg)
Year
LDDV
LDDT
LHDDT
MHDDT
HHDDT
Buses
1975
21.85
20
8.235
6. 069
4 .35
4.78
1990
28.26
23.57
14. 04
7.285
5.933
6.06
1991
28.76
23 . 76
14.195
7.37
6.024
6. 16
1992
29 .32
23 .957
14.353
7.463
6. 118
6.26
1993
29.872
24.134
14.511
7.544
6.214
6.39
1994
30 .4
24 . 277
14.667
7.628
6.312
6. 51
1995
30.924
24 .438
14.846
7.707
6.411
6.62
1996
31.412
24.644
15.01
7.783
6.509
6.73
1997
31.858
24.842
15. 19
7.858
6.606
6.84
1998
32.287
25.079
15.381
7.933
6.689
6 . 97
1999
32.707
25.305
15.562
8.003
6.76
7.08
2000
33 . 114
25.543
15.752
8.07
6.82
7 .20
2001
33.441
25.738
15.918
8. 126
6.901
7.25
2002
33.723
25.901
16.078
8.167
6.965
7 .29
2003
33 .953
26.041
16.199
8. 196
7.012
7.34
2004
34.147
26.153
16.372
8.222
7.049
7 . 40
2005
34.314
26.25
16.469
8.238
7.076
7 . 44
2006
34.444
26.333
16.56
8.254
7.097
7 . 49
2007
34.566
26.405
16.633
8.262
7.113
7 . 54
2008
34.66
26.464
16.695
8.272
7.125
7. 59
2009
34.737
26.518
16.753
8.281
7.135
7 . 63
2010
34.798
26.562
16.789
8.286
7. 142
7 . 67
2011
34.84
26.598
16.817
8.291
7. 148
7 . 72
2012
34.881
26.627
16.843
8.294
7 .152
7.76
2013
34.907
26.654
16.869
8.296
7.154
7 . 81
2014
34 . 925
26.673
16.887
8.301
7. 156
7 . 85
2015
34.938
26.693
16.905
8.303
7.158
7 . 89
2016
34.948
26.702
16.904
8 .304
7 . 160
7 . 94
2017
34.949
26.712
16.914
8.308
7 . 161
7 . 97
2018
34.956
26.717
16.919
8.309
7. 162
8. 02
2019
34.958
26.717
16.918
8.31
7. 163
8. 06
2020
34.957
26.722
16.921
8.31
7. 163
8. 06
2021
34.957
26.722
16.921
8.31
7. 163
8.06
2022
34.957
26.722
16.921
8.31
7 . 163
8.06
2023
34.957
26.722
16.921
8.31
7. 163
8.06
-------
Year
1975
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
6-A-5
Table 6-A-2
Annual Fleet VMT
LDDV
LDDT
LHDDT
MHDDT
HHDDT
4.37E+08
0.00E+00
0.00E+00
7.28E+08
4.73E+10
1.89E+10
8.98E+09
1.45E+10
4.09E+10
6.58E+10
1.75E+10
9.39E+09
1.63E+10
4.17E+10
6.72E+10
1.64E+10
9.80E+09
1.79E+10
4.28E+10
6.85E+10
1.55E+10
1.02E+10
1.94E+10
4.38E+10
6.99E+10
1.49E+10
1.06E+10
2.09E+10
4.51E+10
7.13E+10
1.47E+10
1.09E+10
2.23E+10
4.64E+10
7.29E+10
1.49E+10
1.12E+10
2.36E+10
4.80E+10
7.46E+10
1.52E+10
1.15E+10
2.49E+10
4.98E+10
7.64E+10
1.56E+10
1.20E+10
2.61E+10
5.18E+10
7.84E+10
1.61E+10
1.24E+10
2.73E+10
5.39E+10
8.05E+10
1.68E+10
1.28E+10
2.85E+10
5.63E+10
8.27E+10
1.73E+10
1.32E+10
2.94E+10
5.80E+10
8.52E+10
1.78E+10
1.36E+10
3.03E+10
5.98E+10
8.78E+10
1.83E+10
1.40E+10
3.12E+10
6.15E+10
9.04E+10
1.89E+10
1.44E+10
3.21E+10
6.34E+10
9.31E+10
1.95E+10
1.49E+10
3.31E+10
6.53E+10
9.59E+10
2.00E+10
1.53E+10
3.41E+10
6.72E+10
9.88E+10
2.06E+10
1.58E+10
3.51E+10
6.93E+10
1.02E+11
2.13E+10
1.62E+10
3.61E+10
7.13E+10
1.05E+11
2.19E+10
1.67E+10
3.72E+10
7.35E+10
1.08E+11
2.26E+10
1.72E+10
3.83E+10
7.57E+10
1.11E+11
2.32E+10
1.77E+10
3.95E+10
7.80E+10
1.15E+11
2.39E+10
1.83E+10
4.07E+10
8.03E+10
1.18E+11
2.47E+10
1.88E+10
4.19E+10
8.27E+10
1.22E+11
2.54E+10
1.94E+10
4.31E+10
8.52E+10
1.25E+11
2.62E+10
2.00E+10
4.44E+10
8.77E+10
1.29E+11
2.69E+10
2.06E+10
4.58E+10
9.04E+10
1.33E+11
2.77E+10
2.12E+10
4.72E+10
9.31E+10
1.37E+11
2.86E+10
2.18E+10
4.86E+10
9.59E+10
1.41E+11
2.94E+10
2.25E+10
5.OOE+IO
9.88E+10
1.45E+11
3.03E+10
2.32E+10
5.15E+10
1.02E+11
1.49E+11
3.12E+10
2.39E+10
5.31E+10
1.05E+11
1.54E+11
3.22E+10
2.46E+10
5.47E+10
1.08E+11
1.59E+11
3.31E+10
2.53E+10
5.63E+10
1.11E+11
1.63E+11
-------
Year
1975
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
6-A-6
Table 6-A-3
Fleet Registrations
LDDV
LDDT
LHDDT
MHDDT
HHDDT
50000
0
10000
50000
1030000
1739000
1021000
950000
947000
1324000
1647000
1085000
1089000
991000
1340000
1557000
1147000
1233000
1036000
1356000
1475000
1207000
1375000
1082000
1372000
1403000
1265000
1514000
1130000
1387000
1347000
1320000
1649000
1178000
1406000
1313000
1372000
1782000
1228000
1425000
1292000
1422000
1910000
1280000
1447000
1279000
1471000
2035000
1331000
1470000
1273000
1518000
2154000
1382000
1495000
1278000
1564000
2269000
1433000
1522000
1304342
1641967
2358385
1480953
1567660
1337082
1700702
2441790
1529727
1614690
1400006
1769603
2524688
1577892
1663130
1450056
1830334
2608796
1628111
1713024
1497873
1885771
2695733
1678278
1764415
1563562
1963019
2783637
1729194
1817348
1625492
2039779
2871590
1782360
1871868
1693813
2121129
2961702
1836918
1928024
1760087
2204005
3053785
1892836
1985865
1812889
2270126
3148289
1950392
2045441
1867276
2338229
3245032
2009566
2106804
1923294
2408376
3342383
2070331
2170008
1980993
2480627
3442655
2132863
2235108
2040423
2555046
3545934
2197138
2302162
2101635
2631698
3652312
2263277
2371226
2164685
2710649
3761882
2331252
2442363
2229625
2791968
3874738
2401189
2515634
2296514
2875727
3990980
2473225
2591103
2365409
2961999
4110710
2547422
2668836
2436372
3050859
4234031
2623844
2748901
2509463
3142385
4361052
2702560
2831368
2584747
3236656
4491883
2783636
2916309
2662289
3333756
4626640
2867146
3003799
-------
Year
1975
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
30957
18311
17830
17368
16904
16496
16163
15830
15506
15218
14944
14683
14452
14232
13994
13804
13607
13405
13246
13070
12926
12774
12628
12489
12351
12236
12128
12011
11904
11793
11692
11690
11690
11690
11690
6-A-7
Table 6-A-4
Annual VMT/Vehicle
LDDV
LDDT
LHDDT
MHDDT
HHDI
8740
N/A
N/A
14566
45950
10888
8795
15311
43233
49722
10652
8650
14964
42123
50130
10545
8543
14551
41270
50546
10511
8438
14142
40516
50944
10617
8348
13805
39875
51424
10928
8220
13550
39425
51853
11340
8137
13258
39105
52355
11737
8106
13019
38903
52830
12193
8149
12834
38902
53330
12667
8168
12679
39036
53836
13137
8199
12572
39303
54367
13258
8044
12459
39171
54367
13321
7999
12394
39060
54367
13104
7918
12347
39003
54367
13031
7885
12307
38934
54367
12994
7883
12268
38904
54367
12821
7800
12237
38891
54367
12703
7731
12218
38863
54367
12556
7658
12201
38840
54367
12446
7591
12189
38823
54367
12446
7591
12177
38808
54367
12446
7591
12169
38795
54367
12446
7591
12169
38786
54367
12446
7591
12169
38778
54367
12446
7591
12169
38773
54367
12446
7591
12169
38769
54367
12446
7591
12169
38768
54367
12446
7591
12169
38768
54367
12446
7591
12169
38768
54367
12446
7591
12169
38768
54367
12446
7591
12169
38768
54367
12446
7591
12169
38768
54367
12446
7591
12169
38768
54367
12446
7591
12169
38768
54367
-------
6-A-8
Table 6-A-5
Fleet Fuel Consumption (gal/year)
Year
LDDV
LDDT
LHDDT
MKDDT
HHDDT
BUSES
Total
1975
2.00E+07
0.00E+00
0.00E+00
1.20E+0E
1.09E + 10
6.09E*08
1.16E+
10
1990
6.70E+08
3.81E+08
1.04E+09
5.62E+09
1.11E + 10
8.94E + 08
1.97E
10
1991
6.10E+Q8
3.95E+08
1.15E+09
5.66E+09
1.12E+10
9.20E+08
1.99E
10
1992
5.60E+08
4.09E+08
1.25E+09
5.73E+09
1.12E+10
9.49E+08
2.01E
10
1993
5.19E+08
4.22E+08
1.34E+09
5.81E+09
1.12E+10
9.66E+08
2.03E
10
1994
4.90E+08
4.35E+08
1.43E+09
5.91E+09
1.13E+10
9.91E+-08
2.05E
10
1995
4.76E+08
4.44E+08
1.51E + 09
6.03E + 09
1.14E+10
1.02E+09
2.08E
10
1996
4.74E+08
4.53E+08
1.57E+09
6.17E+09
1.15E+10
1.05E+09
2.12E
10
1997
4.76E+08
4.64E+08
1.64E+09
6.34E+09
1.16E+10
1.07E+09
2.16E
10
1998
4.83E+08
4.78E+08
1.70E+09
6.53E+09
1.17E*10
1.10E+09
2.20E
10
1999
4.93E+08
4.90E+08
1.7 6E+09
6.74E+09
1.19E+10
1.13E+09
2 .25E
10
2000
5.07E+08
5.02E+08
1.81E+09
6.98E+09
1.21E+10
1.16E+09
2.31E
10
2001
5.17E+08
5.13E+08
1.85E+09
7 .14E+09
1.2 3E + 10
1.18E+09
2.35E
10
2002
5.28E+08
5.25E+08
1.88E+09
7.32E+09
1.26E+-10
1.21E+09
2.41E
10
2003
5.40E+08
5.38E+08
1.92E+09
7.51E+09
1.29E+10
1.24E+09
2.46E
10
2004
5.53E+08
5.52E+08
1.96E+09
7.71E+09
1.32E+10
1.27E+09
2.53E
10
2005
5.67E+08
5.66E+08
2.01E+09
7.93E+09
1.36E+10
1.30E+09
2.59E
10
2006
5.82E+08
5.81E+08
2.06E+09
8.15E+09
1.39E+10
1.33E+09
2.66E
10
2007
5.97E+08
5.97E+08
2.11E+09
8.38E+09
1.43E+10
1.36E+09
2.74E
10
2008
6.14E+08
6.14E+08
2 .16E + 09
8.62E+09
1. 47E+-10
1. 39E+09
2.81E
10
2009
6•31E+08
6.31E+08
2.22E+09
8.87E+09
1.51E+10
1.43E+09
2.89E
10
2010
6.48E+08
6.49E+08
2.28E+09
9.13E+09
1.56E+10
1.46E+09
2.97E
10
2011
6.67E+08
6•67E+08
2.3 5E+09
9.40E+09
1.60E+10
1.50E+09
3.06E
10
2012
6.86E+08
6.87E + 08
2.41E+09
9.68E+09
1.65E + 10
1.53E+09
3.15E
10
2013
7.06E+08
7.06E+08
2.48E+09
9.97E+09
1.70E+10
1.57E-t-09
3. 24E
10
2014
7.27E+08
7.27E+08
2.56E+09
1.03E+10
1.7 5E + 10
1.61E+09
3.34E
10
2015
7.49E+08
7.48E+08
2.63E+09
1.06E+10
1.80E+10
1.65E+09
3.43E
10
2016
7.71E+08
7 .71E + 08
2.71E+09
1.09E+10
1.85E+10
1.69E+09
3.54E
10
2017
7.94E+08
7 .93E+08
2.79E+09
1.12E+10
1.91E+10
1.73E+09
3.64E
10
2018
8.18E+08
8.17E + 08
2 . 87E+09
1.15E+10
1.97E+10
1.77E+09
3.75E
10
2019
8.42E+08
8 .42E+08
2 .96E+09
1.19E+10
2.03E+10
1.81E+09
3.86E
10
2020
8.67E+08
8.67E+08
3.04E+Q9
1.2 2E + 10
2.09E+10
1.87E+09
3.98E
10
2021
8.93E+08
8.93E + 08
3.14E+Q9
1.26E+10
2.15E+10
1.92E+09
4.09E
10
2022
9.20E+08
9.19E+08
3.2 3E + 09
1.30E+10
2.21E+10
1.98E+09
4.22E
10
2023
9.48E+08
9 • 47E + 08
3.33E+09
1.34E+10
2.28E+10
2.04E+09
4. 34E
10
-------
6-A-9
Table 6-A-6
Urban Travel Fractions [5]
Vehicle Class
Gasoline
Diesel
Class 2B
0.52
0.64
Class 3
0.44
0.45
Class 4
0 . 44
0 .45
Class 5
0.44
0.45
Class 6
0.31
0.53
Class 7
0 .46
0.51
Class 8A
0.25
0 .42
Class 8B
0.08
0.26
Transit Buses
1.00
1.00
Commercial Buses
0.26
0.26
School Buses
0.34
0.34
-------
Chapter 7
Cost Effectiveness of Fuel Controls
This chapter develops the methodology for and provides
estimates of the cost effectiveness of diesel fuel controls.
Cost effectiveness is defined here to be the net cost per ton
of pollutant removed and is used to relatively rank control
programs.
The chapter is divided into three parts. The first part
gives an overview of the pollutant reductions, and cost and
credit elements used in estimating net costs. The second part
describes the specific methodologies used to estimate net
costs. Included in this section are explanations of the
derivation of the total refinery cost, the technology credit,
the wear credit and fuel consumption credits. The third part
presents and discusses the cost effectiveness results.
A. Overview
The purpose of this section is to present an overview of
the major components of the cost effectiveness equation.
Again, the cost effectiveness is expressed as the net cost per
ton of pollutant removed. This explanation will start by
discussing the pollutants being affected, and then the costs
associated with fuel controls.
The emission analysis in Chapter 6 showed that diesel fuel
controls have an impact on a number of different pollutants:
PM, SO2, HC, and CO. While the control of each of these
pollutants is important, the purpose of the cost effectiveness
analysis in this RIA is to compare diesel fuel controls as
particulate reduction strategies with other particulate
reduction strategies. Therefore, in this analysis, cost
effectiveness will be expressed in terms of dollars per ton of
particulate emission reductions only. The control of emissions
of SO2, HC, and CO, while important, is only of secondary
importance in this analysis, and will be presented merely as a
beneficial "by-product" of fuel control,
Turning first to the net cost of fuel control, it is clear
that the only actual cost is the refinery cost of fuel
controls, as was discussed in Chapter 2. All other cost
components are credits. First, there is a technology credit
for both sulfur and aromatics control, which is the result of
engines not needing as much or as costly aftertreatment
technology. Next, there is a fuel consumption credit which
also results from fewer trap systems in-use with fuel
controls. (The increase in volumetric fuel consumption
associated with aromatics control, due to the reduction in fuel
energy density, has been accounted for in Bonner and Moore's
-------
7-2
refinery modeling, and is included in the refinery costs
presented in Chapter 2.) Finally, there is an engine wear
credit for sulfur control (but not for aromatics control),
which was discussed in Chapter 5. All of these individual
costs and credits are combined to obtain the net cost of each
fuel control scenario.
Another important aspect of the cost effectiveness
analysis is the method and time period over which costs,
credits, and emission reductions are estimated. Two basic
approaches have been used to evaluate motor vehicle controls,
one focusing on a new vehicle over its life and the other
focusing on one or more calendar years. In the vehicle based
approach, both costs and emission reductions are estimated over
the life of the vehicle. In the calendar year or annual
approach, net costs and emission reductions are estimated for
each individual calendar year. The first type of approach is
most useful when the controls primarily affect the technology
of specific model year vehicles. However, commercial fuel
controls affect both existing and new vehicles to varying
degrees, so the calendar year approach is more appropriate.
The annual cost effectiveness of a control strategy (i.e.,
fuel controls) can change dramatically with calendar year. For
example, aromatics control may have a significant impact on
emissions of the in-use fleet now, but as more vehicles are
designed to meet lower emissions standards in the future, it
has correspondingly less effect and the cost effectiveness (in
absolute terms) of aromatics control increases. Therefore, it
is usually useful to estimate cost effectiveness for several
different years so that a single more appropriate cost
effectiveness value may be estimated from these individual
annual values.
The approach used in a number of previous EPA rulemakings
has been to estimate the average cost effectiveness over 33
years. This approach includes the initial, start-up years,
during which cost effectiveness values can be either
particularly low or high, as well as including a number of
years where the controls have essentially reached steady state.
In the 33-year cost effectiveness approach, the net
present values of both costs and emission reductions over the
33-year period are determined and then the ratio taken. This
technique, which weights start-up (near-term) costs and
emission reductions slightly higher than long-term costs and
emission reductions, will be used along with the cost
effectiveness values of several selected calendar years to
compare the various fuel scenarios.
-------
7-3
In this analysis, it was assumed that fuel controls could
be implemented as early as the 1992 calendar year. However,
implementation of controls in a year other than 1992 does not
affect the results of each calendar year's cost effectiveness,
since each year's value is independent of the other years. For
the 33-year discounted analysis a start year and end year must
be selected. Since 1994 is the latest that fuel controls can
be implemented in time for the 0.1 g/BHP-hr PM standard, the
33-year discounted analysis will focus on the years 1994
through 2026.
B. Methodology
This section discusses the specific techniques used to
estimate annual costs and credits. Methods used to estimate
total refinery costs are discussed first, followed by the
engine technology credit, the fuel consumption credit
-------
all
1992
1995
2000
2005
2010
2015
2020
2025
1992
1995
2000
2005
2010
2015
2020
2025
7-4
Table 7-1
Annual Refinery Costs
Maximum Segregation Minimum Segregation
Subsequent Subsequent
Sulfur Aromatics Sulfur Aromatics
Refinery Cost ($/qal)
0.0181 0.0235 0.0225 0.0209
Volume of Fuel Treated
-------
7-5
vehicle class, which in turn is the product of the
class-specific credit times the number of vehicles sold in that
class that year.
The sales data, average credit per vehicle and total
technology credit by vehicle class (and for all vehicles
combined) for several selected calendar years are shown in
Table 7-2. The sales data used for this analysis come from the
MFCM, as described in Chapter 6. As described there during the
calendar years 1990-2000, the annual compounded growth rates
for each class are about 2.75 percent. For calendar years
beyond 2000, this has been rounded to three percent per annum
and applied in a compound fashion.
The average credits- per vehicle were taken directly from
Tables 4-3 and 4-5. In 1992, the total technology credit for
sulfur control is estimated to be nearly $71 million. The
largest credit comes from the LHDDVs, which have the highest
projected sales. The credit for aromatics control, however, is
only about $3.2 million, which is derived entirely from LDDTs.
The absence of credits for the other vehicles is due to the
fact that they already are projected to need no aftertreatment
with sulfur control.
Technology credits for sulfur control from 1995 to 2025
range from $142 - 343 million. The reason for the growth in
the credits is the projected growth in vehicle sales. For
aromatics control over the same period the credit ranges from
$19 - 46 million, which is smaller than the sulfur control
credit by about a factor of 10.
3. Fuel Consumption Credit
As discussed in Chapter 3, vehicles with trap systems are
estimated to use two percent more fuel than those with flow
through catalysts or those without any aftertreatment. The
fuel consumption credit represents the savings in fuel costs
derived from those vehicles for which the necessity of a trap
system is eliminated as a result of fuel control.
To estimate the credit in a particular calendar year, two
items are needed. First, an estimate of the fuel that would
have been consumed by vehicles that would have had traps with
current fuel, but that do not with the control fuel, is
needed. The second item needed is an estimate of the cost of
diesel fuel minus taxes. The fuel consumption credit is the
product of these two numbers.
The first item can be estimated for each model year in a
calendar year by multiplying the change in percent of vehicles
equipped with traps from the base to the control fuel by the
total fuel projected to be consumed by that model year, and by
the trap fuel economy penalty (2 percent). The total fuel
-------
7-6
Table 7-2
Annual Technology Credits
Calendar Year Veh Type
Vehicle
Sales
(lOOO's)
Sulfur Control
Savings
Per
Vehicle($)
Annual
Credit
Aromatics Control
Savings
Per
(Smillion) Vehicle($)
Annual
Credit
(Smillion)
1992
LDDT
LHDDV
MHDDV
HHDDV
Bus
99
171
101
104
38
513
61
151
148
168
156
6.04
25.77
14.97
17.52
6.40
70.70
32
0
0
0
0
3.16
0
0
0
_o
3.16
1995
LDDT
LHDV
MHDV
HHDV
Bus
Total
109
183
114
113
42
57
268
318
324
306
6.21
48.95
36.28
36.56
13.58
141.58
32
27
40
39
49
3.48
4.88
4.57
4.38
1.63
18.94
2000
LDDT
LHDDV
MHDDV
HHDDV
Bus
Total
125
210
134
129
46
644
57
268
318
324
306
7.13
56.17
42.65
41.74
14.88
162.57
32
27
40
39
49
4.00
5.60
5.37
4. 99
1.78
21.74
2010
LDDT
LHDDV
MHDDV
HHDDV
Bus
Total
168
282
180
173
62
865
57
268
318
324
306
9.57
75.49
57.32
56.09
20.00
218.47
32
27
40
39
49
5.37
7 . 53
7.22
6.71
2.39
29.22
2025
LDDT
LHDDV
MHDDV
HHDDV
Bus
Total
262
440
281
270
96
1,216
57
268
318
324
306
14.92
120.0
89.29
87.39
31.76
342.76
32
27
40
39
49
8.37
11.73
11.24
10.46
3.73
45. 53
-------
7-7
consumed by vehicles of a specific model year is the quotient
of total miles driven by the model year and the model year fuel
economy. The total miles driven by a model year is the product
of that year's travel fraction and total VMT for that vehicle
class.
The total credit can then be obtained by summing the
credit gallons for all model years in a calendar year, and for
all vehicle classes, and multiplying this sum by the cost of
diesel fuel minus taxes. This cost was determined to be
73.6^/gal for $20 per barrel crude oil, based on DOE distillate
fuel price information.[1]
The changes in the percentages of trucks equipped with
traps, derived from data in Tables 4-2 and 4-4, are shown in
Table 7-3. Travel fractions for each model year are estimated
as the product of the registration, VMT, and diesel sales
fractions distributions, which were discussed in Chapter 6 and
presented in Tables 6-3 through 6-5. VMT and fuel economy
estimates for each vehicle class were taken from Appendix 6-A.
Total fuel consumption by model year group and vehicle
class and the fuel consumption credit for each vehicle class
for selected calendar years is shown in Table 7-4. The fuel
consumption credit for sulfur control ranges from $12 million
in 1992 to $400 million in 2025. MHDDVs and HHDDVs account for
more than 50 percent of the credit. The fuel credit starts
lower and ends higher than the technology credit, because it
only affects post - 1991 vehicles and grows only as the fleet
turns over. When a truck is sold, its technology credit is
counted only in the year it is sold. However, the fuel credit
for a vehicle accrues over every calendar year it is operated.
In 1992 or 1995, there are few vehicles that are accumulating
fuel credits. As time goes on, however, more of the fleet is
accumulating these credits until in the year 2015 all of the
fleet is accumulating fuel consumption credits. The growth in
the credit between 2015 and 2025 is due to the growth in fuel
consumption over that time period. The fuel credit for
subsequent aromatics control ranges from almost zero in 1992 to
$41 million in 2025, and again is about one-tenth the size of
the fuel credit due to sulfur control.
4. Wear Credit
The wear credit is the result of lower engine wear with
low sulfur fuel. The potential effects of lower engine wear on
oil change interval, engine and vehicle life were discussed in
Chapter 5. Generally, the conclusions of Chapter 5 were that
reduced wear could result in lower engine oil cost and less
frequent oil change intervals, or longer engine and vehicle
life, or longer engine life with fewer total rebuilds.
Benefits were estimated for each of these possible scenarios in
terms of tf/mile and were presented in Tables 5-4, 5-14 and
-------
Table 7-3
Changes in Percentage of Trucks
Equipped With Traps
Percent Equipped Difference
Veh
MYR
Base
Sulfur
Aromatics
Sulfur
Aromatics
LDDT
91-93
25.0
21.7
19.9
3.3
1.8
94+
25.0
21. 6
19.8
3.4
1.8
LHDV
91-93
27.5
0
0
27.5
4.8
94+
100
29 . 5
24.7
70 . 5
4.8
MHDV
91-93
20.6
0
0
20.6
0
94+
98.6
36.0
27.0
62.6
9.0
HHDV
91-93
8.9
0
0
8.9
0
94+
88.0
36.5
32.0
51.5
4.5
Buses
91-93
100.0
67.8
61.4
32.2
6.4
94+
100.0
34 . 0
28.6
66. 0
5.4
-------
7-9
Table 7-4
Fuel Consumption Credits
Calendar
Fuel Consumption
Fuel Consumption
Year
Veh Type
91-93*
94+*
Sulfur
1992
LDDT
57
0
0 . 03
LHDV
392
0
1.59
MHDV
1584
0
4.80
HHDV
2779
0
3 . 64
Bus
321
0
1.52
Total
5133
0
11 .58
1995
LDDT
100
27
0 . 62
LHDV
575
425
6 . 74
MHDV
2154
1602
21 .29
HHDV
3840
2814
26.36
Bus
430
308
5.03
Total
7099
5176
60 . 04
2000
LDDT
79
256
0 .17
LHDV
234
1358
15 . 04
MHDV
868
5139
50 . 00
HHDV
1518
8744
68 . 28
Bus
351
1032
11.69
Total
3050
16529
145.18
2010
LDDT
40
609
0 .32
LHDV
83
2201
23. 18
MHDV
327
8808
82 . 16
HHDV
588
15291
120.00
Bus
127
2331
23.25
Total
1165
29240
248.91
2025
LDDT
0
1005
0.50
LHDV
0
3530
36.63
MHDV
0
14190
130 . 0
HHDV
0
24669
190 . 0
Bus
0
3823
37 . 47
Total
0
47217
394.27
Aromatics
0
0
0
0
0
02
30
0 .32
0 . 03
0.30
2. 12
1 . 86
0 . 65
4.96
0
0
6
5
1.
89
96
81
79
15
14.80
72
55
11. 67
10. 13
1.97
24 . 49
0 .27
2.49
18. 80
16.34
3 . 04
40 .94
Vehicle model years.
-------
7-10
5-16. All that is needed to estimate an annual wear credit is
to multiply these benefit estimates by total VMT per calendar
year for each vehicle class. As discussed in Chapter 5, the
engine wear benefit in which there is an extension in engine
and vehicle life, and the engine wear benefit where there is an
extension in engine life and a reduction in rebuilds are both
applicable to all vehicles, regardless of age, once low sulfur
fuel is implemented. Therefore, total VMT by vehicle class and
calendar year is used in estimating the wear benefits for these
two engine wear benefits analyses. The oil change interval
benefit, however, is applicable only to 1991 and later HDDVs.
Therefore, the VMT of 1991 and later model years is needed to
estimate this benefit.
These two categories of VMTs are shown in Table 7-5. The
first set of VMTs are the total VMTs for each vehicle class in
each calendar year. These come directly from Table 6-A-2. The
second set of VMTs were obtained by multiplying these total
VMTs by the 1991 and later travel fractions (as defined in
Chapter 6) for each vehicle class in each calendar year.
Total wear benefits for each of these three scenarios by
calendar year and vehicle class are shown in Table 7-6. The
column headed "Engine and Vehicle Life" shows the estimated
credits for the wear scenario in which there is both an
extension in engine and vehicle life (benefits are shown in
Chapter 5, Table 5-14). The total credit ranges from $2.1 to
$5.7 billion. This credit is three and six times the refining
cost for sulfur controls with minimum and maximum segregation,
respectively. The column headed "Rebuild Interval and Number
of Rebuilds" shows the credit due to extending rebuild
intervals and reducing the number of rebuilds, but not
extending vehicle life. The total credit ranges from $411
million to $1.1 billion, slightly larger than the refinery cost
of sulfur control for maximum segregation, and about one-half
the refinery cost of sulfur control for minimum segregation.
The last column headed "Oil Cost and Change Interval" shows the
total credit from an increase in oil change intervals and a
slight decrease in oil cost per quart (lower total additive
content due to low sulfur fuel). The credit ranges from $55
million to $611 million. The credit is only applicable to 1991
and later vehicles, which explains the rapid growth in the
credit. In 2025, this credit is between 31 percent (minimum
segregation) and 74 percent (maximum segregation) of the
refinery cost. In all cases, the credits for MHDDVs and HHDDVs
together are at least 65 percent of the total credit.
5. Cost Effectiveness Results
A summary of all the inputs to cost effectiveness and the
resulting cost effectiveness values for sulfur and aromatics
control is shown for several calendar years in Table 7-7.
-------
7-11
Table 7-5
Type
All
1991 &
later
Future VMT Projections
(1010 miles per year)
Calendar
Year LDDV LDDT LHDDV MHDDV HHDDV Buses
1992
1995
2000
2010
2025
1992
1995
2000
2010
2025
1. 64
1.47
1.68
2.26
3 .50
0 .12
0 .34
0 . 93
2.26
3 . 50
0.98
1.09
1.28
1 .72
2.67
0 .15
0 .43
0.91
1.72
2.67
1. 79
2.23
2.85
3 .83
5.96
0 . 55
1.69
2.71
3 .83
5.96
4 .28
4 . 64
5.63
7 .57
11. 70
1.11
3 .22
5.20
7.57
11 . 70
6.85
7.29
8.27
11 .10
17.20
1.68
4 .92
7 . 59
11 .10
17.20
0.86
0 .98
1.21
1 .62
2 .51
0 .15
0 . 52
1 . 02
1 .62
2.51
-------
7-12
Table 7-6
Engine Wear Credits
($million/yr)
Calendar
Year
1992
1995
2000
2010
2025
Type
Engine &
Vehicle
Life
Rebuild
& No.of
Rebuilds
Oil Cost
& Change
Internal
LDDV
149
0
0 , 60
LDDT
99
0
0 . 75
LHDDV
259
0
2 . 76
MHDDV
616
132
21.2
HHDDV
842
253
25.2
Bus
149
26
4.2
Total
2,114
411
54.7
LDDV
133
0
1.7
LDDT
110
0
2.2
LHDDV
323
0
8.5
MHDDV
668
143
61.2
HHDDV
896
269
73.7
Bus
169
29
14.6
Total
2,299
441
161.9
LDDV
153
0
4 . 7
LDDT
129
0
4 . 6
LHDDV
412
0
13.5
MHDDV
810
174
99 . 0
HHDDV
1,016
305
113 .8
Bus
210
37
28 . 4
Total
2,730
516
264.0
LDDV
206
0
11.3
LDDT
173
0
8 . 6
LHDDV
554
0
19. 1
MHDDV
1,089
234
144 . 0
HHDDV
1,363
409
166 . 0
Bus
281
50
45.2
Total
3,666
693
394 .2
LDDV
319
0
17 . 5
LDDT
269
0
13 .4
LHDDV
862
0
29 . 6
MHDDV
1,683
362
223.0
HHDDV
2,112
634
257.7
Bus
435
77
70 . 0
Total
5,680
1,073
611.2
-------
7-13
Table 7-7
Annual Cost Effectiveness of Sulfur
and Aromatics Control
Calendar
Year
1992
1995
2000
2010
2025
Item
Sulfur
Control
Aromatics
Control
Max Seqr
Min Seqr
Max Seqr
Min Seqi
PM Reductions (tons)
36,866
91,696
5,601
5,601
Refinery Cost
364
866
472
804
Technology Credit
71
71
3
3
Fuel Credit
12
12
0
0
Wear Credit*
55
55
0
0
Net Costs
357
729
469
801
Cost Effectiveness($/ton)
6,152
7,947
83,710
142,946
PM Reductions (tons)
35,353
94,526
3,670
3,670
Refinery Cost
376
914
439
849
Technology Credit
142
142
19
19
Fuel Credit
60
60
5
5
Wear Credit*
162
162
0
0
Net Costs
13
551
465
825
Cost Effectiveness($/ton)
381
5,830
126.670
224,846
PM Reductions (tons)
38,140
105,373
1,311
1,311
Refinery Cost
413
1026
543
953
Technology Credit
163
163
22
22
Fuel Credit
145
14 5
15
15
Wear Credit*
264
264
0
0
Net Costs
-154
455
506
917
Cost Effectiveness($/ton)
-4,030
4,314
386,194
699,332
PM Reductions (tons)
49,458
136,369
417
417
Refinery Cost
538
1,323
700
1,230
Tech Credit
218
218
29
29
Fuelnology Credit
246
246
25
25
Wear Credit*
395
395
0
0
Net Costs
-321
464
643
1,174
Cost Effectiveness(J>/ton)
-6,495
3,402
1.54xl06
2 .82x10'
PM Seductions (tons)
76,570
204,313
561
561
Refinery Cost
833
2,000
1,080
1, 860
Technology Credit
340
340
46
46
Fuel Credit
392
392
41
41
Wear Credit*
611
611
0
0
Net Costs
-511
657
995
1,772
Cost Effectiveness($/ton)
-6,674
3,206
1.77xl06
3.16x10
Extension of oil change interval only.
-------
7-14
Costs, emission reductions, and cost effectiveness values for
sulfur and aromatics control, for both the minimum and maximum
(NPRA) segregation scenarios, are shown therein.
As can be seen in the Table, the cost-effectiveness values
shown for sulfur control in the maximum segregation scenario
decrease with time, from $6,152 per ton to -$6,674 per ton in
2025. This is due to the increase in the aftertreatment
technology, fuel consumption, and wear credits in later years,
as 1991 and later model year engines replace the current
fleet. Beginning in the year 2000, net costs for sulfur
control in the maximum segregation scenario actually become
negative, indicating that societal savings will outweigh the
cost to refiners of producing low sulfur fuel. Under the
minimum segregation scenario, cost effectiveness values also
decrease with time, from $7,947 per ton in 1992 to $3,206 per
ton in 2025. In the minimum segregation case, fuel economy,
technology, and engine wear credits do not quite offset the
greater costs to the refining sector associated with treating a
larger volume of fuel.
The estimates in Table 7-7 show the cost effectiveness of
sulfur control assuming reduced engine wear will result only in
an increase in oil change interval. Cost effectiveness results
under the other two wear benefits scenarios are shown in Table
7-8. The long term (2025) cost effectiveness with the
"increase in engine life and reduction in rebuilds engine wear"
scenario ranges from -$12,948 to $860 per ton. For the
"increase in engine and vehicle life" scenario, long term cost
effectiveness ranges from -$72,996 to -$21,589/ton. These
estimates indicate that the method of translating the impact of
reduced sulfur level on engine wear into an economic benefit
can have a large impact on the cost effectiveness of sulfur
control.
The calculated cost effectiveness of aromatics control, as
shown in Table 7-7, continually increases over time. As newer
technology engines replace the current fleet, the emission
reductions resulting from aromatics control continue to
decrease, while refining costs continue to grow. Aromatics
cost effectiveness ranges from $83,710 per ton in 1992 to more
than $3 million per ton in 2025.
The emission reductions used in the above analysis of the
cost effectiveness of aromatics control were based on the
results of VE-l testing on a Cummin's NTCC engine as discussed
in Chapter 4. However, the analysis in Chapter 4 also
discussed emission reductions on a Caterpillar engine that was
tested in a joint Mobil/Caterpillar testing program. The
emission reductions resulting from aromatics control were much
higher on this engine than on the former. While the
Caterpillar engine tested is slightly older technology, the
data obtained from it may be representative of older engines
-------
7-15
Table 7-8
Effect of Higher Engine Wear Credits on
Cost Effectiveness of Sulfur Control
Sulfur Control Cost
Effectiveness,$/ton
Wear Benefit Type
CYR
Maximum
Minimum
Engine Life & Rebuilds
1992
-3,723
3,976
1995
-7,772
2,781
2000
-10,873
1,837
2010
-12,770
1,127
2025
-12,948
860
Engine and Vehicle Life
1992
-49,768
-14,536
1995
-60,150
-16,809
2000
-68,786
-19,125
2010
-72,758
-20,630
2025
-72,996
-21,589
-------
7-16
still in the fleet in the mid-1990" s. Thus, as a sensitivity
case, the percentage emission reductions resulting from fuel
aromatics control on the Caterpillar engine have been applied
to all pre-1991 heavy-duty engines. New fleet particulate
emissions were developed and new cost effectiveness values
generated. The results are shown in Table 7-9. Estimated
particulate reductions for 1992 are increased by a factor of
three, while long term reductions (2025) are the same due to
the sensitivity emission benefits being applicable to only
pre-1991 engines. The 1992 cost effectiveness of aromatics
control with this sensitivity case is about one-third the cost
effectiveness in the previous case, but still over $25,000 per
ton of PM. Long term cost effectiveness is identical to the
previous case, ranging from $1.7 to $3.2 million per ton.
The 33-year cost effectiveness analysis for both sulfur
and aromatics control is shown in Table 7-10. Cost
effectiveness results for sulfur control are shown for all
three engine wear benefit scenarios, as well as results
assuming that no engine wear benefits exists. The 33-year cost
effectiveness of sulfur control ranges from $2,800 to $6,800
per ton if no engine wear benefit is assumed, and is
significantly less when wear benefits are claimed. For
aromatics control two cases are shown, the first based on
emission reductions from the Cummins engine, and the second
case, a sensitivity case, based on emission reductions from the
Mobil/Caterpillar study. The 33-year cost effectiveness values
based on the Cummins engine range from $310,000 to $560,000 per
ton. The sensitivity case values are less than one-half of the
first case, but still are in excess of $130,000 per ton.
The cost effectiveness of aromatics control ($310-560,000
per ton) is significantly higher than that of particulate trap
technology as estimated in the NOx/Particulate RIA.[2] In that
document, the cost-effectiveness of the 1994 heavy-duty
particulate standard was estimated there to be at most $18,300
per ton ($18,700 in 1986 dollars), a difference of
approximately $300,000 per ton. However, as seen in Chapter 6,
in addition to particulate control, the aromatics control
scenario evaluated here also results in significant reductions
in HC and CO (10,348 and 28,794 annualized tons, respectively,
over the 33-year period). In order to render fuel aromatics
control equivalent in cost-effectiveness to the 1994 heavy-duty
particulate standard, the 33-year discounted annual cost of
aromatics control would have to be only $33.5 million, rather
than the $560 million shown in Table 7-10. If the remaining
$526.5 million were apportioned equally to the control of HC
and CO, the resulting cost effectiveness of control for these
two pollutants would be $25,400 and $9,100 per urban ton,
respectively. These values are significantly greater than
those of control strategies which have been implemented or
which are currently under consideration by the Agency. Thus,
even considering the added benefit of HC and CO control,
aromatics control is not cost effective relative to the 1994
particulate standard.
-------
7-17
Table 7-9
Effect of a Higher Emissions Effect on the
Cost Effectiveness Of Aromatics Control
Calendar Year
1992
1995
2000
2010
2025
Emission Reductions(tons)
PM
17,046
9,020
2,930
417
561
HC
25,314
16,547
9,895
7,312
10,852
Cost Effectiveness($/ton)
Max
27,505
51,539
172,799
1,542,498
1,772,763
Min
46,970
91,484
312,909
2,815,306
3,157,770
-------
7-18
Table 7-10
33-Year Cost Effectiveness Analysis
Engine Wear Level of Segregation
Control
Credit
Item
Maximum
Minimum
Sulfur
Oil Interval
Emissions(tons)
Costs(106)
C/E($/ton)
42,751
-170
-3,906
116,588
500
4,304
Eng,Veh Life
Emission(tons)
Costs(106$)
C/E($/ton)
42,751
-2,900
-68,148
116,588
-2,200
-19,253
Eng Life, Reb
Emissions(tons)
Costs(106$)
C/E($/ton)
42,751
-460
-10,867
116,588
200
1,752
No Wear Credit
Emissions(tons)
Costs(106$)
C/E($/ton)
42,751
120
2,826
116,588
790
6, 773
Aromatics
Emissions(tons)
Costs(106$)
C/E($/ton)
1,790
560
310,751
1, 790
1,010
560,378
Aromatics
High Emissions
Sensitivity
Emissions(tons)
Cost
C/E($/ton)
4,128
560
135,789
4, 128
1,010
244,868
-------
7-19
References, Chapter 7
1. "Petroleum Marketing Monthly," Energy Information
Administration, DOE/EIA-0380(88/06), June, 1988.
2. "Regulatory Impact Analysis, Oxides of Nitrogen
Pollutant Specific Study and Summary and Analysis of Comments -
Control of Air Pollution from New Motor Vehicles and New Motor
Vehicle Engines: Gaseous Emission Regulations for 1987 and
Later Model Year Light-Duty Vehicles, and for 1988 and Later
Model Year Light-Duty Trucks and Heavy-Duty Engines;
Particulate Emission Regulations for 1988 and Later Model Year
Heavy-Duty Diesel Engines," EPA, OAR, OMS, March 1985. Docket
A-80-18.
------- |